AAAI2026 AAAI2026 accepted papers AAAI2026 paper list AI paper notes top conference papers 3D Vision Image Generation Medical Imaging Multimodal VLM Model Compression Reinforcement Learning Autonomous Driving AI Safety

🤖 AAAI2026 Accepted Papers¶

1381 AAAI2026 paper notes covering 3D Vision (79), Image Generation (79), Medical Imaging (76), Multimodal VLM (75), Model Compression (60), Reinforcement Learning (58), Autonomous Driving (56), AI Safety (45) and other 52 areas. Each note has TL;DR, motivation, method, experiments, highlights, and limitations — 5-minute reads of core ideas.

💡 LLM Reasoning (37)¶

MathSmith: Towards Extremely Hard Mathematical Reasoning by Forging Synthetic Problems with a Reinforced Policy: This paper proposes MathSmith, a framework that generates hard mathematical problems by randomly sampling concept pairs from PlanetMath, applying 9 predefined difficulty strategies, and jointly optimizing structural validity, reasoning complexity, and answer consistency via GRPO-based reinforcement learning. The resulting high-difficulty synthetic problems significantly improve LLM mathematical reasoning on AIME and OlympiadBench.
Trade-offs in Large Reasoning Models: An Empirical Analysis of Deliberative and Adaptive Reasoning over Foundational Capabilities: This paper systematically evaluates the negative impact of deliberative reasoning on foundational capabilities (helpfulness and harmlessness) in Large Reasoning Models (LRMs) such as DeepSeek-R1, QwQ, and OpenThinker. It finds that deliberative reasoning significantly degrades instruction-following and safety, and proposes adaptive reasoning modes—Zero-Thinking, Less-Thinking, and Summary-Thinking—that effectively mitigate these deficiencies.
CMMCoT: Enhancing Complex Multi-Image Comprehension via Multi-Modal Chain-of-Thought and Memory Augmentation: This paper proposes the CMMCoT framework, which constructs interleaved multimodal multi-step reasoning chains (with visual region token supervision) and a test-time retrieval-based memory augmentation module (RIFREM) to enhance slow-thinking reasoning in multi-image scenarios without increasing model parameters. Built on Qwen2.5-VL-7B, the method achieves an average improvement of 1.4 points on multi-image benchmarks.
Beyond ReAct: A Planner-Centric Framework for Complex Tool-Augmented LLM Reasoning: This paper proposes a Planner-centric Plan-Execute framework that transforms complex queries into DAG-based execution plans. Through two-stage SFT+GRPO training of a dedicated Planner model, the approach surpasses reactive methods such as ReAct on ComplexTool-Plan and StableToolBench, achieving higher success rates with fewer inference steps.
ExtendAttack: Attacking Servers of LRMs via Extending Reasoning: This paper proposes ExtendAttack, a resource exhaustion attack targeting Large Reasoning Models (LRMs): by randomly converting characters in the prompt into multi-base ASCII encodings, the attack forces models to perform extensive character-by-character decoding before answering, increasing o3's response length by more than 2.7× and doubling latency, while keeping answer accuracy largely intact.
Relation-R1: Progressively Cognitive Chain-of-Thought Guided Reinforcement Learning for Unified Relation Comprehension: This paper proposes Relation-R1, the first unified framework for binary and N-ary relation comprehension, combining progressively cognitive CoT-guided SFT with GRPO multi-reward optimization. With only 3B parameters, it surpasses 13B models, achieving 21.20% Mean (+6.87%) on PSG and state-of-the-art performance across all metrics on SWiG (Grnd-all 30.18%, +14.48%).
Text-to-Scene with Large Reasoning Models: This paper proposes Reason-3D, which leverages the multi-step spatial reasoning capabilities of large reasoning models (LRMs) to achieve zero-shot text-to-3D scene generation via semantic-voting-based object retrieval and a two-stage layout strategy (autoregressive placement + collision-aware refinement). The system achieves an Elo score of 2248 in human evaluation, substantially outperforming Holodeck (1500) and LayoutVLM (1650).
Graph of Verification: Structured Verification of LLM Reasoning with Directed Acyclic Graphs: This paper proposes Graph of Verification (GoV), a structured verification framework that models LLM reasoning processes as directed acyclic graphs (DAGs). Through a flexible Node Block architecture, GoV enables multi-granularity verification—ranging from atomic-level steps in formal tasks to paragraph-level verification in natural language narratives—and substantially outperforms both holistic verification and other decomposed verification methods on both structured and loosely structured reasoning benchmarks.
Small Language Models for Efficient Agentic Tool Calling: Outperforming Large Models with Targeted Fine-tuning: By applying a single-epoch supervised fine-tuning (SFT) on OPT-350M, this work achieves a 77.55% pass rate on ToolBench, substantially outperforming large-model baselines such as ChatGPT-CoT (26%) and ToolLLaMA-DFS (30.18%), demonstrating that small models with targeted fine-tuning can significantly surpass general-purpose large models on specific tasks.
Well Begun, Half Done: Reinforcement Learning with Prefix Optimization for LLM Reasoning: This paper identifies the Beginning Lock-in Effect (BLE) in LLM reasoning — the initial reasoning steps significantly determine subsequent trajectories and final outcomes. Based on this finding, the paper proposes PPPO, a method that optimizes only prefix tokens (approximately 26% of all tokens), achieving accuracy improvements of up to 18.02% while reducing output token counts by up to 18.35%.

Browse all 37 LLM Reasoning papers →

🦾 LLM Agent (33)¶

AutoGLM: Autonomous Foundation Agents for GUIs: AutoGLM builds a GUI foundation agent for web browsers and Android devices on top of ChatGLM. By introducing an intermediate interface design that decouples planning from grounding, and proposing a self-evolving online curriculum reinforcement learning framework, the system achieves a 55.2% success rate on VAB-WebArena-Lite, substantially surpassing GPT-4o's 18.2%.
TongUI: Internet-Scale Trajectories from Multimodal Web Tutorials for Generalized GUI Agents: TongUI proposes a framework that automatically converts multimodal web tutorials (videos and illustrated articles) into GUI operation trajectories, constructing the million-scale GUI-Net-1M dataset for fine-tuning Qwen2.5-VL. The resulting models surpass or approach state-of-the-art methods such as UI-TARS across multiple grounding and navigation benchmarks.
AgentSwift: Efficient LLM Agent Design via Value-guided Hierarchical Search: This paper proposes AgentSwift, a framework that automatically discovers high-performance LLM agent designs through a hierarchical search space (jointly optimizing agentic workflows and functional components), a lightweight value model for predicting agent performance, and an uncertainty-guided MCTS search strategy, achieving an average improvement of 8.34% across 7 benchmarks.
When Refusals Fail: Unstable Safety Mechanisms in Long-Context LLM Agents: This paper systematically investigates how long-context padding affects the safety behavior of LLM agents. Models claiming support for 1M–2M token windows exhibit performance collapse exceeding 50% at 100K tokens. Refusal rates fluctuate in unpredictable directions (GPT-4.1-nano rises from 5% to 40%; Grok 4 Fast drops from 80% to 10%), revealing critical safety vulnerabilities in long-context agent systems.
Agent-SAMA: State-Aware Mobile Assistant: This paper proposes Agent-SAMA, which for the first time introduces a finite state machine (FSM) into mobile GUI agents, modeling UI screens as states and user actions as transitions. Four specialized agents collaborate to achieve state-aware task planning, execution verification, and error recovery, improving success rate by up to 12% and recovery rate by 13.8% on cross-app benchmarks.
ProBench: Benchmarking GUI Agents with Accurate Process Information: ProBench is proposed as the first mobile GUI Agent benchmark that evaluates both final state and operational process: 200+ challenging tasks cover 34 mainstream Chinese and English apps. A Process Provider (Structure Description Converter + MLLM Summarizer) automatically captures accurate intermediate process information. Evaluation reveals that even the strongest model, Gemini 2.5 Pro, completes only 40.1% of tasks, exposing three prevalent issues: insufficient grounding, poor awareness of action history, and oversimplified task planning.
History-Aware Reasoning for GUI Agents: This paper proposes the HAR framework, which transforms the reasoning paradigm of GUI agents from "history-unaware" to "history-aware" by constructing reflective learning scenarios, synthesizing error-correction guidelines, and designing a hybrid RL reward function incorporating a Memory-Augmented Reward (MAR). A 3B model trained under this framework surpasses larger models on multiple benchmarks including AITW, Mind2Web, and GUI-Odyssey.
Prune4Web: DOM Tree Pruning Programming for Web Agent: This paper proposes Prune4Web, a programmatic DOM pruning approach that achieves 25–50× candidate element reduction via "LLM-generated scoring function parameters + fixed heuristic template execution." The three-stage pipeline (Planner decomposes subtasks → Programmatic Filter generates scoring functions to prune DOM → Grounder executes actions) enables a 3B model to achieve 52.4% Step SR on Multimodal-Mind2Web, surpassing all baselines of the same parameter scale and even some 9.6B/32B models, while improving low-level grounding accuracy from 46.8% to 88.28%.
A2Flow: Automating Agentic Workflow Generation via Self-Adaptive Abstraction Operators: This paper proposes A2Flow, a framework that automatically extracts reusable abstract execution operators from expert data via a three-stage pipeline (case generation → functional clustering → deep extraction), replacing manually predefined operators. Combined with an operator memory mechanism that accumulates intermediate outputs to assist node decision-making, A2Flow outperforms AFLOW and other state-of-the-art methods across 8 benchmarks while reducing resource consumption by 37%.
AutoTool: Efficient Tool Selection for Large Language Model Agents: This paper proposes AutoTool, a graph-based tool selection framework that exploits tool usage inertia to construct a Tool Inertia Graph (TIG). By leveraging statistical structure, AutoTool bypasses redundant LLM inference for tool selection and parameter filling, reducing inference overhead by up to 30% while maintaining task completion rates.

Browse all 33 LLM Agent papers →

👥 Multi-Agent (26)¶

Assemble Your Crew: Automatic Multi-agent Communication Topology Design via Autoregressive Graph Generation: This paper proposes ARG-Designer, which reformulates multi-agent system topology design as a conditional autoregressive graph generation task. Rather than pruning from template graphs, the model incrementally generates agent nodes and communication edges from scratch. ARG-Designer achieves state-of-the-art performance across 6 benchmarks (average 92.78%), reduces token consumption by approximately 50% compared to G-Designer, and supports role expansion without retraining.
Scalable and Accurate Graph Reasoning with LLM-Based Multi-Agents: This paper proposes GraphAgent-Reasoner (GAR), inspired by distributed graph computation theory. It decomposes graph problems into node-centric subtasks assigned to multiple agents, which collaborate through neighbor message passing. GAR extends the graph scale tractable by LLMs from 100 nodes to 1,000 nodes, and significantly outperforms existing state-of-the-art methods on polynomial-time graph reasoning tasks.
MedLA: A Logic-Driven Multi-Agent Framework for Complex Medical Reasoning with Large Language Models: This paper proposes MedLA, the first multi-agent medical reasoning framework based on syllogistic logic trees. Each agent organizes its reasoning as an explicit logic tree composed of syllogistic nodes (major premise–minor premise–conclusion). Multiple agents align and revise their logic trees at the premise level through graph-guided multi-round discussions. MedLA outperforms all baselines by 7.4% on MedDDx (8B model) and achieves an average accuracy of 69.9% on medical QA benchmarks with an 8B model, surpassing 70B RAG-based models.
SafeSieve: From Heuristics to Experience in Progressive Pruning for LLM-based Multi-Agent Communication: SafeSieve is proposed as a progressive adaptive multi-agent communication pruning framework. Through a two-stage edge scoring mechanism combining semantic-heuristic initialization and history-feedback-driven refinement, together with 0-extension clustering, SafeSieve achieves 94.01% average accuracy across 6 benchmarks while reducing token consumption by 12.4%–27.8%, and demonstrates inherent robustness against prompt injection attacks.
iMAD: Intelligent Multi-Agent Debate for Efficient and Accurate LLM Inference: iMAD proposes a framework for selectively triggering multi-agent debate (MAD): a single agent first generates a structured response with self-critique, from which 41 interpretable linguistic/semantic features are extracted; a lightweight MLP classifier trained with the FocusCal loss then determines whether to trigger MAD. Across 6 QA/VQA benchmarks, iMAD reduces token overhead by up to 92% while improving accuracy by up to 13.5%.
Shadows in the Code: Exploring the Risks and Defenses of LLM-based Multi-Agent Software Development Systems: The first systematic security analysis of LLM-based multi-agent software development systems (ChatDev/MetaGPT/AgentVerse): proposes the IMBIA attack framework covering two threat scenarios (malicious user + benign agents / benign user + malicious agent) and 12 malicious behaviors across 5 malware families, achieving an attack success rate (ASR) of up to 93% on ChatDev, with the Adv-IMBIA adversarial defense reducing ASR by 40–73%.
FinRpt: Dataset, Evaluation System and LLM-based Multi-agent Framework for Equity Research Report Generation: This paper is the first to systematically define the task of automated Equity Research Report (ERR) generation. It constructs the FinRpt dataset (6,825 high-quality bilingual reports integrating 7 categories of financial data), proposes an 11-metric evaluation framework, and designs the FinRpt-Gen generation framework with 9 collaborative agents featuring a three-stage enhancement pipeline (rating correction / expert review / language polishing). Human evaluation shows that generated reports approach expert-written quality.
LungNoduleAgent: A Collaborative Multi-Agent System for Precision Diagnosis of Lung Nodules: This paper proposes LungNoduleAgent, the first collaborative multi-agent system for lung nodule analysis. It simulates the clinical workflow through a three-stage pipeline—Nodule Spotter, Simulated Radiologist, and Doctor Agent System—and substantially outperforms mainstream VLMs (GPT-4o, Claude 3.7 Sonnet) and medical agents (MedAgent-Pro) on CT report generation and malignancy grading tasks.
EcoAgent: An Efficient Device-Cloud Collaborative Multi-Agent Framework for Mobile Automation: This paper proposes EcoAgent, a closed-loop device-cloud collaborative multi-agent framework for mobile automation. By combining Dual-ReACT two-level reasoning and planning, lightweight on-device verification feedback, and a Pre-Understanding text compression module, EcoAgent achieves success rates comparable to fully cloud-based agents on AndroidWorld while substantially reducing latency (3.9s vs. 15.3s), cloud invocations (−89%), and upstream data volume (−48.6×).
BAMAS: Structuring Budget-Aware Multi-Agent Systems: This paper proposes the BAMAS framework, which employs Integer Linear Programming (ILP) to select the optimal LLM combination under budget constraints, and uses a reinforcement learning policy to choose the best collaboration topology (Linear/Star/Feedback/Planner-Driven). BAMAS achieves accuracy comparable to state-of-the-art multi-agent systems on GSM8K, MBPP, and MATH, while reducing costs by up to 86%.

Browse all 26 Multi-Agent papers →

⚖️ Alignment & RLHF (17)¶

Intrinsic Barriers and Practical Pathways for Human-AI Alignment: An Agreement-Based Complexity Analysis: This paper formalizes AI alignment as an $\langle M,N,\varepsilon,\delta\rangle$-agreement multi-objective optimization problem, proves information-theoretic lower bounds on alignment (encoding "all human values" is fundamentally intractable) from a communication complexity perspective, and derives explicit achievable algorithms and tight upper bounds for unbounded/bounded rational agents, revealing the theoretical basis for the global inevitability of reward hacking in large state spaces.
Align to Structure: Aligning Large Language Models with Structural Information: This paper proposes Structural Alignment, a method that integrates linguistic discourse structure frameworks—surface-level text structure scoring and an RST-based discourse motif classifier—into PPO reinforcement learning training, and introduces a discourse motif-based dense reward mechanism. This enables LLMs to generate more coherent, human-like long-form text, outperforming standard RLHF models on academic essay writing and long document summarization tasks.
DeCoRL: Decoupling Reasoning Chains via Parallel Sub-Step Generation and Cascaded Reinforcement for Interpretable and Scalable RLHF: DeCoRL transforms CoT reasoning from monolithic sequential processing into an "orchestra-style" modular parallel collaboration — nine specialized sub-models (parsing / semantic / entity / fact-checking / style / quality / computation / verification / integration) generate reasoning sub-steps in parallel, coordinated via dual reward attribution (local quality + contribution) and cascaded DRPO optimization, achieving 80.8% on RM-Bench (surpassing all baselines), a 3.8× inference speedup, and a 22.7% improvement in interpretability.
Enhancing Uncertainty Estimation in LLMs with Expectation of Aggregated Internal Belief: This paper proposes EAGLE, a method that estimates uncertainty by aggregating logits from multiple intermediate hidden layers of an LLM and computing the expectation of the resulting confidence distribution. EAGLE requires no additional trainable parameters and reduces ECE from 12.6% to 3.2% while improving AUROC from 59.0% to 61.6% across multiple datasets and models.
Importance-Aware Data Selection for Efficient LLM Instruction Tuning: This paper proposes MIWV (Model Instruction Weakness Value), a metric that measures the importance of each instruction sample for improving model capability by comparing LLM loss with and without a one-shot ICL demonstration. Using only 1% (520 samples) of the Alpaca dataset, the method comprehensively outperforms fine-tuning on the full 52,002 samples.
Differentiated Directional Intervention: A Framework for Evading LLM Safety Alignment: This work deconstructs the internal representations of LLM safety alignment from the conventional "single refusal direction" into two functionally independent directions — a harm detection direction and a refusal execution direction — and proposes the DBDI framework, which applies adaptive projection elimination and direct steering to intervene on each direction separately, achieving a 97.88% attack success rate (ASR) on Llama-2.
GRAM-R²: Self-Training Generative Foundation Reward Models for Reward Reasoning: This paper proposes GRAM-R², a generative foundation reward model that elicits reward reasoning capabilities on unlabeled data via self-training. The model simultaneously produces preference labels and reasoning rationales, consistently outperforming both discriminative and generative baselines across multiple downstream tasks including response ranking, task adaptation, and RLHF.
LaF-GRPO: In-Situ Navigation Instruction Generation for the Visually Impaired via GRPO with LLM-as-Follower Reward: This paper proposes the LaF-GRPO framework, which employs an LLM to simulate the responses of visually impaired users to navigation instructions as a reward signal. By applying GRPO-based post-training to a VLM, the framework generates more precise and safer navigation instructions for the visually impaired. The authors also construct the NIG4VI benchmark dataset comprising 27k samples.
W2S-AlignTree: Weak-to-Strong Inference-Time Alignment for Large Language Models via Monte Carlo Tree Search: This paper proposes W2S-AlignTree, the first inference-time alignment framework that integrates Monte Carlo Tree Search (MCTS) with the weak-to-strong generalization (W2SG) paradigm. It leverages step-level proxy value functions derived from a weak model to guide the generation of a strong model at inference time, achieving significant improvements over baselines across sentiment control, summarization, and instruction-following tasks — with a 15.9% gain on the Llama3-8B summarization task.
AlignTree: Efficient Defense Against LLM Jailbreak Attacks: AlignTree leverages internal LLM activation features — combining linear refusal directions with nonlinear SVM signals — to train a lightweight random forest classifier that efficiently detects jailbreak attacks with negligible computational overhead, achieving state-of-the-art reductions in attack success rate (ASR).

Browse all 17 Alignment & RLHF papers →

🔒 LLM Safety (41)¶

Can Editing LLMs Inject Harm?: This paper reframes knowledge editing as a novel LLM security threat termed Editing Attack, systematically investigating the feasibility of injecting misinformation and bias into LLMs via three editing methods—ROME, FT, and ICE—and demonstrating that such attacks are both highly effective and remarkably stealthy.
LLM Targeted Underperformance Disproportionately Impacts Vulnerable Users: Systematic experiments demonstrate that mainstream LLMs (GPT-4, Claude 3 Opus, Llama 3-8B) exhibit significant discriminatory degradation in information accuracy, truthfulness, and refusal rates toward users with lower English proficiency, lower educational attainment, and non-US backgrounds, making the most vulnerable users the least reliably served.
FedALT: Federated Fine-Tuning through Adaptive Local Training with Rest-of-World LoRA: FedALT is proposed to maintain a trainable Individual LoRA (updated locally) and a frozen Rest-of-World (RoW) LoRA (averaged from other clients) for each client, combined with an adaptive MoE mixer that dynamically balances local and global knowledge. This design fundamentally eliminates cross-client interference caused by FedAvg aggregation, achieving significant improvements over SOTA on heterogeneous-task federated LLM fine-tuning.
PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration: This paper proposes PRISM, a framework that dynamically routes user prompts to one of three inference modes—cloud-only, edge-only, or collaborative—via a context-aware soft gating mechanism. In the collaborative mode, an adaptive two-layer local differential privacy (LDP) scheme and semantic sketch collaboration are employed to achieve a three-way balance among privacy, utility, and efficiency.
RadarLLM: Empowering Large Language Models to Understand Human Motion from Millimeter-Wave Point Cloud Sequence: This paper proposes RadarLLM, the first end-to-end framework leveraging large language models for semantic-level human motion understanding from millimeter-wave radar point cloud sequences. The framework comprises a motion-guided radar tokenizer based on Aggregate VQ-VAE and a radar-aware language model, along with a physics-aware simulation pipeline for generating large-scale paired radar-text data.
AgentSense: Virtual Sensor Data Generation Using LLM Agents in Simulated Home Environments: LLM-driven embodied agents are instantiated to "live" in simulated smart home environments, generating virtual ambient sensor data for pre-training HAR models, which yields significant gains in activity recognition under low-resource settings.
BadThink: Triggered Overthinking Attacks on Chain-of-Thought Reasoning in Large Language Models: This paper proposes BadThink — the first training-time backdoor attack targeting CoT reasoning efficiency. By iteratively optimizing verbose reasoning templates via an LLM, it constructs poisoned data that causes the victim model, upon trigger activation, to generate reasoning chains inflated by over 17× (on MATH-500), while preserving final answer correctness and maintaining strong stealthiness.
From Single to Societal: Analyzing Persona-Induced Bias in Multi-Agent Interactions: This paper presents the first systematic study of persona-induced bias in LLM-based multi-agent interactions. Through controlled experiments on collaborative problem solving and persuasion tasks, three key findings are revealed: (1) different personas exhibit significant divergence in trustworthiness and insistence (dominant groups such as males and White individuals are perceived as less trustworthy); (2) agents display pronounced in-group favoritism; and (3) these biases persist and tend to amplify in multi-turn, multi-agent settings.
WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking: This paper proposes WaterMod, an LLM text watermarking method based on modular arithmetic ($\text{rank} \bmod k$) that partitions the vocabulary into modular residue classes after sorting tokens by probability. Under both zero-bit ($k=2$) and multi-bit ($k>2$) watermarking settings, WaterMod achieves high detection rates and low quality degradation within a unified framework, requiring no external thesaurus or hashing tricks.
Cross-Modal Unlearning via Influential Neuron Path Editing in Multimodal Large Language Models: This paper proposes MIP-Editor, which localizes influential neuron paths encoding forget-target knowledge in MLLMs via cross-layer gradient integration (text branch) and Fisher integration (visual branch), then edits these neurons using path-based Representation Misdirection Unlearning (RMisU), achieving up to 87.75% forget rate and 54.26% improvement in general knowledge retention on MLLMU-Bench.

Browse all 41 LLM Safety papers →

👻 Hallucination Detection (15)¶

LLM-CAS: Dynamic Neuron Perturbation for Real-Time Hallucination Correction: LLM-CAS is the first work to formulate real-time LLM hallucination correction as a hierarchical reinforcement learning (HRL) problem. It trains an RL agent to dynamically select optimal neuron perturbation strategies at inference time — the high-level policy selects a functional network category, while the low-level policy selects perturbation type and magnitude. Combined with adaptive masking and causal tracing for precise neuron localization, LLM-CAS achieves a 10.98% improvement on StoryCloze, outperforming static and dynamic baselines such as ITI, CAA, and SADI.
InEx: Hallucination Mitigation via Introspection and Cross-Modal Multi-Agent Collaboration: This paper proposes InEx, a framework that iteratively verifies and corrects MLLM outputs via internal introspective reasoning (TVER-driven uncertainty-aware visual augmentation) and external cross-modal multi-agent collaboration (textual self-reflection + image editing verification + visual self-reflection), achieving an 8.9% improvement on POPE and consistently outperforming OPERA/VCD/ICD across multiple hallucination and general benchmarks.
PASE: Leveraging the Phonological Prior of WavLM for Low-Hallucination Generative Speech Enhancement: This paper proposes PASE, a framework that leverages robust phonological priors embedded in pretrained WavLM via Denoising Representation Distillation (DRD) to suppress linguistic hallucinations, while employing a dual-stream representation (high-level phonetic + low-level acoustic) to eliminate acoustic hallucinations, simultaneously achieving state-of-the-art performance in both perceptual quality and content fidelity.
Bridging Day and Night: Target-Class Hallucination Suppression in Unpaired Image Translation: This paper is the first to systematically address the "target-class hallucination" problem in unpaired day-to-night image translation. By combining a dual-head discriminator (style head + SAM2 pseudo-label segmentation head) for hallucination detection and class-prototype contrastive learning for suppression, the method improves mAP from 15.08 to 17.40 (+15.5%) on BDD100K day-to-night domain adaptation detection, with traffic light AP improving by 31.7%.
Causally-Grounded Dual-Path Attention Intervention for Object Hallucination Mitigation in LVLMs: This paper proposes Owl, a framework that models visual and textual attention as mediating variables within a structural causal model, introduces the VTACR metric to quantify cross-modal attention imbalance, and designs VTACR-guided adaptive attention modulation combined with a dual-path contrastive decoding strategy, achieving state-of-the-art hallucination mitigation on POPE and CHAIR benchmarks.
When Hallucination Costs Millions: Benchmarking AI Agents in High-Stakes Adversarial Financial Markets: This paper introduces the CAIA benchmark, which leverages cryptocurrency markets as a natural adversarial laboratory to evaluate 17 state-of-the-art LLMs on agent capabilities in high-stakes adversarial environments. Results reveal that frontier models achieve only 67.4% accuracy (GPT-5) compared to a human baseline of 80%, and expose systematic tool selection failures.
Hallucinate Less by Thinking More: Aspect-Based Causal Abstention for Large Language Models: This paper proposes ABCA (Aspect-Based Causal Abstention), a pre-generation abstention framework that employs dual-agent debate to identify "aspect variables" (e.g., discipline, legal context, temporal frame) for activating distinct knowledge branches within LLMs. It applies the AIPW doubly robust estimator to compute causal effects and uses Centroid Angular Deviation (CAD) to detect knowledge conflicts (Type-1) or knowledge insufficiency (Type-2), achieving 91.4% accuracy on TruthfulQA and 96.4% unanswerable question identification rate—far surpassing the baseline of 44%.
MUG: Multi-agent Undercover Gaming — Hallucination Removal via Counterfactual Test for Multimodal Reasoning: MUG reframes Multi-Agent Debate (MAD) as a "Who's Undercover" social reasoning game — by introducing information asymmetry through counterfactual image editing (modifying the reference image), one agent is assigned the edited image $I^-$ as the "undercover," while other agents hold the original image $I^+$ and identify the undercover (i.e., the hallucination source) via reasoning and voting. On HallusionBench, Qwen2.5VL-7B improves from 46.4% to 53.8%.
Listen Like a Teacher: Mitigating Whisper Hallucinations using Adaptive Layer Attention and Knowledge Distillation: A two-stage framework is proposed: Adaptive Layer Attention (ALA) fuses multi-layer representations from the Whisper encoder to enhance noise robustness, while Multi-Objective Knowledge Distillation (MOKD) aligns the semantic and attention distributions of a clean-speech teacher with a noisy-speech student — achieving significant reductions in hallucination rate and WER on multilingual noisy ASR benchmarks.
Ground What You See: Hallucination-Resistant MLLMs via Caption Feedback, Diversity-Aware Sampling, and Conflict Regularization: This paper identifies three root causes of hallucination in RL-based MLLM training—visual misinterpretation, limited exploration diversity, and sample conflict—and addresses each with Caption Reward, reward-variance-guided sample selection, and NTK-similarity-based InfoNCE regularization, achieving significant hallucination reduction across multiple benchmarks.

Browse all 15 Hallucination Detection papers →

📊 LLM Evaluation (16)¶

Lost in Benchmarks? Rethinking Large Language Model Benchmarking with Item Response Theory: This paper proposes PSN-IRT (Pseudo-Siamese Network for IRT), an enhanced Item Response Theory framework that jointly estimates LLM ability parameters and four-parameter item characteristics (difficulty / discrimination / guessing / feasibility). Applied to 41,871 items across 11 benchmarks, the framework reveals systemic issues including widespread saturation, insufficient difficulty ceilings, and data contamination. Item subsets selected by PSN-IRT achieve a ranking consistency of Kendall $\tau = 1.00$.
Benchmarking LLMs for Political Science: A United Nations Perspective: This paper presents UNBench, the first comprehensive LLM evaluation benchmark for political science grounded in UN Security Council records from 1994 to 2024. It encompasses four interrelated tasks—resolution drafting, voting simulation, adoption prediction, and representative statement generation—to systematically assess LLMs' ability to understand and simulate complex political dynamics.
Towards a Common Framework for Autoformalization: This paper systematically surveys existing work on autoformalization across mathematics, logical reasoning, planning, and knowledge representation, and proposes a unified cross-disciplinary definitional framework. Autoformalization is defined as the semantically equivalent transformation from informal language to formal reasoning languages, with the goal of facilitating methodology sharing across research communities and accelerating the development of next-generation AI reasoning systems.
Low-Rank Curvature for Zeroth-Order Optimization in LLM Fine-Tuning: This paper proposes LOREN, a curvature-aware zeroth-order optimization method that captures the anisotropic curvature of the loss landscape via a low-rank block-diagonal preconditioner, combined with REINFORCE Leave-One-Out (RLOO) variance reduction. LOREN achieves higher accuracy and faster convergence in LLM fine-tuning while reducing peak memory by up to 27.3% compared to MeZO-Adam.
ConInstruct: Evaluating Large Language Models on Conflict Detection and Resolution in Instructions: This paper proposes ConInstruct, a benchmark for evaluating LLMs' ability to detect and resolve conflicting constraints in instructions. Results show that most proprietary models can detect conflicts reasonably well but rarely notify users explicitly, with DeepSeek-R1 and Claude-4.5-Sonnet achieving the best conflict detection performance (F1 of 91.5% and 87.3%, respectively).
MCTS-SQL: Light-Weight LLMs can Master the Text-to-SQL through Monte Carlo Tree Search: This paper proposes MCTS-SQL, enabling lightweight LLMs (e.g., Qwen-1.5B) to achieve strong Text-to-SQL performance via Monte Carlo Tree Search — a three-component architecture (Selector for schema pruning + Direct Generator for initial SQL generation + MCTS-Refiner for iterative refinement), combined with a prefix caching mechanism that reduces inference time by 53%. Qwen-1.5B achieves 40.69% execution accuracy on BIRD, surpassing ChatGPT-3.5.
OptScale: Probabilistic Optimality for Inference-time Scaling: This paper proposes OptScale, a probabilistic optimality framework that models the probability distribution of verifier scores to derive a theoretical lower bound on the optimal number of samples, dynamically determining the minimum number of samples required per problem and substantially reducing computational overhead while preserving inference accuracy.
Test-time Diverse Reasoning by Riemannian Activation Steering: This paper proposes SPREAD, an unsupervised test-time activation steering framework that maximizes the total volume spanned by hidden activations across multiple reasoning paths by solving a Riemannian optimization problem on a product of spherical manifolds. SPREAD improves reasoning diversity and accuracy in Best-of-N sampling, outperforming temperature sampling baselines on mathematical reasoning benchmarks.
BCWildfire: A Long-term Multi-factor Dataset and Deep Learning Benchmark for Boreal Wildfire Risk Prediction: This paper introduces BCWildfire, a multimodal wildfire risk prediction dataset covering 240 million hectares of British Columbia, Canada over a 25-year span, encompassing 38 driving factors. It conducts a systematic benchmark evaluation of time series forecasting models across four paradigms—CNN, Linear, Transformer, and Mamba—revealing the performance ceiling of current models and the key influential factors in wildfire prediction.
DiCaP: Distribution-Calibrated Pseudo-labeling for Semi-Supervised Multi-Label Learning: This paper proposes DiCaP (Distribution-Calibrated Pseudo-labeling), which estimates the posterior correctness rate of pseudo-labels to calibrate their weights, introduces a dual-threshold mechanism to separate confident and ambiguous regions with differentiated strategies, and surpasses the state of the art by up to 4.27% in semi-supervised multi-label learning.

Browse all 16 LLM Evaluation papers →

⚡ LLM Efficiency (9)¶

MoETTA: Test-Time Adaptation Under Mixed Distribution Shifts with MoE-LayerNorm: This paper proposes MoETTA, a test-time adaptation framework that reparameterizes LayerNorm into multiple structurally decoupled expert branches. A routing mechanism assigns samples from different domains to different experts, enabling multi-directional parameter updates and overcoming the limitations of a single adaptation path under mixed distribution shifts. The paper also introduces two more realistic evaluation benchmarks—potpourri and potpourri+—and achieves state-of-the-art performance across all settings.
How Many Experts Are Enough? Towards Optimal Semantic Specialization for Mixture-of-Experts: This paper proposes MASS, a framework that adaptively expands the MoE expert pool via gradient-based semantic drift detection, combined with a Top-p confidence routing strategy, to automatically discover the optimal number of experts without hyperparameter search while enhancing semantic differentiation across experts.
Judge Q: Trainable Queries for Optimized Information Retention in KV Cache Eviction: This paper proposes Judge Q, which introduces trainable soft tokens into the model vocabulary and trains their attention patterns to align with those of actual decoding tokens, enabling them to replace local-window queries for evaluating KV cache importance during the prefill stage. This approach better preserves global information, achieving ~1-point improvement on LongBench and 3+ points on RULER.
InterMoE: Individual-Specific 3D Human Interaction Generation via Dynamic Temporal-Selective MoE: This paper proposes InterMoE, a Dynamic Temporal-Selective MoE architecture for text-driven two-person 3D interaction motion generation that addresses individual identity preservation and semantic fidelity. A Synergistic Router fuses semantic and kinematic features to guide routing, while Dynamic Temporal Selection enables each expert to adaptively select key temporal frames. The method achieves a 9% FID reduction on InterHuman and 22% on InterX.
Harnessing the Unseen: The Hidden Influence of Intrinsic Knowledge in Long-Context Language Models: This paper presents the first systematic study of how parametric knowledge influences generation in long-context language models (LCLMs), finding that such influence grows with context length and that methods designed to improve extrinsic retrieval suppress parametric recall. Based on these findings, the paper proposes the Hybrid Needle-in-a-Haystack (Hybrid NIAH) benchmark to jointly evaluate both capabilities.
HN-MVTS: HyperNetwork-based Multivariate Time Series Forecasting: This paper proposes HN-MVTS, which employs a HyperNetwork to generate channel-specific weights for the final prediction layer, striking a balance between channel-independent (CI) and channel-dependent (CD) modeling. As a plug-and-play module, it improves forecasting accuracy of various backbone models including DLinear, PatchTST, and TSMixer without incurring additional inference overhead.
Scaling and Transferability of Annealing Strategies in Large Language Model Training: This paper proposes a model-agnostic predictive framework that decomposes training loss into a forward-effect term (learning rate integral $S$), an annealing momentum term (Adam-style momentum integral $M$), and a model-size term $N$. It demonstrates that annealing strategies can be transferred from small models/small batches to large models/large batches, achieving a prediction MAPE below 2%.
Connectivity-Guided Sparsification of 2-FWL GNNs Preserving Full Expressivity: Co-Sparsify proposes a connectivity-aware sparsification framework that restricts 3-node interactions to biconnected components and 2-node interactions to connected components, eliminating provably redundant computation. It preserves full 2-FWL expressivity while substantially improving efficiency, achieving state-of-the-art results on synthetic substructure counting tasks and benchmarks including ZINC and QM9.
Resource Efficient Sleep Staging via Multi-Level Masking and Prompt Learning: This paper proposes MASS (Mask-Aware Sleep Staging), a framework that achieves reliable sleep staging using only 10% of the original EEG signal through a multi-level masking strategy and hierarchical prompt learning mechanism, providing a practical solution for resource-constrained wearable sleep monitoring systems.

📚 Pretraining (9)¶

Rectified Noise: A Generative Model Using Positive-incentive Noise: This paper proposes Rectified Noise (ΔRN), which leverages the positive-incentive noise (π-noise) framework to learn a set of beneficial noise signals and inject them into the velocity field of a pretrained Rectified Flow model, achieving a reduction in FID from 10.16 to 9.05 on ImageNet-1k with only 0.39% additional parameters.
GranAlign: Granularity-Aware Alignment Framework for Zero-Shot Video Moment Retrieval: This paper proposes GranAlign, a training-free granularity-aware alignment framework that addresses the core challenge of semantic granularity mismatch in zero-shot video moment retrieval (ZVMR). By rewriting queries into simplified and detailed variants and matching them against query-agnostic and query-aware video descriptions respectively, GranAlign achieves a 3.23% improvement in mAP@avg on QVHighlights.
ELSPR: Evaluator LLM Training Data Self-Purification on Non-Transitive Preferences: ELSPR models pairwise preferences of LLM evaluators as tournament graphs, identifies non-transitive preferences via strongly connected components (SCCs), proposes a normalized directed graph structural entropy metric, and filters problematic training data through graph reconstruction — resulting in a 13.8% reduction in non-transitivity and a 0.088 decrease in structural entropy, while the discarded data achieves only 34.4% human agreement (vs. 52.6% for retained data).
Perspective from a Broader Context: Can Room Style Knowledge Help Visual Floorplan Localization?: This paper proposes leveraging room style knowledge — obtained via unsupervised clustering pretraining in the form of a room discriminator — to resolve ambiguities caused by repetitive structures in visual floorplan localization (FLoc), achieving state-of-the-art performance on two standard benchmarks: Gibson and Structured3D.
PrefixGPT: Prefix Adder Optimization by a Generative Pre-trained Transformer: PrefixGPT frames prefix adder optimization as a sequence generation problem. A customized GPT model is pretrained to learn design rules, then fine-tuned via RL to generate optimized designs, achieving state-of-the-art area-delay product (ADP) with robustness to initialization.
TRACE: A Generalizable Drift Detector for Streaming Data-Driven Optimization: This paper proposes TRACE, a transferable concept drift detector based on attention-based sequence learning. By tokenizing statistical features and employing a dual-attention encoder, TRACE learns drift patterns that generalize across tasks, enabling deployment on unseen datasets and integration as a plug-and-play module into streaming data-driven optimization algorithms.
No-Regret Strategy Solving in Imperfect-Information Games via Pre-Trained Embedding: This paper proposes the Embedding CFR algorithm, which maps information sets in imperfect-information games to a continuous low-dimensional embedding space (rather than discrete clusters), achieving faster exploitability convergence and higher-quality strategy solving under the same space budget.
Beyond Cosine Similarity: Magnitude-Aware CLIP for No-Reference Image Quality Assessment: This paper proposes MA-CLIP, which discovers and exploits the magnitude information of CLIP image features as a complementary perceptual quality cue. Combined with cosine similarity, it achieves training-free adaptive dual-cue fusion for image quality assessment.
Learning Procedural-aware Video Representations through State-Grounded Hierarchy Unfolding: This paper proposes a Task-Step-State (TSS) three-level semantic framework that introduces "state" as a visual grounding layer within the conventional task-step hierarchy, and designs a progressive pretraining strategy following a U-shaped path (Task→Step→State→Step→Task) to unfold the TSS hierarchy stage by stage. The approach achieves comprehensive state-of-the-art performance on task recognition, step recognition, and step forecasting tasks on the COIN and CrossTask datasets.

✏️ Knowledge Editing (4)¶

Is the Information Bottleneck Robust Enough? Towards Label-Noise Resistant Information Bottleneck Learning: This paper identifies the inherent vulnerability of the Information Bottleneck (IB) principle under label noise and proposes LaT-IB, which decomposes representations into a clean-label subspace and a noisy-label subspace. Combined with a Minimal-Sufficient-Clean (MSC) criterion and a three-stage training framework, LaT-IB significantly outperforms existing IB methods across diverse noise conditions.
Hybrid-DMKG: A Hybrid Reasoning Framework over Dynamic Multimodal Knowledge Graphs for Multimodal Multihop QA with Knowledge Editing: This paper proposes the MMQAKE benchmark and the Hybrid-DMKG framework, which constructs a dual-channel hybrid reasoning mechanism — combining relation link prediction with RAG-augmented LVLM inference — over a dynamic multimodal knowledge graph, supplemented by a background reflection decision module. The approach significantly outperforms existing methods on 2–5 hop multimodal knowledge editing QA (H-Acc of 29.90% on LLaVA, surpassing IKE by 13.52 percentage points).
Model Editing as a Double-Edged Sword: Steering Agent Ethical Behavior: This paper frames the steering of agent ethical behavior as a model editing task (Behavior Editing), proposes a three-tier BehaviorBench grounded in psychological moral theory, and validates on 9 open-source and 20 closed-source models that model editing can precisely steer agents toward either benevolent or malicious behavior, with a single edit potentially causing global moral alignment drift.
Multiplicative Orthogonal Sequential Editing for Language Models (MOSE): This paper proposes MOSE (Multiplicative Orthogonal Sequential Editing), which injects new knowledge by left-multiplying the parameter matrix with an orthogonal matrix (rather than via additive updates), strictly preserving the Frobenius norm and condition number of the edited matrix. MOSE achieves a 12.08% performance improvement in sequential editing while retaining 95.73% of general capabilities.

💬 LLM (Other) (29)¶

Control Illusion: The Failure of Instruction Hierarchies in Large Language Models: This paper systematically demonstrates that the system/user prompt separation mechanism in current LLMs fails to establish reliable instruction priority, and finds that social hierarchy priors acquired during pretraining (authority, expertise, consensus) exert stronger control over model behavior than explicit system/user role markers.
PERSIST: Persistent Instability in LLM's Personality Measurements: The PERSIST framework systematically evaluates personality measurement stability across 29 LLMs (1B–685B) on over 2 million responses, revealing a "reasoning paradox" in which CoT reasoning increases variability while reducing perplexity, as well as a scale-dependent effect whereby conversational history exerts opposite influences on large versus small models—collectively indicating that current LLMs lack the architectural foundation for behavioral consistency.
ICL-Router: In-Context Learned Model Representations for LLM Routing: This paper proposes ICL-Router, a two-stage training framework (query reconstruction + ICL model routing) that encodes LLM capability profiles as in-context vectors, enabling scalable dynamic model routing. New models can be incorporated without retraining the router, achieving state-of-the-art performance on both in-distribution and out-of-distribution tasks.
Vision Transformers are Circulant Attention Learners: This paper discovers that self-attention matrices in ViTs inherently learn Block Circulant with Circulant Blocks (BCCB) patterns, and proposes Circulant Attention, which achieves $O(N\log N)$ complexity via 2D FFT, yielding consistent improvements on ImageNet classification, COCO detection, and ADE20K segmentation.
TransMamba: A Sequence-Level Hybrid Transformer-Mamba Language Model: This paper proposes TransMamba, a sequence-level Transformer-Mamba hybrid architecture that dynamically switches between Attention and SSM computation at different token positions via shared QKV/CBx parameters and a Memory Converter, achieving efficiency advantages for both short and long sequences.
LoopLLM: Transferable Energy-Latency Attacks in LLMs via Repetitive Generation: This paper proposes LoopLLM, a framework that launches energy-latency attacks by inducing LLMs into repetitive generation modes. Through repetition-inducing prompt optimization and token-aligned ensemble optimization, LoopLLM achieves over 90% of maximum output length across 12 open-source and 2 commercial LLMs, with approximately 40% improvement in cross-model transferability.
ParetoHqD: Fast Offline Multiobjective Alignment of Large Language Models Using Pareto High-Quality Data: This paper proposes ParetoHqD, which represents human preferences as preference directions in objective space (rather than linear scalarization), and performs two-stage SFT on high-quality data selected near the Pareto front. Using only 42% of the GPU time, it achieves multi-objective LLM alignment performance superior to five baselines.
CoEvo: Continual Evolution of Symbolic Solutions Using Large Language Models: This paper proposes CoEvo, a framework that integrates LLMs with evolutionary search methodology to achieve continual open-ended evolution of symbolic solutions through a dynamic knowledge library and multi-representation spaces (natural language / mathematical formulas / code), significantly outperforming existing symbolic regression methods on the AI Feynman benchmark.
Whispering Agents: An Event-Driven Covert Communication Protocol for the Internet of Agents: This paper presents the first formal definition of a "Covert Event Channel" in the Internet of Agents (IoA) and proposes the ΠCCAP protocol, which embeds secret data across the storage, timing, and behavioral dimensions of agent conversations, achieving high-capacity, high-robustness covert communication that is imperceptible to LLM-based censors.
ProFuser: Progressive Fusion of Large Language Models: ProFuser is proposed to comprehensively identify the strengths of each source model across different dimensions via dual-mode advantage assessment (training-mode Min-CE + inference-mode Reward Model voting), and then integrates the complementary capabilities of heterogeneous LLMs into a single target model through a progressive fusion strategy (inference mode first → training mode second, as an easy-to-hard curriculum), achieving an average improvement of 1.65% across 6 benchmarks covering knowledge, reasoning, and safety.

Browse all 29 LLM (Other) papers →

📖 NLP Understanding (1)¶

Language Models and Logic Programs for Trustworthy Tax Reasoning: This paper reframes tax law reasoning as a semantic parsing task, where LLMs translate statutory text and case facts into Prolog logic programs that are subsequently executed by a symbolic solver. By combining gold-standard statute translations, retrieval-augmented case examples, and self-consistency checks, the system achieves 86/100 accuracy on the SARA dataset while reducing estimated deployment cost to $15.78 per person — less than 6% of the average U.S. tax filing cost.

✍️ Text Generation (3)¶

AutoMalDesc: Large-Scale Script Analysis for Cyber Threat Research: This paper proposes AutoMalDesc, an automated static analysis framework that employs an iterative self-paced learning pipeline — starting from 900 expert-annotated seed samples, fine-tuning Llama-3.3-70B via LoRA to generate pseudo-labels, applying multi-stage quality filtering to obtain 101K samples, and training a V2 model — to achieve automated malware classification and behavior description across five scripting languages, improving Batch script detection accuracy from 52.7% to 82.4%.
C3TG: Conflict-aware, Composite, and Collaborative Controlled Text Generation: This paper proposes the C3TG framework, which achieves fine-grained multi-attribute controllable text generation through a two-stage approach: in the generation stage, weighted KL divergence is used to fuse attribute distributions and adjust token probabilities; in the optimization stage, an energy function (combining classifier scores and conflict penalty terms) drives iterative rewriting via a Feedback Agent. C3TG achieves 90.4% attribute accuracy across 17 attribute subcategories while substantially reducing toxicity.
Structured Language Generation Model: Loss Calibration and Formatted Decoding for Efficient Text: This paper proposes the SLGM framework, which reformulates structured prediction tasks for generative language models as classification problems via three components: structured input format, format loss, and format-aware decoding. Without introducing additional model parameters, SLGM significantly improves structural prediction performance of sub-1B models across 13 datasets spanning 5 task categories, including NER, RE, and SRL.

🗣️ Dialogue Systems (5)¶

Canoe: Teaching LLMs to Maintain Contextual Faithfulness via Synthetic Tasks and RL: This paper proposes the Canoe framework, which synthesizes four types of verifiable short-form QA data from Wikidata triples and applies Dual-GRPO (incorporating accuracy reward, long-form proxy reward, and format reward) to jointly optimize faithfulness in both short- and long-form generation. The approach improves Llama-3-8B by an average of 22.6% across 11 downstream tasks, surpassing GPT-4o.
TalkSketch: Multimodal Generative AI for Real-time Sketch Ideation with Speech: This paper proposes TalkSketch, a system that integrates hand-drawn sketches with real-time speech input into a multimodal AI chatbot, enabling designers to simultaneously draw and verbalize ideas during early-stage ideation. The system addresses the problem that text-based prompting in existing GenAI tools disrupts the creative workflow.
Auto-PRE: An Automatic and Cost-Efficient Peer-Review Framework for Language Generation Evaluation: This paper proposes the Auto-PRE framework, which selects qualified LLM evaluators through an automatic qualification exam across three dimensions—consistency, pertinence, and self-confidence—achieving state-of-the-art evaluation performance without human annotation while significantly reducing costs.
Chatsparent: An Interactive System for Detecting and Mitigating Cognitive Fatigue in LLMs: This paper presents Chatsparent, an interactive system that monitors three token-level fatigue signals during LLM inference in real time—attention decay, embedding drift, and entropy collapse—aggregates them into a unified fatigue index, and automatically applies lightweight interventions (prompt re-injection, attention reset, entropy-regularized decoding, self-reflection checkpoints) when fatigue thresholds are triggered, transforming passive chat interaction into an active diagnostic experience.
Emergent Persuasion: Will LLMs Persuade Without Being Prompted?: This paper investigates whether LLMs spontaneously exhibit persuasive behavior without being explicitly prompted to do so. It finds that activation steering fails to reliably induce persuasive tendencies, whereas SFT fine-tuning on benign persuasion data causes models to exhibit emergent persuasive behavior on harmful topics, revealing latent post-training safety risks.

🌐 Multilingual & Translation (9)¶

Mitigating Content Effects on Reasoning in Language Models through Fine-Grained Activation Steering: This paper applies activation steering to mitigate content effects in LLMs — the tendency to conflate content believability with formal logical validity. The proposed K-CAST (kNN-based Conditional Activation Steering) method achieves up to 15% improvement in formal reasoning accuracy on models unresponsive to standard static steering.
How Does Alignment Enhance LLMs' Multilingual Capabilities? A Language Neurons Perspective: This paper proposes a ternary neuron classification scheme (language-specific / language-related / universal) and decomposes multilingual LLM inference into a four-stage framework. It finds that multilingual alignment improves performance by increasing language-related neurons (while reducing language-specific ones), and further demonstrates a "spontaneous multilingual alignment" effect on untrained languages.
X-MuTeST: A Multilingual Benchmark for Explainable Hate Speech Detection and A Novel LLM-consulted Explanation Framework: This paper proposes the X-MuTeST framework, which combines LLM semantic reasoning with a two-stage training strategy enhanced by n-gram attention for explainable multilingual hate speech detection. It also introduces the first token-level human-annotated rationale benchmark datasets for Hindi and Telugu.
Focusing on Language: Revealing and Exploiting Language Attention Heads in Multilingual Large Language Models: This paper proposes LAHIS, a method that efficiently identifies language-specific and language-general attention heads in multilingual LLMs using only a single forward-backward pass. It demonstrates that manipulating these heads enables cross-lingual attention transfer, mitigates off-target language generation, and improves multilingual QA performance with only 14–20 trainable parameters.
GloCTM: Cross-Lingual Topic Modeling via a Global Context Space: This paper proposes GloCTM, a dual-path VAE architecture (local language path + global context path) that enforces cross-lingual alignment at four levels—Polyglot Augmentation (cross-lingual neighbor-based input expansion), KL divergence internal alignment, unified decoder structural alignment, and CKA semantic alignment—achieving state-of-the-art topic quality and cross-lingual alignment on three cross-lingual datasets.
ViDia2Std: A Parallel Corpus and Methods for Low-Resource Vietnamese Dialect-to-Standard Translation: ViDia2Std constructs the first manually annotated Vietnamese dialect-to-standard parallel corpus covering all 63 provinces of Vietnam (13,000+ sentence pairs), evaluates multiple seq2seq models on the dialect normalization task, and demonstrates that dialect normalization as a preprocessing step significantly improves downstream task performance in machine translation and sentiment analysis.
Bridging the Multilingual Safety Divide: Efficient, Culturally-Aware Alignment for Global South Languages: This paper synthesizes multiple empirical studies to reveal critical failures of LLM safety mechanisms in low-resource and code-mixed settings, and proposes a resource-aware blueprint grounded in parameter-efficient safety steering, culturally driven preference data, and community-participatory alignment.
MIDB: Multilingual Instruction Data Booster for Enhancing Cultural Equality in Multilingual Instruction Synthesis: This paper proposes MIDB (Multilingual Instruction Data Booster), a unified model trained on 36.8k expert-annotated revision samples, which automatically repairs content errors, machine translation defects, and localization deficiencies in multilingual synthetic instruction data, significantly improving instruction data quality across 16 languages and enhancing downstream LLM multilingual/cultural understanding capabilities.
NADIR: Differential Attention Flow for Non-Autoregressive Transliteration in Indic Languages: This paper proposes NADIR, a non-autoregressive (NAR) multilingual transliteration architecture combining a differential Transformer with a Mixture-of-Experts (MoE) module. NADIR achieves over 13× inference speedup on Indic language transliteration tasks while substantially reducing hallucination errors common in NAR models (repetition, substitution, omission, and insertion), narrowing the accuracy gap with autoregressive counterparts.

🔍 Information Retrieval & RAG (21)¶

ComoRAG: A Cognitive-Inspired Memory-Organized RAG for Stateful Long Narrative Reasoning: Inspired by the metacognitive regulation mechanism of the prefrontal cortex, this paper proposes the ComoRAG framework, which achieves stateful multi-step reasoning via a dynamic memory workspace and iterative probe queries, significantly outperforming existing RAG methods on long narrative understanding tasks (200K+ tokens).
ConvMix: A Mixed-Criteria Data Augmentation Framework for Conversational Dense Retrieval: This paper proposes ConvMix, a mixed-criteria data augmentation framework that leverages LLMs to perform scalable relevance annotation augmentation from both query and document directions, combined with clustering-based diversity selection and Fisher information-based in-distribution supervision, to systematically improve conversational dense retrieval performance.
Mem-PAL: Towards Memory-based Personalized Dialogue Assistants for Long-term User-Agent Interaction: This paper proposes H2Memory, a four-layer hierarchical heterogeneous memory structure (Log Graphs / Background Memory / Topic Outlines / Principles), validated on the PAL-Set dataset (100 users × 8.4 months of interaction), improving BLEU-1 on demand paraphrasing and solution recommendation tasks from 13.59 to 26.67.
Cog-RAG: Cognitive-Inspired Dual-Hypergraph with Theme Alignment Retrieval-Augmented Generation: This paper proposes Cog-RAG, which constructs a dual-hypergraph index comprising a theme hypergraph and an entity hypergraph to simulate the human "top-down" cognitive process via a two-stage retrieval strategy (theme first, then details), achieving global-to-local semantic alignment for generation.
OPERA: A Reinforcement Learning--Enhanced Orchestrated Planner-Executor Architecture for Reasoning-Oriented Multi-Hop Retrieval: This paper proposes OPERA, a hierarchical framework comprising a Goal Planning Module and a Reason-Execute Module, combined with MAPGRPO—a training algorithm specifically designed for multi-agent settings—to substantially improve performance on reasoning-oriented multi-hop retrieval tasks.
PRECISE: Reducing the Bias of LLM Evaluations Using Prediction-Powered Ranking Estimation: This paper extends the Prediction-Powered Inference (PPI) framework to sub-instance-level ranking metrics (e.g., Precision@K), achieving unbiased ranking metric estimation using only 30–100 human annotations combined with large-scale LLM judgments. The computational complexity is reduced from $O(2^{|C|})$ to $O(2^K)$, and the approach has been successfully deployed to guide an LLM-based query rewriting system in an Indian e-commerce search setting.
SR-KI: Scalable and Real-Time Knowledge Integration into LLMs via Supervised Attention: This paper proposes SR-KI, a framework that injects structured knowledge bases into LLM KV caches via a two-stage training procedure (retrieval layer localization + attention supervision loss). On a single A100 40GB GPU, SR-KI supports injection of up to 40K knowledge base entries, achieves a compression ratio of up to 99.75% via top-100 selection, and maintains an average Recall@10 above 88%.
Do Retrieval Augmented Language Models Know When They Don't Know?: This paper systematically analyzes the refusal calibration problem in RAG models, finding that RALMs exhibit an over-refusal rate exceeding 55% when all retrieved documents are irrelevant (even when the model's internal knowledge suffices to answer), and proposes a mechanism combining uncertainty estimation with refusal-aware fine-tuning to balance refusal behavior and answer quality.
Magnitude Matters: A Superior Class of Similarity Metrics for Holistic Semantic Understanding: This paper proposes two parameter-free, magnitude-aware vector similarity metrics—Overlap Similarity (OS) and Hyperbolic Tangent Similarity (HTS)—that achieve significantly lower MSE than Cosine Similarity and Dot Product on classification tasks (paraphrase detection, natural language inference) across 4 sentence embedding models and 8 NLP benchmarks, without any additional training overhead.
PRIME: Planning and Retrieval-Integrated Memory for Enhanced Reasoning: Inspired by dual-process cognitive theory, PRIME is a multi-agent reasoning framework in which a Quick Thinking Agent (System 1) rapidly generates intuitive answers, a Reflection Agent evaluates their confidence, and—when uncertainty is detected—six specialized System 2 agents (Planning / Search / Reading / Hypothesis / Integration / Decision) are triggered for deep knowledge-retrieval reasoning. The framework enables open-source LLaMA 3 to approach GPT-4o performance on medical and multi-hop QA benchmarks.

Browse all 21 Information Retrieval & RAG papers →

💻 Code Intelligence (10)¶

ReCode: Updating Code API Knowledge with Reinforcement Learning: This paper proposes ReCode, a framework that trains LLMs via rule-based reinforcement learning (rather than SFT) to correctly leverage API update documentation provided in the prompt for code version migration, enabling a 7B model to surpass 32B models on CodeUpdateArena.
Unintended Misalignment from Agentic Fine-Tuning: Risks and Mitigation: This paper demonstrates that fine-tuning LLMs on benign agentic data causes unintended safety misalignment (attack success rate increases by 32–38%), and proposes PING (Prefix Injection Guard)—an iterative generate-and-evaluate approach that automatically discovers natural language prefixes to guide fine-tuned agents toward refusing harmful requests, achieving an average refusal rate improvement of 66% (Web) and 44% (Code) while preserving task performance (degradation of only 1.8%).
TAPA: Training-Free Adaptation of Programmatic Agents via LLM-Guided Program Synthesis in Dynamic Environments: TAPA positions LLMs as "intelligent modulators" of the symbolic action space rather than direct decision-makers. Through LLM-guided program synthesis, it dynamically adapts the symbolic actions of programmatic agents without retraining, achieving strong performance in cybersecurity DDoS defense (77.7% network uptime) and swarm intelligence formation control.
Why Do Open-Source LLMs Struggle with Data Analysis? A Systematic Empirical Study: This paper systematically investigates the capability bottlenecks of open-source LLMs in data analysis tasks. It decomposes data analysis into three dimensions—data comprehension, code generation, and strategic planning—and identifies strategic planning as the decisive factor, rather than coding or data comprehension. A strategy-guided data synthesis approach is proposed, enabling fine-tuned 7B/14B models to achieve performance competitive with GPT-4o.
Extracting Events Like Code: A Multi-Agent Programming Framework for Zero-Shot Event Extraction: This paper proposes Agent-Event-Coder (AEC), which reformulates zero-shot event extraction as a software engineering workflow. Four specialized agents (Retrieval→Planning→Coding→Verification) collaborate to perform extraction, while event schemas are encoded as executable Python classes to enable compiler-style deterministic validation and dual-loop iterative correction. AEC comprehensively outperforms zero-shot baselines across 5 domains and 6 LLMs.
MoSE: Hierarchical Self-Distillation Enhances Early Layer Embeddings: This paper proposes ModularStarEncoder (MoSE), a 1B-parameter multi-exit encoder that significantly enhances early-layer representations via a novel self-distillation mechanism in which higher layers guide the training of lower layers. MoSE surpasses all open-source models on code understanding tasks such as CodeSearchNet while supporting flexible compute–accuracy tradeoff deployment.
DiffBench Meets DiffAgent: End-to-End LLM-Driven Diffusion Acceleration Code Generation: This paper proposes DiffBench (an evaluation benchmark comprising 604 diffusion model acceleration tasks across 5 difficulty levels) and DiffAgent (a closed-loop framework integrating Planning, Coding, and Debugging agents with a genetic algorithm-based selector). On Claude Sonnet 4, the framework improves the pass rate for diffusion acceleration code generation from 54.30% to 81.59%, achieving a 68.27% success rate on complex optimization tasks.
EquaCode: A Multi-Strategy Jailbreak Approach for Large Language Models via Equation Solving and Code Completion: This paper proposes EquaCode, a multi-strategy jailbreak method that decomposes malicious queries into a cross-domain combination of equation solving ($B+C+x=A$) and code completion (completing the solve() method of a Solver class), achieving an average attack success rate of 92.78% on the GPT series and approaching 100% on the latest models (Gemini/DeepSeek/Grok).
SPAN: Benchmarking and Improving Cross-Calendar Temporal Reasoning of Large Language Models: This paper proposes SPAN, a cross-calendar temporal reasoning benchmark (6 calendars × 10 reasoning directions × 100-year range × 37,380 instances). Baseline LLMs achieve an average accuracy of only 34.5% (none exceeding 80%), revealing two systematic failure modes—Future-Date Degradation and Calendar Asymmetry Bias. A tool-augmented Time Agent achieves 95.31%, demonstrating that cross-calendar reasoning requires external tools rather than parametric knowledge.
Towards Better Code Understanding in Decoder-Only Models with Contrastive Learning: This paper proposes CL4D, a contrastive learning framework that adapts pretrained decoder-only code generation models to code understanding tasks (code search, clone detection) via continued pretraining, achieving performance comparable to or better than encoder-only models of equivalent scale without retraining them from scratch.

🎨 Image Generation (79)¶

Margin-aware Preference Optimization for Aligning Diffusion Models without Reference: This paper proposes MaPO (Margin-aware Preference Optimization), a reference-free preference alignment method that aligns T2I diffusion models by directly optimizing the likelihood margin between preferred and dispreferred outputs under the Bradley-Terry model. MaPO outperforms DPO and task-specific methods across 5 domains, including style adaptation, safety generation, and general preference alignment.
LongLLaDA: Unlocking Long Context Capabilities in Diffusion LLMs: This paper presents the first systematic study of long-context capabilities in diffusion large language models (diffusion LLMs), revealing stable perplexity under direct extrapolation and a "local awareness" phenomenon. It further proposes LongLLaDA, a training-free method that successfully extends the context window by 6× (to 24k tokens) via NTK-based RoPE extrapolation.
MP1: MeanFlow Tames Policy Learning in 1-step for Robotic Manipulation: This work introduces the MeanFlow paradigm to the robot learning domain for the first time. By incorporating 3D point cloud inputs and a Dispersive Loss, MP1 generates action trajectories in a single network forward pass (1-NFE), achieving state-of-the-art success rates with an inference latency of only 6.8 ms on robotic manipulation tasks.
X2Edit: Revisiting Arbitrary-Instruction Image Editing through Self-Constructed Data and Task-Aware Representation Learning: A 3.7M high-quality editing dataset covering 14 task categories is constructed, and a lightweight (0.9B parameter) plug-and-play editing module based on Task-Aware MoE-LoRA and Contrastive Learning is proposed, achieving performance comparable to 12B fully fine-tuned models.
Mass Concept Erasure in Diffusion Models with Concept Hierarchy: This paper proposes a grouped erasure strategy based on supertype-subtype concept hierarchy and Supertype-Preserving LoRA (SuPLoRA). By freezing the down-projection matrix (orthogonal to the supertype subspace) and training only the up-projection matrix, the method achieves an optimal balance between erasure effectiveness and generation quality in large-scale, multi-domain concept erasure.
RelaCtrl: Relevance-Guided Efficient Control for Diffusion Transformers: This paper proposes the RelaCtrl framework, which quantifies the sensitivity of each DiT layer to control information via a ControlNet Relevance Score, and uses this analysis to guide the placement and modeling capacity of control blocks. A Two-Dimensional Shuffle Mixer (TDSM) is introduced to replace self-attention and FFN, achieving controllable generation quality superior to PixArt-δ with only 15% of its parameters and computational cost.
Multi-Aspect Cross-modal Quantization for Generative Recommendation: This paper proposes MACRec, which introduces multi-aspect cross-modal interaction at both the semantic ID learning stage and the generative model training stage. Through cross-modal quantization (contrastive learning-enhanced residual quantization) and multi-aspect alignment (implicit + explicit), MACRec significantly improves recommendation performance while reducing ID collision rates.
TruthfulRAG: Resolving Factual-level Conflicts in Retrieval-Augmented Generation with Knowledge Graphs: This paper proposes TruthfulRAG, a framework that, for the first time, leverages knowledge graphs (KGs) to resolve conflicts between retrieved knowledge and LLM parametric knowledge at the factual level in RAG systems. The framework improves generation accuracy and trustworthiness through triple extraction, query-aware graph retrieval, and an entropy-based conflict filtering mechanism.
T-LoRA: Single Image Diffusion Model Customization Without Overfitting: This paper proposes T-LoRA, a timestep-dependent low-rank adaptation framework that addresses overfitting in single-image diffusion model customization. The framework dynamically adjusts the effective LoRA rank across diffusion timesteps (smaller rank at high-noise timesteps, larger rank at low-noise timesteps) and employs orthogonal initialization (Ortho-LoRA) via random matrix SVD to ensure information independence among adaptation components, achieving an optimal balance between concept fidelity and text alignment.
T2I-RiskyPrompt: A Benchmark for Safety Evaluation, Attack, and Defense on Text-to-Image Model: This paper constructs T2I-RiskyPrompt — a comprehensive benchmark comprising 6,432 valid risky prompts spanning 6 major categories and 14 subcategories, each annotated with hierarchical labels and detailed risk rationales. A reason-driven MLLM-based risk detection method is proposed (achieving 91.8% accuracy with a 3B model), and a systematic evaluation is conducted across 8 T2I models, 9 defense methods, 5 safety filters, and 5 attack strategies.

Browse all 79 Image Generation papers →

🎬 Video Generation (11)¶

OmniVDiff: Omni Controllable Video Diffusion for Generation and Understanding: This paper proposes OmniVDiff, a unified controllable video diffusion framework that jointly models multiple visual modalities (RGB, depth, segmentation, Canny) in color space and introduces an Adaptive Modality Control Strategy (AMCS). Within a single diffusion model, OmniVDiff simultaneously supports three task types—text-conditioned generation, X-conditioned generation, and video understanding—achieving state-of-the-art performance on VBench.
GenVidBench: A 6-Million Benchmark for AI-Generated Video Detection: This paper introduces GenVidBench—the first large-scale AI-generated video detection dataset with 6.78 million videos, featuring cross-source and cross-generator properties, covering 11 state-of-the-art video generators, and providing rich semantic annotations.
Mask2IV: Interaction-Centric Video Generation via Mask Trajectories: This paper proposes Mask2IV, a two-stage decoupled framework that first predicts mask motion trajectories of the interactor and object, then generates video conditioned on these trajectories. The approach enables controllable, interaction-centric video generation without dense mask annotations, supporting both human-object interaction and robot manipulation scenarios.
FilmWeaver: Weaving Consistent Multi-Shot Videos with Cache-Guided Autoregressive Diffusion: FilmWeaver is proposed as a framework that guides autoregressive diffusion models via a dual-level cache (Shot Cache + Temporal Cache), enabling multi-shot video generation of arbitrary length with cross-shot consistency.
DreamRunner: Fine-Grained Compositional Story-to-Video Generation with Retrieval-Augmented Motion Adaptation: This paper proposes DreamRunner, a framework that achieves fine-grained controllable multi-character multi-event story video generation via LLM-based dual-level planning, retrieval-augmented motion prior learning, and a spatial-temporal region-based 3D attention injection module (SR3AI).
Phased One-Step Adversarial Equilibrium for Video Diffusion Models: This paper proposes V-PAE (Video Phased Adversarial Equilibrium), a two-phase distillation framework consisting of stability priming followed by unified adversarial equilibrium, which compresses large-scale video diffusion models (e.g., Wan2.1-I2V-14B) to single-step generation, achieving a 100× speedup and surpassing existing acceleration methods by 5.8% in average quality on VBench-I2V.
MoFu: Scale-Aware Modulation and Fourier Fusion for Multi-Subject Video Generation: This paper proposes MoFu, which addresses two fundamental challenges in multi-subject video generation—scale inconsistency and permutation sensitivity—through two core modules: Scale-Aware Modulation (SMO, an LLM-guided scale-aware modulation mechanism) and Fourier Fusion (an FFT-based permutation-invariant feature fusion strategy). The work additionally introduces the MoFu-1M training dataset and the MoFu-Bench evaluation benchmark.
Seeing the Unseen: Zooming in the Dark with Event Cameras: This paper proposes RetinexEVSR, the first event-driven low-light video super-resolution (LVSR) framework. Through a Retinex-inspired bidirectional fusion strategy (RBF)—which first uses illumination maps to guide event feature denoising (IEE), then leverages enhanced event features to recover reflectance details (ERE)—the method achieves a 2.95 dB gain on the SDSD benchmark while reducing runtime by 65%.
3D4D: An Interactive Editable 4D World Model via 3D Video Generation: This paper proposes 3D4D, an interactive 4D visualization framework integrating WebGL and Supersplat rendering. A four-module backend pipeline converts static images and text prompts into editable 4D scenes, while a VLM-guided foveated rendering strategy enables 60fps real-time interaction, achieving state-of-the-art performance on both CLIP Consistency and CLIP Score.
SphereDiff: Tuning-free Omnidirectional Panoramic Image and Video Generation via Spherical Latent Representation: This paper proposes SphereDiff, which defines a spherical latent representation (uniformly distributed via Fibonacci Lattice) to replace conventional equirectangular projection (ERP), combined with a dynamic sampling algorithm and distortion-aware weighted averaging. Without any fine-tuning, SphereDiff leverages pretrained diffusion models such as SANA and LTX Video to generate seamless, low-distortion 360° panoramic images and videos.

Browse all 11 Video Generation papers →

🧩 Multimodal VLM (75)¶

SpeakerLM: End-to-End Versatile Speaker Diarization and Recognition with Multimodal Large Language Models: SpeakerLM is the first multimodal large language model designed specifically for end-to-end Speaker Diarization and Recognition (SDR). Through an audio encoder–projector–LLM architecture and a flexible speaker enrollment mechanism, it significantly outperforms cascaded baseline systems on multiple public benchmarks (absolute cpCER reduction up to 13.82%) and demonstrates strong robustness on out-of-domain test sets.
Exo2Ego: Exocentric Knowledge Guided MLLM for Egocentric Video Understanding: This paper proposes Exo2Ego, a framework that learns a mapping between the exocentric (third-person) and egocentric (first-person) domains to transfer rich exocentric knowledge encoded in MLLMs to egocentric video understanding. Combined with a newly constructed dataset of 1.1M synchronized ego-exo clip-text pairs (Ego-ExoClip) and 600K instruction-tuning samples (EgoIT), Exo2Ego achieves state-of-the-art open-source performance across 8 egocentric video benchmarks.
Cross-modal Proxy Evolving for OOD Detection with Vision-Language Models: This paper proposes CoEvo, a training-free and annotation-free test-time framework that dynamically updates positive and negative proxy caches through a bidirectional sample-conditioned text/visual proxy co-evolution mechanism. On ImageNet-1K, CoEvo improves AUROC by 1.33% and reduces FPR95 by 45.98% (from 18.92% to 10.22%) over the strongest negative-label baseline, achieving state-of-the-art zero-shot OOD detection.
Explore How to Inject Beneficial Noise in MLLMs: This paper proposes the Multimodal Noise Generator (MuNG), which dynamically generates "beneficial noise" from image-text pairs via a variational inference framework and injects it into the frozen visual features of an MLLM. The approach suppresses task-irrelevant semantics and enhances cross-modal representation alignment, requiring only ~1% additional parameters while outperforming full fine-tuning and PEFT methods such as LoRA.
To Align or Not to Align: Strategic Multimodal Representation Alignment for Optimal Performance: By introducing a controllable contrastive learning module to systematically regulate alignment strength $\lambda$, and employing the Partial Information Decomposition (PID) framework to quantify the redundancy–uniqueness–synergy structure between modalities, this work reveals that the utility of explicit alignment is highly data-dependent: alignment is beneficial when redundancy dominates, harmful when uniqueness dominates, and an optimal $\lambda^*$ exists in mixed scenarios.
Multimodal DeepResearcher: Generating Text-Chart Interleaved Reports From Scratch with Agentic Framework: This paper proposes Multimodal DeepResearcher, a four-stage agentic framework for generating text-chart interleaved research reports from scratch. It introduces Formal Description of Visualization (FDV) to enable LLMs to learn and produce diverse charts, and employs an Actor-Critic iterative refinement mechanism (LLM generates D3.js code → browser rendering → multimodal LLM review). The system achieves an 82% overall win rate (Claude 3.7) on the newly constructed MultimodalReportBench and a 100% win rate in human evaluation.
VipAct: Visual-Perception Enhancement via Specialized VLM Agent Collaboration and Tool-use: VipAct proposes a multi-agent collaboration framework that significantly improves VLM performance on fine-grained visual perception tasks through three-tier collaboration: an Orchestrator Agent (task analysis, planning, and coordination), specialized agents (captioning, comparison, and visual prompt interpretation), and vision expert models (depth estimation, object detection, segmentation, etc.). The framework improves accuracy on Blink from 63.74% (zero-shot GPT-4o) to 73.79%.
ClearAIR: A Human-Visual-Perception-Inspired All-in-One Image Restoration: Inspired by human visual perception (HVP), this paper proposes ClearAIR, a coarse-to-fine unified image restoration framework that progressively recovers image quality through four stages — MLLM-based quality assessment → semantic region perception → degradation type identification → internal clue reuse — achieving state-of-the-art performance across multiple degradation tasks.
Bridging the Copyright Gap: Do Large Vision-Language Models Recognize and Respect Copyrighted Content?: This paper presents the first systematic evaluation of LVLMs' ability to recognize and respect copyrighted content in multimodal contexts. It constructs a large-scale benchmark of 50,000 multimodal query–content pairs, finds that 11 out of 12 SOTA LVLMs fail to refuse infringing requests even when explicit copyright notices are present, and proposes CopyGuard—a tool-augmented framework that raises the infringement rejection rate from ~3% to ~62%.
MCMoE: Completing Missing Modalities with Mixture of Experts for Incomplete Multimodal Action Quality Assessment: This paper presents the first exploration of incomplete multimodal action quality assessment (AQA), proposing the MCMoE framework. An Adaptive Gated Modality Generator (AGMG) completes missing modalities, while a Mixture of Experts (MoE) module with soft routing dynamically fuses unimodal and cross-modal joint representations within a unified single-stage training paradigm. MCMoE achieves state-of-the-art performance under both complete and incomplete modality settings across three public AQA benchmarks, with only 4.90M parameters.

Browse all 75 Multimodal VLM papers →

🧠 VLM Reasoning (10)¶

AStar: Boosting Multimodal Reasoning with Automated Structured Thinking: This paper proposes AStar, a training-free multimodal reasoning paradigm that constructs a library of high-level "thought card" reasoning templates from 500 seed samples. At inference time, the most suitable templates are adaptively retrieved to guide structured reasoning in MLLMs. A 7B model achieves 53.9% accuracy on MathVerse (surpassing GPT-4o's 50.2%), requiring only 50 minutes of preprocessing and no model training.
SToLa: Self-Adaptive Touch-Language Framework with Tactile Commonsense Reasoning in Open-Ended Scenarios: SToLa proposes the first Mixture-of-Experts (MoE)-based touch-language framework, which employs a dynamic routing mechanism to manage the modality gap between tactile and linguistic inputs. The work also introduces TactileBench, an open-ended tactile commonsense reasoning dataset covering 8 physical properties and 4 interaction characteristics. With only 7B parameters, SToLa achieves state-of-the-art performance on the PhysiCLeAR benchmark, surpassing the 13B Octopi model.
AbductiveMLLM: Boosting Visual Abductive Reasoning Within MLLMs: Inspired by the dual-mode human cognitive process of verbal abduction and pictorial imagination, this paper proposes AbductiveMLLM, which enhances visual abductive reasoning in MLLMs via two collaborative components — a Reasoner (causal contrastive learning for hypothesis selection) and an Imaginer (diffusion-model-based pictorial reasoning) — achieving state-of-the-art performance on the VAR and YouCookII benchmarks.
CrossVid: A Comprehensive Benchmark for Evaluating Cross-Video Reasoning in Multimodal Large Language Models: This paper introduces CrossVid, the first comprehensive benchmark for systematically evaluating the Cross-Video Reasoning (CVR) capabilities of multimodal large language models (MLLMs). CrossVid encompasses 10 tasks across 4 dimensions, 5,331 videos, and 9,015 QA pairs. Experiments reveal that the current best-performing model, Gemini-2.5-Pro, achieves only 50.4% accuracy, far below the human performance of 89.2%.
FinMMDocR: Benchmarking Financial Multimodal Reasoning with Scenario Awareness, Document Understanding, and Multi-Step Computation: This paper introduces FinMMDocR, a bilingual multimodal reasoning benchmark targeting real-world financial scenarios. It comprises 1,200 expert-annotated numerical reasoning questions spanning 12 implicit financial scenario types, 9 categories of long documents (averaging 50.8 pages), and reasoning chains averaging 11 steps. The strongest MLLM (o4-mini-high) achieves only 58% accuracy, exposing critical deficiencies of existing models in complex financial reasoning.
Graph-of-Mark: Promote Spatial Reasoning in Multimodal Language Models with Graph-Based Visual Prompting: This paper proposes Graph-of-Mark (GoM), a training-free pixel-level visual prompting method that explicitly encodes inter-object spatial relationships by overlaying a depth-aware scene graph (comprising nodes and directed edges) directly onto input images, achieving up to an 11 percentage point improvement in zero-shot spatial reasoning accuracy for multimodal language models on VQA and grounding tasks.
Leveraging Textual Compositional Reasoning for Robust Change Captioning: This paper proposes CORTEX, a framework that introduces VLM-generated compositional reasoning text as explicit cues, combined with an Image-Text Dual Alignment (ITDA) module, to enhance purely visual change captioning methods in understanding structured semantics such as object relationships and spatial configurations.
Yes FLoReNce, I Will Do Better Next Time! Agentic Feedback Reasoning for Humorous Meme Detection: This paper proposes FLoReNce, a framework that models humorous meme understanding as a closed-loop control system. Through a feedback loop consisting of a Judge agent, a PID controller, and a non-parametric knowledge base, the system retrieves similar past experiences at inference time to modulate prompts, enabling a frozen VLM to perform adaptive reasoning without fine-tuning, substantially improving both prediction accuracy and explanation quality.
Concept-RuleNet: Grounded Multi-Agent Neurosymbolic Reasoning in Vision Language Models: This paper proposes Concept-RuleNet, a three-agent collaborative neurosymbolic reasoning framework that conditions symbol generation and rule construction on visual concepts extracted from training images. It addresses the symbol hallucination and non-representativeness issues of existing methods (e.g., Symbol-LLM) that rely solely on class labels, achieving an average accuracy improvement of ~5% across 5 OOD benchmarks and reducing hallucinated symbols by up to 50%.
Tri-Bench: Stress-Testing VLM Reliability on Spatial Reasoning under Camera Tilt and Object Interference: Tri-Bench is a compact benchmark comprising 400 real-world photographs of triangles. By systematically controlling two factors — camera pose (planar vs. tilted) and object interference — it evaluates the spatial geometric reasoning capabilities of four leading VLMs. The results reveal that models default to 2D image-plane cues rather than genuine 3D geometry, even when explicit reference-frame guardrails are provided in the prompt, with accuracy on minority-class shapes dropping to near 0%.

⚡ VLM Efficiency (5)¶

Filter, Correlate, Compress: Training-Free Token Reduction for MLLM Acceleration: This paper proposes FiCoCo, a three-stage framework (Filter–Correlate–Compress) that identifies redundant tokens via integrated vision-aware and semantic-aware redundancy metrics, adaptively recycles information from discarded tokens via inter-token correlation, and achieves training-free MLLM acceleration. On LLaVA-NeXT, FiCoCo achieves a 14.7× FLOPs reduction while retaining 93.6% of performance, and consistently outperforms FastV, SparseVLM, and other state-of-the-art methods across five MLLM architectures.
Global Compression Commander: Plug-and-Play Inference Acceleration for High-Resolution Large Vision-Language Models: This paper proposes GlobalCom², a plug-and-play, training-free token compression framework tailored for high-resolution VLMs with dynamic cropping architectures. It leverages the global thumbnail as a "commander" to guide differentiated compression across local crop regions, achieving >90% of original performance while compressing 90% of visual tokens.
TinyChemVL: Advancing Chemical Vision-Language Models via Efficient Visual Token Reduction and Complex Reaction Tasks: TinyChemVL is a chemistry-domain VLM with only 4B parameters. It compresses visual tokens to 1/16 of the original count via an adaptive token merging and pruning strategy, introduces reaction-level tasks and the ChemRxn-V benchmark, and achieves state-of-the-art performance on both molecular- and reaction-level visual chemistry tasks while significantly improving inference and training speed.
EM-KD: Distilling Efficient Multimodal Large Language Model with Unbalanced Vision Tokens: This paper proposes EM-KD, a distillation framework that leverages the Hungarian algorithm to address the vision token count imbalance between teacher and student models. By combining Vision Semantic Distillation (VSD) and Vision-Language Affinity Distillation (VLAD), EM-KD transfers knowledge from a vanilla teacher to an efficient student MLLM, achieving an average score of 50.4 across 11 benchmarks at 144 tokens/patch — surpassing LLaVA-NeXT with 576 tokens (49.4) while delivering nearly 2× inference speedup.
Rethinking Visual Token Reduction in LVLMs under Cross-Modal Misalignment: This paper identifies three forms of cross-modal misalignment (causal, semantic, and spatial) in text-guided visual token importance estimation within LVLMs, and proposes VisionDrop—a training-free progressive token pruning framework that relies exclusively on visual self-attention. The framework performs multi-stage compression across both the visual encoder and LLM decoder, retaining over 91% of original performance while keeping only 5.6% of tokens.

🎵 Audio & Speech (30)¶

DiffA: Large Language Diffusion Models Can Listen and Understand: Proposed DIFFA, the first Large Audio-Language Model based on a diffusion language model. By using a frozen LLaDA-8B backbone, a lightweight dual-adapter architecture, and a two-stage training pipeline, it achieves performance competitive with autoregressive baselines on MMSU, MMAU, and VoiceBench using only 960 hours of ASR data and 127 hours of synthetic instruction data.
Listening Between the Frames: Bridging Temporal Gaps in Large Audio-Language Models: This paper proposes TimeAudio, which endows large audio-language models (LALMs) with precise temporal localization capabilities and end-to-end long audio understanding through three key modules: Temporal Markers, Absolute Time-aware Encoding, and Segment-level Token Merging. It also constructs the FTAR dataset for instruction tuning in fine-grained temporal reasoning.
Say More with Less: Variable-Frame-Rate Speech Tokenization via Adaptive Clustering and Implicit Duration Coding: VARSTok is proposed, which is the first fully dynamic variable-frame-rate speech tokenizer. Through temporal-aware density peak clustering and implicit duration coding, it achieves adaptive token allocation, outperforming fixed-frame-rate baselines while using fewer tokens.
Improving Multimodal Sentiment Analysis via Modality Optimization and Dynamic Primary Modality Selection: Proposes the MODS framework, which eliminates redundant non-verbal modalities via Graph-based Dynamic Compression (GDC), and designs a sample-level dynamic primary modality selector (MSelector) along with a primary-modality-centric cross-attention (PCCA) mechanism to achieve adaptive dominant modality selection per sample in MSA.
CCFQA: A Benchmark for Cross-Lingual and Cross-Modal Speech and Text Factuality Evaluation: CCFQA is proposed—the first cross-lingual and cross-modal benchmark covering 8 languages and 14,400 fully parallel speech-text factuality QA samples. It supports four task settings (QA/XQA/SQA/XSQA) and systematically reveals the factual inconsistency of existing MLLMs under language and modality switching. Concurrently, LLM-SQA is introduced, achieving cross-lingual spoken QA transfer via English bridging with only five shots, obtaining an F1 of 51.4 on XSQA, outperforming GPT-4o-mini-Audio (45.7).
Do LLMs Feel? Teaching Emotion Recognition with Prompts, Retrieval, and Curriculum Learning: This work proposes the PRC-Emo framework, which integrates explicit/implicit emotion prompts, a dedicated retrieval repository, and curriculum learning strategies to comprehensively enhance LLM performance on the Emotion Recognition in Conversation (ERC) task, achieving state-of-the-art (SOTA) results on the IEMOCAP and MELD benchmarks.
DualSpeechLM: Towards Unified Speech Understanding and Generation via Dual Speech Token Modeling: This paper proposes the DualSpeechLM framework, which leverages an understanding-driven speech tokenizer (USTokenizer) to extract high-level semantic tokens as LLM inputs and acoustic tokens as outputs. This approach simultaneously optimizes speech understanding and generation capabilities within a unified, end-to-end framework.
Hearing More with Less: Multi-Modal Retrieval-and-Selection Augmented Conversational LLM-Based ASR: MARS proposes a multi-modal retrieval-and-selection approach to select the most relevant historical context for conversational LLM-ASR (instead of using a fixed set of preceding sentences or the entire history). It outperforms the SOTA system TEA-ASLP, which is trained on 179K hours of data, using only 1.5K hours of training data.
HQ-SVC: Towards High-Quality Zero-Shot Singing Voice Conversion in Low-Resource Scenarios: This work proposes the HQ-SVC framework, which jointly extracts content and speaker features based on a decoupled audio codec (FACodec). Combined with an Enhanced Voice Adapter (EVA) to fuse acoustic features such as pitch and energy, progressive optimization is performed using DDSP and a diffusion model. Using a single RTX 3090 and less than 80 hours of singing voice data, the framework achieves zero-shot singing voice conversion quality that outperforms large-scale training baselines, while additionally supporting speech super-resolution tasks.
End-to-end Contrastive Language-Speech Pretraining Model For Long-form Spoken Question Answering: This paper proposes CLSR, an end-to-end contrastive language-speech retriever. By first converting acoustic representations into text-like representations and then aligning them with text, it efficiently extracts question-relevant segments from long-form audio, providing RAG support for downstream LALM-based long spoken question answering.

Browse all 30 Audio & Speech papers →

🔎 AIGC Detection (2)¶

Optimized Algorithms for Text Clustering with LLM-Generated Constraints: Proposes the LSCK-HC framework, which utilizes LLMs to generate must-link/cannot-link constraints in set formats (rather than traditional pairwise constraints), and integrates with a penalized local search clustering algorithm. It achieves clustering accuracy comparable to SOTA on 5 short text datasets while reducing the number of LLM queries by over 20 times.
BAID: A Benchmark for Bias Assessment of AI Detectors: Proposes the BAID benchmark dataset (208k sample pairs covering 7 bias dimensions and 41 subgroups) to systematically evaluate the fairness performance of four open-source AI text detectors across various demographic and linguistic subgroups, revealing significant recall disparities of detectors on dialects, informal English, and minority texts.

🧊 3D Vision (79)¶

Griffin: Aerial-Ground Cooperative Detection and Tracking Dataset and Benchmark: Introduces Griffin, an aerial-ground cooperative (AGC) 3D perception dataset and benchmark framework. It comprises 250+ dynamic scenes (37K+ frames), achieving realistic UAV dynamics, variable cruising altitudes (20–60m), and occlusion-aware annotations through CARLA-AirSim co-simulation, alongside a systematic robustness evaluation protocol.
MoBGS: Motion Deblurring Dynamic 3D Gaussian Splatting for Blurry Monocular Video: MoBGS proposes an end-to-end dynamic deblurring 3D Gaussian Splatting framework. Through two core modules, Blur-adaptive Latent Camera Estimation (BLCE) and Latent Camera-induced Exposure Estimation (LCEE), it reconstructs sharp spatiotemporal novel views from blurry monocular videos, significantly outperforming existing SOTA methods on the Stereo Blur dataset.
VGGT-DP: Generalizable Robot Control via Vision Foundation Models: The paper proposes VGGT-DP, a bio-inspired visuomotor policy framework that combines the pre-trained 3D-aware foundation model VGGT as a vision encoder with Diffusion Policy. Through three key designs—frame-wise token reuse, random token pruning, and proprioception-guided vision learning—it significantly outperforms DP and DP3 baselines on high-precision manipulation tasks in MetaWorld.
OpenScan: A Benchmark for Generalized Open-Vocabulary 3D Scene Understanding: This paper proposes the Generalized Open-Vocabulary 3D Scene Understanding (GOV-3D) task and the corresponding OpenScan benchmark, extending 3D scene understanding from object categories to eight linguistic property dimensions, revealing the severe limitations of existing OV-3D methods in understanding abstract object properties.
SplatSSC: Decoupled Depth-Guided Gaussian Splatting for Semantic Scene Completion: This paper proposes SplatSSC, which addresses the issue of inefficient random initialization and floating artifacts from outlier primitives in the object-centric paradigm through a depth-guided Gaussian primitive initialization strategy and a Decoupled Gaussian Aggregator (DGA). It achieves an IoU gain of 6.3% and a mIoU gain of 4.1% on Occ-ScanNet, while reducing latency and memory costs by over 9.3%.
NURBGen: High-Fidelity Text-to-CAD Generation through LLM-Driven NURBS Modeling: This paper proposes NURBGen, the first text-to-CAD generation framework based on NURBS surface representation. By fine-tuning an LLM to convert natural language descriptions into structured NURBS parameter JSONs, and introducing a hybrid representation (untrimmed NURBS + analytical primitives) alongside the large-scale partABC dataset, it significantly outperforms existing methods in geometric fidelity and dimensional accuracy.
Redundant Queries in DETR-Based 3D Detection: Unnecessary and Prunable: Proposes GPQ (Gradually Pruning Queries) to progressively prune a large number of redundant object queries in DETR-based 3D detectors based on classification scores. Removing queries requires no extra learnable parameters and can be directly accomplished by fine-tuning on pre-trained checkpoints, achieving up to a 67.86% FLOPs reduction and a 65.16% inference time reduction on edge devices.
Rethinking Rainy 3D Scene Reconstruction via Perspective Transforming and Brightness Tuning: Proposes the OmniRain3D dataset (the first rainy 3D scene dataset that simultaneously models perspective heterogeneity and brightness dynamicity) alongside the REVR-GSNet end-to-end framework (joint recursive brightness enhancement + Gaussian primitives optimization + GS-guided deraining) to reconstruct high-fidelity clean 3D scenes from rain-degraded images.
Sparse4DGS: 4D Gaussian Splatting for Sparse-Frame Dynamic Scene Reconstruction: This paper proposes Sparse4DGS, the first 4D dynamic scene reconstruction method designed for sparse-frame inputs. Through two core modules, Texture-Aware Deformation Regularization (TADR) and Texture-Aware Canonical Optimization (TACO), it guides the Gaussian distribution to focus on texture-rich areas, achieving high-quality dynamic novel view synthesis with only 5–30 sparse input frames.
MeshSplat: Generalizable Sparse-View Surface Reconstruction via Gaussian Splatting: MeshSplat is proposed as the first generalizable sparse-view surface reconstruction framework based on 2D GS. By introducing a weighted Chamfer Distance loss to regularize depth predictions and an uncertainty-based normal prediction network to align 2D GS orientations, it learns geometric priors from novel view synthesis in a self-supervised manner, achieving state-of-the-art performance in sparse-view mesh reconstruction and cross-dataset generalization.

Browse all 79 3D Vision papers →

🎯 Object Detection (29)¶

SM3Det: A Unified Model for Multi-Modal Remote Sensing Object Detection: SM3Det introduces the M2Det task for remote sensing (multi-modal datasets + multi-task object detection), employing a grid-level sparse MoE backbone and a Dynamic Sub-module Optimization (DSO) mechanism to handle SAR/optical/infrared modalities with both horizontal and oriented bounding box detection in a single unified model, substantially outperforming three independently trained modality-specific models combined.
CountVid: Open-World Object Counting in Videos: This paper proposes CountVid, a model, and the VideoCount benchmark, presenting the first systematic study of open-world video object counting—given a text or image description specifying target objects, the system enumerates all unique instances in a video. By combining an image counting model with a promptable video segmentation and tracking model, CountVid addresses challenges such as occlusion and re-appearance, achieving substantial improvements over strong baselines across diverse scenarios including TAO, MOT20, penguin colonies, and X-ray metal crystallization.
Correcting False Alarms from Unseen: Adapting Graph Anomaly Detectors at Test Time: This paper proposes TUNE, a plug-and-play test-time adaptation framework that addresses the "normality shift" problem in graph anomaly detection—caused by the emergence of new normal node categories—by transforming node features via a graph aligner. It leverages the degree of aggregation contamination as an unsupervised adaptation signal and significantly enhances the generalization of various pretrained GAD models across 10 real-world datasets.
Harnessing Vision-Language Models for Time Series Anomaly Detection: A two-stage zero-shot time series anomaly detection framework is proposed: ViT4TS employs a lightweight ViT to perform multi-scale cross-patch matching on line-chart renderings of time series for candidate anomaly interval localization, while VLM4TS leverages GPT-4o with global temporal context to validate and refine detection results. The framework achieves F1-max surpassing the best baseline by 24.6% across 11 benchmarks, with token consumption only 1/36 of existing LLM-based methods.
AerialMind: Towards Referring Multi-Object Tracking in UAV Scenarios: This paper introduces AerialMind, the first large-scale Referring Multi-Object Tracking (RMOT) benchmark dataset for UAV scenarios, and proposes HawkEyeTrack (HETrack), a method that achieves language-guided multi-object tracking in aerial UAV scenes via a co-evolutionary fusion encoder and a scale-adaptive contextual refinement module.
When Trackers Date Fish: A Benchmark and Framework for Underwater Multiple Fish Tracking: This paper proposes MFT25, a large-scale underwater multiple fish tracking dataset (15 sequences, 408K annotations), and SU-T, a tracking framework combining UKF with FishIoU, achieving state-of-the-art performance of 34.1 HOTA and 44.6 IDF1. Statistical analyses further reveal fundamental differences between fish tracking and terrestrial object tracking.
Commonality in Few: Few-Shot Multimodal Anomaly Detection via Hypergraph-Enhanced Memory: This paper proposes CIF, which leverages hypergraphs to extract intra-class structural commonalities from a small number of training samples, guiding memory bank construction and retrieval for few-shot multimodal industrial anomaly detection, achieving state-of-the-art performance.
YOLO-IOD: Towards Real Time Incremental Object Detection: This work is the first to systematically integrate incremental object detection (IOD) into the YOLO real-time framework. It identifies three types of knowledge conflict, proposes a three-module solution (CPR + IKS + CAKD), and introduces the more realistic LoCo COCO benchmark for evaluation.
PromptMoE: Generalizable Zero-Shot Anomaly Detection via Visually-Guided Prompt Mixing of Experts: PromptMoE shifts prompt learning from a monolithic paradigm to a compositional one. Through a visually-guided Mixture of Experts (MoE) mechanism, it dynamically assembles instance-adaptive normal/abnormal state prompts from a learnable semantic primitive bank, achieving state-of-the-art zero-shot anomaly detection (ZSAD) performance across 15 industrial and medical datasets.
SimROD: A Simple Baseline for Raw Object Detection with Global and Local Enhancements: This paper proposes SimROD, an extremely lightweight (only 0.003M parameters) RAW image object detection method that surpasses complex state-of-the-art approaches on multiple RAW detection benchmarks through global Gamma enhancement (4 learnable parameters) and green channel-guided local enhancement.

Browse all 29 Object Detection papers →

✂️ Segmentation (29)¶

LWGANet: Addressing Spatial and Channel Redundancy in Remote Sensing Visual Tasks with Light-Weight Grouped Attention: To address spatial redundancy (large homogeneous backgrounds) and channel redundancy (inefficient single feature space under extreme scale variation) in remote sensing imagery, LWGANet is proposed as a lightweight backbone that achieves efficient multi-scale feature representation via Top-K sparse global feature interaction (TGFI) and heterogeneous grouped attention (LWGA) modules, achieving SOTA across 4 remote sensing task categories on 12 datasets.
RSVG-ZeroOV: Exploring a Training-Free Framework for Zero-Shot Open-Vocabulary Visual Grounding in Remote Sensing Images: This paper proposes RSVG-ZeroOV, a training-free framework that integrates cross-attention maps from a VLM and self-attention maps from a diffusion model via an "Overview-Focus-Evolve" three-stage strategy, enabling zero-shot open-vocabulary visual grounding in remote sensing images.
Breaking the Stealth-Potency Trade-off in Clean-Image Backdoors with Generative Trigger Optimization: This paper proposes Generative Clean-Image Backdoors (GCB), which employs a Conditional InfoGAN (C-InfoGAN) to automatically discover naturally occurring, task-irrelevant features within images as backdoor triggers. GCB achieves high attack success rates (ASR ≥ 90%) at extremely low poison rates (≤ 0.5%) with negligible degradation of clean accuracy (CA drop ≤ 1%), thereby becoming the first method to break the inherent stealth-potency trade-off in clean-image backdoor attacks.
Empowering DINO Representations for Underwater Instance Segmentation via Aligner and Prompter: This paper is the first to introduce DINOv2 into underwater instance segmentation. Through two adaptation modules—AquaStyle Aligner (Fourier frequency-domain style injection) and ObjectPrior Prompter (binary mask prior prompting)—the proposed DiveSeg achieves efficient domain adaptation and substantially outperforms SAM-based methods on the UIIS and USIS10K benchmarks with fewer parameters.
S5: Scalable Semi-Supervised Semantic Segmentation in Remote Sensing: This paper proposes the S5 framework, which for the first time extends semi-supervised semantic segmentation into a pre-training paradigm for remote sensing foundation models (RSFMs). By constructing the million-scale RS4P-1M dataset and introducing a MoE-based multi-dataset fine-tuning strategy, S5 achieves state-of-the-art performance across multiple remote sensing segmentation and detection benchmarks.
Generalizable Slum Detection from Satellite Imagery with Mixture-of-Experts: This paper proposes GRAM (Generalized Region-Aware Mixture-of-Experts), a two-stage test-time adaptation framework. In the first stage, a MoE architecture is used to train region-specialized experts on million-scale satellite imagery from 12 cities. In the second stage, reliable pseudo-labels are selected via cross-region prediction consistency for self-training, enabling generalization of slum segmentation to unseen African cities.
Target Refocusing via Attention Redistribution for Open-Vocabulary Semantic Segmentation: An Explainability Perspective: This paper systematically investigates CLIP's internal mechanisms from an explainability perspective, revealing a "distraction" phenomenon in which CLIP allocates substantial attention resources to target-irrelevant tokens in deeper layers. The proposed training-free method RF-CLIP performs attention redistribution to refocus dispersed resources onto target regions, achieving state-of-the-art performance across 8 benchmarks while maintaining inference efficiency.
Causal-Tune: Mining Causal Factors from Vision Foundation Models for Domain Generalized Semantic Segmentation: This paper proposes Causal-Tune, a causality-driven VFM fine-tuning strategy that decomposes VFM features into causal (domain-invariant) and non-causal (domain-specific) components via DCT frequency-domain transformation and Gaussian band-pass filtering. Learnable tokens are applied exclusively to the causal components for refinement, effectively suppressing VFM artifacts and improving generalization in domain generalized semantic segmentation.
JoDiffusion: Jointly Diffusing Image with Pixel-Level Annotations for Semantic Segmentation Promotion: This paper proposes JoDiffusion, a framework that jointly diffuses images and pixel-level annotation masks in latent space, enabling—for the first time—simultaneous generation of semantically consistent image–annotation pairs conditioned solely on text prompts. JoDiffusion substantially outperforms existing Image2Mask and Mask2Image methods on Pascal VOC, COCO, and ADE20K.
RS2-SAM2: Customized SAM2 for Referring Remote Sensing Image Segmentation: This paper proposes RS2-SAM2, a framework that injects textual information into the SAM2 image encoding process via a Bidirectional Hierarchical Fusion Module (BHFM) and designs a Mask Prompt Generator (MPG) to supply SAM2 with dense prompts, achieving state-of-the-art performance on referring remote sensing image segmentation.

Browse all 29 Segmentation papers →

🖼️ Image Restoration (10)¶

Clear Nights Ahead: Towards Multi-Weather Nighttime Image Restoration: This paper is the first to define and explore the multi-weather nighttime image restoration task. It constructs the AllWeatherNight dataset (8K training + 1K synthetic test + 1K real-world test) and proposes the ClearNight unified framework, which simultaneously removes compound degradations—haze, rain streaks, raindrops, snow, and flare—in a single stage via Retinex dual-prior guidance and weather-aware dynamic specificity–commonality collaboration. With only 2.84M parameters, ClearNight comprehensively surpasses state-of-the-art methods.
Depth-Synergized Mamba Meets Memory Experts for All-Day Image Reflection Separation: This paper proposes DMDNet, which employs a depth-aware scanning strategy (DAScan) to guide Mamba toward salient structures, incorporates a depth-synergized state space model (DS-SSM) to suppress ambiguous feature propagation, and introduces a memory expert compensation module (MECM) to leverage cross-image historical knowledge, achieving all-day (daytime + nighttime) image reflection separation.
MFmamba: A Multi-function Network for Panchromatic Image Resolution Restoration Based on State-Space Model: This paper proposes MFmamba, a multi-function network built upon a UNet++ backbone that integrates a Mamba Upsampling Block (MUB), Dual Pooling Attention (DPA), and a Multi-scale Hybrid Cross Block (MHCB). Using only panchromatic (PAN) images as input, the unified framework simultaneously supports three tasks: super-resolution, spectral restoration, and joint SR with colorization.
ICLR: Inter-Chrominance and Luminance Interaction for Natural Color Restoration in Low-Light Image Enhancement: Targeting two overlooked statistical distribution issues in the HVI color space — large distribution discrepancy between chrominance and luminance branches leading to insufficient complementary feature extraction, and weak inter-chrominance correlation causing gradient conflicts — this paper proposes the ICLR framework. It introduces a Dual-stream Interaction Enhancement Module (DIEM) and a Covariance Correction Loss (CCL) to address these issues from the perspectives of fusion enhancement and statistical distribution optimization, respectively, achieving state-of-the-art performance on the LOL benchmark series.
SpatioTemporal Difference Network for Video Depth Super-Resolution: Motivated by the statistical observation that spatially non-smooth regions and temporally varying regions in video depth super-resolution (VDSR) follow long-tail distributions, this paper proposes STDNet. The method incorporates a spatial difference branch (learning spatial difference representations for intra-frame RGB-D adaptive aggregation) and a temporal difference branch (exploiting temporal difference representations for motion compensation in changing regions). On the TarTanAir dataset at ×16 super-resolution, RMSE is reduced from 112.04 cm to 96.80 cm, outperforming state-of-the-art methods by an average of 27.6%–32.6%.
Temporal Inconsistency Guidance for Super-resolution Video Quality Assessment: This paper proposes TIG-SVQA, a framework that, for the first time, incorporates temporal inconsistency as an explicit guidance signal for super-resolution video quality assessment. The framework introduces an Inconsistency-Highlighted Spatial Module (IHSM) and an Inconsistency-Guided Temporal Module (IGTM), achieving SRCC scores of 0.950, 0.942, and 0.939 on the SFD, MFD, and Combined-VSR datasets, respectively, surpassing all existing IQA/VQA methods.
TMDC: A Two-Stage Modality Denoising and Complementation Framework for Multimodal Sentiment Analysis: This paper proposes TMDC, a two-stage framework in which the first stage learns denoised modality-specific and modality-common representations on complete data, and the second stage leverages denoised representations from available modalities to reconstruct missing ones — marking the first joint treatment of noise and missing modalities in MSA.
Blur-Robust Detection via Feature Restoration: An End-to-End Framework for Prior-Guided Infrared UAV Target Detection: This paper proposes JFD3, an end-to-end dual-branch framework that performs deblurring in the feature domain rather than the image domain, and leverages frequency structure priors to guide the detection network, achieving high-accuracy real-time infrared UAV target detection under motion blur conditions.
SD-PSFNet: Sequential and Dynamic Point Spread Function Network for Image Deraining: SD-PSFNet is a cascaded CNN-based deraining network driven by a dynamic PSF mechanism. It models the optical effects of raindrops via a multi-scale learnable PSF dictionary, combined with a sequential restoration architecture featuring adaptive gated fusion. The method achieves SOTA performance of 33.12 dB on Rain100H and 42.28 dB on RealRain-1k-L, yielding a cumulative gain of 5.04 dB (13.5%) over the baseline MPRNet.
RefiDiff: Progressive Refinement Diffusion for Efficient Missing Data Imputation: RefiDiff proposes a four-stage framework (pre-processing → warm-up → diffusion → polishing) that progressively unifies the predictive and generative imputation paradigms for the first time. Combined with a Mamba-based denoising network, it achieves state-of-the-art performance across 9 datasets while running 4× faster than DIFFPUTER.

🛰️ Remote Sensing (7)¶

TDCNet: Spatio-Temporal Context Learning with Temporal Difference Convolution for Moving IRSTD: This paper proposes TDCNet, which unifies temporal difference and 3D convolution into a single Temporal Difference Convolution (TDC). Through re-parameterization, TDC introduces zero additional inference overhead. Combined with TDC-guided spatio-temporal attention (TDCSTA), TDCNet achieves an F1 of 97.12% (AP50 93.83%) on the newly constructed IRSTD-UAV dataset, which contains 15,106 frames of real infrared UAV imagery.
Debiasing Machine Learning Predictions for Causal Inference Without Additional Ground Truth Data: To address the attenuation of causal treatment effects caused by regression-to-the-mean in ML-based satellite poverty predictions, this paper proposes two post-processing correction methods that require no additional labeled data — Linear Calibration Correction (LCC) and Tweedie local unshrinking — enabling a single prediction map to be reused across multiple downstream causal studies (the "One Map, Many Trials" paradigm). Tweedie correction achieves near-unbiased treatment effect estimation on both simulated and real DHS data.
UniABG: Unified Adversarial View Bridging and Graph Correspondence for Unsupervised Cross-View Geo-Localization: This paper proposes UniABG, a two-stage unsupervised cross-view geo-localization framework that employs View-Aware Adversarial Bridging (VAAB) to eliminate the domain gap between UAV and satellite views, followed by Heterogeneous Graph Filtering Calibration (HGFC) to purify cross-view correspondences. UniABG achieves 93.29% Satellite→Drone AP on University-1652, surpassing most supervised methods.
Consistency-based Abductive Reasoning over Perceptual Errors of Multiple Pre-trained Models in Novel Environments: This paper models conflicting predictions from multiple pre-trained perception models in novel environments as a consistency-based abductive reasoning problem. Error detection rules and domain constraints for each model are encoded as logic programs, and an optimal hypothesis is sought that maximizes prediction coverage while keeping the inconsistency rate below a threshold. The approach achieves an average F1 improvement of 13.6% across 15 aerial test sets.
M3SR: Multi-Scale Multi-Perceptual Mamba for Efficient Spectral Reconstruction: This paper proposes M3SR, a Mamba-based multi-scale multi-perceptual architecture that integrates spatial, frequency, and spectral branches in parallel within a U-Net multi-scale structure. With only 2.17M parameters and 100.9G FLOPs, M3SR surpasses existing state-of-the-art methods on four spectral reconstruction benchmarks.
Perceive, Act and Correct: Confidence Is Not Enough for Hyperspectral Classification: This paper proposes the CABIN framework, which employs a closed-loop cognitive perceive–act–correct learning mechanism. By replacing naive confidence with epistemic uncertainty to guide sample selection and pseudo-label management in semi-supervised hyperspectral image classification, CABIN significantly outperforms fully supervised baselines while using only 75% of the labeled data.
Machine Learning for Sustainable Rice Production: Region-Scale Monitoring of Water-Saving Practices in Punjab, India: This paper proposes a dimensional classification approach that decouples the recognition of water-saving rice practices into two independent binary classification tasks — a seeding dimension (DSR vs. PTR) and an irrigation dimension (AWD vs. CF). Using only Sentinel-1 SAR imagery, the method achieves seeding F1=0.80 and irrigation F1=0.74, and performs large-scale inference over 3 million+ parcels in Punjab, with district-level adoption rates strongly correlated with government statistics (Spearman ρ=0.69).

🧑 Human Understanding (20)¶

ReAlign: Text-to-Motion Generation via Step-Aware Reward-Guided Alignment: This paper proposes ReAlign (Reward-guided sampling Alignment), which employs a step-aware reward model and a reward-guided sampling strategy to dynamically steer sampling trajectories toward distributions with high text-motion alignment during diffusion inference, significantly improving the generation quality of various motion generation methods without fine-tuning any diffusion model. Using MLD as a baseline, R@1 improves by 17.9% and FID improves by 58.8%.
New Synthetic Goldmine: Hand Joint Angle-Driven EMG Data Generation Framework for Micro-Gesture Recognition: This paper proposes SeqEMG-GAN, a conditional adversarial generation framework driven by hand joint angle sequences. Through the joint design of an angle encoder, a two-level context encoder (featuring the novel Ang2Gist unit), a deep convolutional generator, and a multi-view discriminator, the framework synthesizes high-fidelity EMG signals from joint kinematic trajectories, enabling zero-shot generation for unseen gestures. Mixing synthetic and real data for training improves classification accuracy from 57.77% to 60.53%.
CLIP-FTI: Fine-Grained Face Template Inversion via CLIP-Driven Attribute Conditioning: This paper presents the first approach to leverage CLIP-extracted fine-grained facial semantic attribute embeddings for Face Template Inversion (FTI). A cross-modal feature interaction network fuses leaked templates with attribute embeddings and projects them into the StyleGAN latent space, synthesizing identity-consistent face images with richer attribute details. The method surpasses state-of-the-art in recognition accuracy, attribute similarity, and cross-model attack transferability.
mmPred: Radar-based Human Motion Prediction in the Dark: This work is the first to introduce millimeter-wave radar into human motion prediction (HMP), proposing mmPred — a diffusion-based framework that employs dual-domain historical motion representations (time-domain pose refinement TPR + frequency-domain dominant motion FDM) and a Global Skeleton Transformer (GST) to effectively suppress radar-specific noise and temporal inconsistency, surpassing SOTA methods by 8.6% and 22% on the mmBody and mm-Fi datasets, respectively.
Generating Attribute-Aware Human Motions from Textual Prompt: This paper proposes AttrMoGen, a framework that decouples action semantics from human attributes (age, gender, etc.) via a Structural Causal Model (SCM)-based Causal Information Bottleneck, enabling attribute-aware human motion generation from text prompts. The authors also introduce HumanAttr, the first large-scale text-motion dataset with extensive attribute annotations.
Facial-R1: Aligning Reasoning and Recognition for Facial Emotion Analysis: This paper proposes Facial-R1, a three-stage alignment training framework (SFT → RL → Data Synthesis) that aligns the reasoning process of VLMs with emotion recognition outcomes by treating AU and emotion labels as verifiable reward signals. The framework achieves state-of-the-art performance on 8 benchmarks and introduces the FEA-20K dataset.
PA-FAS: Towards Interpretable and Generalizable Multimodal Face Anti-Spoofing via Path-Augmented Reinforcement Learning: This paper proposes PA-FAS, a framework that addresses two critical bottlenecks of the SFT+RL paradigm in multimodal FAS — insufficient reasoning path diversity and reasoning shortcut — via a Reasoning Path Augmentation strategy and an answer shuffling mechanism, achieving the first unified solution for multimodal fusion, domain generalization, and interpretability simultaneously.
KineST: A Kinematics-guided Spatiotemporal State Space Model for Human Motion Tracking from Sparse Signals: This paper proposes KineST, a kinematics-guided state space model that reconstructs whole-body motion from sparse HMD signals via a kinematic tree bidirectional scanning strategy and hybrid spatiotemporal representation learning, surpassing state-of-the-art methods in both accuracy and temporal consistency.
Toward Gaze Target Detection in Young Autistic Children: To address the severe class imbalance in gaze target detection for autistic children—where face-directed gaze accounts for only 6.6% of samples—this paper proposes the Socially Aware Coarse-to-Fine (SACF) framework. A fine-tuned Qwen2.5-VL serves as a social-context-aware gate that routes inputs to either a socially aware or a socially agnostic expert model. Evaluated on the newly introduced AGT dataset, the framework substantially improves face gaze detection performance (Face L2 reduced by 13.9% on Sharingan; F1 improved from 0.753 to 0.761).
CoordAR: One-Reference 6D Pose Estimation of Novel Objects via Autoregressive Coordinate Map Generation: This paper proposes CoordAR, which formulates 3D-3D correspondence estimation in single-reference-view 6D pose estimation as an autoregressive generation problem over discrete tokens. Through coordinate map tokenization, modality-decoupled encoding, and an autoregressive Transformer decoder, CoordAR substantially outperforms existing single-view methods on multiple benchmarks and demonstrates strong robustness to challenging scenarios such as symmetry and occlusion.

Browse all 20 Human Understanding papers →

📹 Video Understanding (27)¶

TSPO: Temporal Sampling Policy Optimization for Long-form Video Language Understanding: This paper formulates keyframe selection and language generation as a joint decision-making process, and optimizes a lightweight temporal agent's sampling policy end-to-end via GRPO-based reinforcement learning. It achieves state-of-the-art results on four long-form video understanding benchmarks (LongVideoBench +5.0%, MLVU +6.0% on LLaVA-Video-7B) and transfers zero-shot to other Video-MLLMs.
Rethinking Progression of Memory State in Robotic Manipulation: An Object-Centric Perspective: This paper proposes LIBERO-Mem, a benchmark comprising 10 non-Markovian robotic manipulation tasks, and Embodied-SlotSSM, an object-centric memory VLA framework combining Slot Attention with state space models, to address the failure of visuomotor policies in long-horizon tasks that require object-level historical reasoning under partial observability.
Predicting Video Slot Attention Queries from Random Slot-Feature Pairs: This paper proposes RandSF.Q, which significantly improves query prediction quality in video object-centric learning (OCL) by leveraging next-frame features for informative query prediction and learning transition dynamics from randomly sampled slot-feature pairs. The method surpasses state-of-the-art approaches by up to 10 points on object discovery benchmarks.
Learning to Tell Apart: Weakly Supervised Video Anomaly Detection via Disentangled Semantic Alignment: This paper proposes DSANet, which enhances the discriminability between normal and anomalous features in weakly supervised video anomaly detection (WS-VAD) at two levels: coarse-grained self-guided normal pattern modeling (SG-NM) and fine-grained disentangled contrastive semantic alignment (DCSA). DSANet achieves state-of-the-art performance with 86.95% AP (+1.14%) on XD-Violence and 13.01% fine-grained mAP (+3.39%) on UCF-Crime.
Causality Matters: How Temporal Information Emerges in Video Language Models: Through systematic ablation experiments, this work demonstrates that the temporal understanding capability of VideoLMs does not originate from positional encoding (PE), but rather emerges from the sequence sensitivity of causal attention masks. Temporal information is constructed layer by layer along a causal pathway of "inter-frame interaction → last-frame aggregation → query integration," based on which two lossless inference acceleration strategies are proposed.
APVR: Hour-Level Long Video Understanding with Adaptive Pivot Visual Information Retrieval: This paper proposes APVR, a training-free dual-granularity visual information retrieval framework. At the frame level, it iteratively retrieves keyframes (up to 1024) via query expansion and spatiotemporal semantic confidence scoring; at the token level, it compresses visual tokens through query-aware attention-driven selection. APVR overcomes memory limitations to process hour-long videos, achieving improvements of up to 9.5%, 4.6%, and 9.7% on LongVideoBench, VideoMME, and MLVU, respectively.
LiViBench: An Omnimodal Benchmark for Interactive Livestream Video Understanding: This paper presents LiViBench, the first omnimodal benchmark for interactive livestream video understanding (3,168 videos, 3,175 MCQs, 24 tasks), introduces a multi-agent seed-guided semi-automatic annotation pipeline, and develops LiVi-LLM-7B — a specialized model featuring a Video-to-Comment Retrieval (VCR) module and two-stage instruction tuning — which surpasses 72B open-source models at the 7B scale.
HeadHunt-VAD: Hunting Robust Anomaly-Sensitive Heads in MLLM for Tuning-Free Video Anomaly Detection: This paper proposes HeadHunt-VAD, which systematically identifies a sparse set of anomaly-sensitive and stable attention heads within a frozen MLLM, bypassing the information loss inherent in text-based outputs. Using a lightweight classifier, it achieves efficient tuning-free video anomaly detection, establishing state-of-the-art performance among tuning-free methods on UCF-Crime and XD-Violence.
ReaSon: Reinforced Causal Search with Information Bottleneck for Video Understanding: This paper proposes a Causal Information Bottleneck (CIB) theoretical framework that formalizes keyframe selection as an information-theoretic problem jointly optimizing predictive sufficiency and causal necessity. Built upon CIB, the ReaSon reinforcement learning framework trains a selection policy using three CIB-aligned rewards (answer reward, cycle-consistency reward, and counterfactual reward), significantly outperforming existing methods under constrained frame budgets.
SUGAR: Learning Skeleton Representation with Visual-Motion Knowledge for Action Recognition: This paper proposes the SUGAR paradigm, which leverages GPT-generated motion descriptions and visual descriptions as prior knowledge to supervise skeleton encoders via contrastive learning, producing more discriminative representations. These representations are then fed into an LLM (LLaMA2-7B) with untouched pretrained weights as the classifier, complemented by a newly designed Temporal Query Projection (TQP) module for efficient skeleton-based action classification and zero-shot inference.

Browse all 27 Video Understanding papers →

🚗 Autonomous Driving (56)¶

DriveSuprim: Towards Precise Trajectory Selection for End-to-End Planning: DriveSuprim is proposed to address key challenges in selection-based end-to-end planning—difficulty in distinguishing similar trajectories, orientation bias, and instability of hard labels—via a coarse-to-fine trajectory filtering paradigm, rotation-based data augmentation, and a self-distillation soft label framework, achieving state-of-the-art (SOTA) performance on NAVSIM v1/v2 and Bench2Drive.
FastDriveVLA: Efficient End-to-End Driving via Plug-and-Play Reconstruction-based Token Pruning: FastDriveVLA is proposed to train a lightweight plug-and-play ReconPruner module (only 0.07B parameters) via MAE-style foreground pixel reconstruction. By employing an adversarial foreground-background reconstruction strategy, the method prioritizes the retention of foreground tokens critical for driving decisions. It achieves state-of-the-art performance across all pruning ratios on the nuScenes open-loop planning benchmark, and can be transferred to different VLA models sharing the same visual encoder without retraining.
WorldRFT: Latent World Model Planning with Reinforcement Fine-Tuning for Autonomous Driving: WorldRFT is a planning-oriented latent world model framework that integrates VGGT-based spatial encoding, hierarchical planning decomposition with local-aware iterative refinement, and GRPO-based collision-aware reinforcement fine-tuning. It reduces collision rate by 83% on nuScenes (0.30% → 0.05%) and achieves near-LiDAR SOTA performance using camera only on NavSim (87.8 vs. 88.1 PDMS).
Hierarchical Prompt Learning for Image- and Text-Based Person Re-Identification: This paper proposes HPL, a unified framework that decouples I2I and T2I tasks via a Task-Routed Transformer (dual classification tokens), and employs hierarchical prompt learning (identity-level + instance-level pseudo-text tokens) combined with cross-modal prompt regularization, achieving simultaneous state-of-the-art performance on both image-to-image and text-to-image person re-identification within a single model for the first time.
LiDARCrafter: Dynamic 4D World Modeling from LiDAR Sequences: This paper proposes LiDARCrafter, the first 4D generative world model targeting LiDAR, which achieves controllable 4D LiDAR sequence generation and editing through a pipeline of text → scene graph → three-branch diffusion layout → range-image diffusion generation → autoregressive temporal extension, comprehensively surpassing existing methods on nuScenes.
CompTrack: Information Bottleneck-Guided Low-Rank Dynamic Token Compression for Point Cloud Tracking: Proposes CompTrack—the first 3D single object tracking framework to address the dual challenges of spatial and informational redundancy in LiDAR point clouds simultaneously: the Spatial Foreground Predictor (SFP) filters background noise based on information entropy, and the Information Bottleneck-guided Dynamic Token Compression (IB-DTC) module estimates the effective rank via online SVD to compress the foreground into compact proxy tokens; achieves state-of-the-art performance on nuScenes and Waymo while running in real-time at 90 FPS.
PriorDrive: Enhancing Online HD Map Construction with Unified Vector Priors: This paper proposes PriorDrive, a framework that encodes multiple types of vectorized prior maps (SD maps, outdated HD maps, historical prediction maps) into a unified representation via a Unified Vector Encoder (UVE) and Hybrid Prior Representation (HPQuery), and integrates them into various online mapping models. It achieves a +14.3 mAP improvement on nuScenes and is compatible with both query-based and non-query-based mapping architectures.
DiffRefiner: Coarse to Fine Trajectory Planning via Diffusion Refinement with Semantic Interaction for End to End Autonomous Driving: This paper proposes DiffRefiner, a coarse-to-fine two-stage framework that first generates coarse trajectories using a discriminative Proposal Decoder and then refines them iteratively using a diffusion model. Combined with a fine-grained semantic interaction module, it achieves SOTA performance on both NAVSIM v2 and Bench2Drive benchmarks.
STRIDE-QA: Visual Question Answering Dataset for Spatiotemporal Reasoning in Urban Driving Scenes: This paper constructs STRIDE-QA, the largest spatiotemporal reasoning VQA dataset in autonomous driving (270K frames, 16M QA pairs), defines three categories of spatiotemporal reasoning tasks (object-centric spatial / ego-centric spatial / ego-centric spatiotemporal), and demonstrates that fine-tuning a VLM raises localization success rate from near zero to 55% and temporal localization consistency from 0 to 28.4%.
MambaSeg: Harnessing Mamba for Accurate and Efficient Image-Event Semantic Segmentation: MambaSeg is proposed, employing a dual-branch parallel Mamba encoder to process RGB images and event streams respectively, with a Dual-Dimension Interaction Module (DDIM) for fine-grained cross-modal fusion along both spatial and temporal dimensions. It achieves state-of-the-art performance of 77.56%/75.10% mIoU on DDD17 and DSEC with only 25.44M parameters, offering substantially better efficiency than Transformer-based approaches.

Browse all 56 Autonomous Driving papers →

🤖 Robotics & Embodied AI (30)¶

SemanticVLA: Semantic-Aligned Sparsification and Enhancement for Efficient Robotic Manipulation: This paper proposes the SemanticVLA framework, which integrates three modules — a Semantic-guided Dual-encoder Pruner (SD-Pruner), a Semantic-complementary Hierarchical Fuser (SH-Fuser), and a Semantic-conditioned Action Coupler (SA-Coupler) — to substantially reduce visual redundancy while enhancing instruction–vision–action alignment. On the LIBERO benchmark, SemanticVLA achieves a 97.7% success rate, surpassing OpenVLA by 21.1%, while reducing training cost and inference latency by 3.0× and 2.7×, respectively.
H-GAR: A Hierarchical Interaction Framework via Goal-Driven Observation-Action Refinement for Robotic Manipulation: This paper proposes H-GAR, a hierarchical goal-driven framework that first predicts a goal observation and then synthesizes intermediate observations, while refining coarse-grained actions via a historical action memory bank. This design enables explicit bidirectional interaction between observations and actions, achieving state-of-the-art performance on both simulated and real-robot manipulation tasks.
Actor-Critic for Continuous Action Chunks: A Reinforcement Learning Framework for Long-Horizon Robotic Manipulation with Sparse Reward: AC3 proposes an actor-critic framework that directly learns continuous action sequences (action chunks), stabilizing long-horizon robotic manipulation under sparse rewards via an asymmetric actor update rule—updating the actor only from successful trajectories—and self-supervised anchor-based intrinsic rewards. The method achieves superior success rates over existing approaches across 25 tasks on BiGym and RLBench.
TTF-VLA: Temporal Token Fusion via Pixel-Attention Integration for Vision-Language-Action Models: TTF-VLA proposes a training-free temporal token fusion method that selectively reuses visual tokens from historical frames via a dual-dimension mechanism combining grayscale pixel difference and attention-based semantic detection, improving inference quality of VLA models on robotic manipulation tasks with an average gain of 4.0 percentage points on LIBERO.
Object-Centric Latent Action Learning: This paper proposes an object-centric latent action learning framework that leverages self-supervised object decomposition (VideoSAUR) to disentangle task-relevant entities from visual distractions (e.g., dynamic backgrounds), reducing the performance degradation of LAPO on distracted videos by approximately 50%. A linear action probe is used to automatically select control-relevant slots.
PanoNav: Mapless Zero-Shot Object Navigation with Panoramic Scene Parsing and Dynamic Memory: This paper proposes PanoNav, a mapless zero-shot object navigation framework that uses only RGB images. It unlocks the spatial reasoning capability of MLLMs through Panoramic Scene Parsing and introduces a Dynamic Bounded Memory Queue to prevent local deadlock.
Towards Affordance-Aware Robotic Dexterous Grasping with Human-like Priors: This paper proposes AffordDex, a two-stage framework: the first stage pre-trains human hand motion priors (natural motion trajectories) via imitation learning; the second stage refines the policy through reinforcement learning using a residual module and VLM-guided Negative Affordance Annotation (NAA), achieving dexterous robotic grasping that is both human-like in naturalness and functionally correct (e.g., avoiding the blade and grasping the handle of a knife). The method significantly outperforms state-of-the-art approaches across multiple generalization levels.
Scalable Multi-Objective and Meta Reinforcement Learning via Gradient Estimation: This paper proposes PolicyGradEx, which efficiently estimates policy adaptation performance on arbitrary task subsets via first-order gradient approximation and surrogate models, constructs a task affinity matrix, and performs task grouping through convex optimization. PolicyGradEx outperforms state-of-the-art baselines by an average of 16% on multi-objective RL and meta-RL benchmarks, with a speedup of up to 26×.
UrbanNav: Learning Language-Guided Urban Navigation from Web-Scale Human Trajectories: This paper proposes UrbanNav, which leverages web-scale urban walking videos (1,500+ hours from YouTube, yielding 3 million instruction–trajectory–landmark triplets) to train a language-guided urban navigation policy via an automated annotation pipeline and robust filtering mechanism, achieving an 83.3% navigation success rate in real-world deployment.
SpatialActor: Exploring Disentangled Spatial Representations for Robust Robotic Manipulation: This paper proposes SpatialActor, a framework that explicitly disentangles semantic and geometric representations. It introduces a Semantic-Guided Geometry Module (SGM) that adaptively fuses noisy depth features with a pretrained depth estimation expert prior, and a Spatial Transformer (SPT) that encodes low-level spatial position cues. SpatialActor achieves 87.4% success rate on 50+ RLBench tasks (SOTA +6.0%) and outperforms RVT-2 by 19.4% under heavy-noise conditions.

Browse all 30 Robotics & Embodied AI papers →

🎮 Reinforcement Learning (58)¶

Reasoning with Exploration: An Entropy Perspective: This paper analyzes the positive correlation between exploratory reasoning behaviors in LLMs (pivotal tokens, self-reflection, rare behaviors) and high-entropy regions from an entropy perspective. It proposes a minimalist entropy-based advantage shaping method—requiring only a single line of code modification—that significantly enhances the Pass@K reasoning capability ceiling of LLMs.
Reasoning or Memorization? Unreliable Results of Reinforcement Learning Due to Data Contamination: This paper conducts a systematic data leakage audit revealing severe data contamination of the Qwen2.5 series on standard math benchmarks such as MATH-500. It demonstrates that recent findings claiming "spurious rewards can improve mathematical reasoning" are artifacts of contamination, and constructs a fully uncontaminated benchmark, RandomCalculation, to verify that only correct reward signals yield genuine reasoning improvements.
InfiGUI-G1: Advancing GUI Grounding with Adaptive Exploration Policy Optimization: To address the exploration bottleneck in semantic alignment for GUI grounding, this paper proposes the Adaptive Exploration Policy Optimization (AEPO) framework. AEPO enforces broad exploration via a multi-answer generation strategy, dynamically guides learning through an adaptive exploration reward function, and ensures exploration quality via a collinearity penalty mechanism, significantly improving multimodal large language model performance on complex GUI grounding tasks.
QiMeng-Kernel: Macro-Thinking Micro-Coding Paradigm for LLM-Based High-Performance GPU Kernel Generation: This paper proposes MTMC (Macro Thinking Micro Coding), a hierarchical framework that decouples GPU kernel generation into two stages: a lightweight RL-trained LLM generates high-level optimization actions (Macro Thinking), while a general-purpose LLM incrementally implements each action (Micro Coding). This design separates correctness from performance concerns, achieving near-100% accuracy and a 2.2× speedup over expert-optimized PyTorch Eager kernels on KernelBench.
MARS: Multi-Agent Adaptive Reasoning with Socratic Guidance for Automated Prompt Optimization: This paper proposes MARS, a five-agent framework for automated prompt optimization (APO): a Planner generates task-specific optimization trajectories; a Teacher-Critic-Student triad conducts Socratic dialogue for iterative prompt refinement (simulating pseudo-gradient descent in text space); and a Target agent executes the prompt and provides feedback. The entire process is modeled as a POMDP. MARS outperforms the previous SOTA (PE2) by 6.04% on general tasks and 6.42% on domain-specific tasks across 17 datasets, requiring only 1-shot training data.
One-Step Generative Policies with Q-Learning: A Reformulation of MeanFlow: This paper reformulates MeanFlow from visual generation into a generative policy for offline RL. It proposes a residual-form direct noise-to-action mapping that achieves expressive one-step sampling and enables stable joint optimization with a Q-function in a single training stage, achieving strong performance across 73 tasks on OGBench and D4RL.
Intention-Guided Cognitive Reasoning for Egocentric Long-Term Action Anticipation: This paper proposes INSIGHT, a two-stage unified framework for egocentric long-term action anticipation (LTA). Stage one enhances action representations via hand-object interaction (HOI) region feature extraction and verb-noun co-occurrence matrices; stage two introduces a GRPO-based reinforcement learning cognitive reasoning module that simulates a structured "perceive → reason → answer" cognitive process for intention inference and action prediction.
Where to Start Alignment? Diffusion Large Language Model May Demand a Distinct Position: This paper presents the first systematic safety analysis of diffusion large language models (dLLMs), revealing that—unlike autoregressive LLMs—middle tokens are more critical to safety in dLLMs, and that attackers are fundamentally constrained by the model's inherent sequential generation tendency from manipulating these positions. Based on this asymmetry, the paper proposes MOSA (Middle-tOken Safety Alignment) as a defense method.
DRMD: Deep Reinforcement Learning for Malware Detection under Concept Drift: This paper is the first to reformulate Android malware detection as a one-step Markov Decision Process (MD-MDP) and trains a PPO-based deep reinforcement learning agent, DRMD, that unifies sample classification, rejection, and active learning within a single policy. The approach achieves average AUT improvements of 8.66 (classification only) and 10.90 (with rejection) in multi-year temporal evaluations, significantly outperforming conventional supervised learning classifiers under concept drift.
A Course Correction in Steerability Evaluation: Revealing Miscalibration and Side Effects in LLMs: This paper proposes a multi-dimensional objective-space framework for evaluating LLM steerability, decomposing steering error into miscalibration and side effects (orthogonality). Experiments on text rewriting reveal that even the strongest LLMs produce severe side effects; prompt engineering proves ineffective, best-of-N sampling is prohibitively costly, and RL fine-tuning yields improvements but does not fully resolve the problem.

Browse all 58 Reinforcement Learning papers →

🎁 Recommender Systems (27)¶

Inductive Generative Recommendation via Retrieval-based Speculation: This paper identifies a critical limitation of Generative Recommendation (GR) models — their inability to recommend items unseen during training — and proposes SpecGR, a plug-and-play framework in which an inductively capable drafter model proposes candidate items (including new ones) while the GR model serves as a verifier to rank and validate candidates. A guided re-drafting mechanism further improves verification efficiency, achieving state-of-the-art overall performance across three datasets.
From IDs to Semantics: A Generative Framework for Cross-Domain Recommendation with Adaptive Semantic Tokenization: This paper proposes GenCDR, a framework that introduces the generative semantic ID paradigm into LLM-driven cross-domain recommendation for the first time, via two core modules: domain-adaptive semantic tokenization and cross-domain autoregressive recommendation. GenCDR effectively addresses the non-transferability of item IDs and insufficient domain-personalized modeling in conventional approaches.
Interpretable Reward Model via Sparse Autoencoder: This paper proposes SARM (Sparse Autoencoder-enhanced Reward Model), which integrates a pretrained sparse autoencoder into a reward model to map hidden-layer activations into an interpretable, sparse, monosemantic feature space. This design enables feature-level reward attribution and dynamic preference steering, while achieving the highest overall score among all models on RewardBench 2.
SlideTailor: Personalized Presentation Slide Generation for Scientific Papers: This paper defines a new task of preference-guided paper-to-slide generation and proposes the SlideTailor framework, which distills content preferences from user-provided paper–slide example pairs and aesthetic preferences from .pptx templates. A chain-of-speech mechanism aligns slide content with intended spoken narratives. On the self-constructed PSP benchmark, SlideTailor achieves an overall score of 75.8% and a human-evaluation win rate of 81.63%, significantly outperforming existing methods.
Align³GR: Unified Multi-Level Alignment for LLM-based Generative Recommendation: This paper proposes Align³GR, a unified three-level alignment framework that systematically bridges the semantic-behavioral gap between LLMs and recommender systems at the token level (dual-side SCID), the behavior modeling level (multi-task SFT), and the preference level (progressive DPO).
AutoPP: Towards Automated Product Poster Generation and Optimization: This paper proposes AutoPP, the first pipeline to unify automated product poster generation with CTR-feedback-driven optimization in a single framework. It employs a unified design module to jointly design background, text, and layout; an element rendering module for efficient and controllable poster generation; and Isolated DPO (IDPO) to achieve element-level click-through rate optimization.
FreqRec: Exploiting Inter-Session Information with Frequency-enhanced Dual-Path Networks for Sequential Recommendation: This paper proposes FreqRec, a dual-path architecture that applies frequency-domain transformations along the batch axis and the time axis to capture group-level consumption rhythms across sessions and fine-grained individual user interests, respectively. A frequency-domain consistency loss is introduced to explicitly align predicted and ground-truth frequency spectra. FreqRec achieves up to 7.38% improvement in NDCG@10 over the strongest baseline on three Amazon datasets.
Moral Change or Noise? On Problems of Aligning AI With Temporally Unstable Human Feedback: Through a longitudinal study involving 400+ participants across 3–5 sessions in the domain of kidney transplant allocation, this paper reveals significant temporal instability in human moral preferences (6–20% response change rate) and demonstrates that such instability substantially degrades the predictive performance of AI alignment models, thereby challenging the validity of current alignment approaches that assume static preferences.
When Top-ranked Recommendations Fail: Modeling Multi-Granular Negative Feedback for Explainable and Robust Video Recommendation: This paper proposes ENF (Explainable Negative Feedback), a framework comprising three collaborative MLLM Agents (Profile Agent, Video Agent, and Reason Agent) and a progressive S-GRPO reinforcement learning training strategy. ENF is the first approach to achieve explainable prediction and root-cause analysis of implicit negative feedback in video recommendation systems. Deployed on Tencent's news platform, it achieves a 6.2% increase in average watch duration and a 9.4% decrease in quick-skip rate.
Evaluating LLMs for Police Decision-Making: A Framework Based on Police Action Scenarios: This paper proposes PAS (Police Action Scenarios), an LLM evaluation framework for policing contexts. The framework comprises five stages: scenario definition, reference answer construction, LLM response generation, core metric extraction, and performance interpretation. An evaluation dataset is constructed from 8,000+ official Korean police documents. The study finds that commercial LLMs (GPT-4, Gemini, Claude) perform significantly below reference answers on policing tasks, particularly in factual accuracy and logical correctness.

Browse all 27 Recommender Systems papers →

🔄 Self-Supervised Learning (16)¶

Spikingformer: A Key Foundation Model for Spiking Neural Networks: This paper proposes Spikingformer, which integrates MS Residual with Self-Attention in a spike-driven manner to address the non-spike computation introduced by SEW Residual in Spikformer, while preserving global modeling capability.
From Pretrain to Pain: Adversarial Vulnerability of Video Foundation Models without Finetuning: This paper proposes Transferable Video Attack (TVA), which generates adversarial perturbations solely by exploiting the embedding space of open-source Video Foundation Models (VFMs), without any knowledge of downstream tasks, and effectively attacks downstream models and multimodal LLMs across 24 video tasks.
Improving Region Representation Learning from Urban Imagery with Noisy Long-Caption Supervision: This paper proposes UrbanLN, a framework that improves urban region representation learning from LLM-generated captions via a long-caption-aware positional encoding interpolation strategy and a dual-level (data and model) noise suppression mechanism.
FineXtrol: Controllable Motion Generation via Fine-Grained Text: This paper proposes FineXtrol, a framework that leverages temporally annotated, fine-grained body-part text descriptions as control signals. By combining a dual-branch ControlNet architecture with hierarchical contrastive learning to enhance the discriminability of the text encoder, FineXtrol achieves efficient, user-friendly, and precise controllable human motion generation, significantly outperforming existing methods on multi-body-part control benchmarks on HumanML3D.
Towards LLM-Empowered Knowledge Tracing via LLM-Student Hierarchical Behavior Alignment in Hyperbolic Space: This paper proposes L-HAKT, a framework that for the first time integrates LLM dual-agent design with hyperbolic geometry for knowledge tracing. A Teacher Agent parses exercise semantics and constructs a hierarchical knowledge graph, while a Student Agent simulates individual learning behaviors to generate synthetic interaction data. Hyperbolic contrastive learning is employed to calibrate the distributional gap between synthetic and real data. L-HAKT achieves an AUC of up to 80.29% across four educational datasets, with an AUC improvement of 13.03% over the GKT baseline on EdNet.
GOAL: Geometrically Optimal Alignment for Continual Generalized Category Discovery: Grounded in Neural Collapse theory, this paper replaces dynamic classifiers with a fixed Equiangular Tight Frame (ETF) classifier and achieves continual generalized category discovery via supervised alignment and confidence-guided unsupervised alignment, reducing forgetting by 16.1% and improving novel category discovery by 3.2% across four benchmarks.
Robust Tabular Foundation Models: This paper proposes RTFM — a model-agnostic adversarial training framework that performs min-max optimization over the parameter space of a synthetic data generator, maximizing the "optimality gap" between a tabular foundation model (TFM) and classical tree-based models. Using fewer than 100,000 additional synthetic datasets, RTFM significantly improves TabPFN V2 across multiple tabular benchmarks.
BCE3S: Binary Cross-Entropy Based Tripartite Synergistic Learning for Long-tailed Recognition: BCE3S is proposed, a binary cross-entropy (BCE)-based tripartite synergistic learning framework that integrates BCE-based joint learning, BCE-based contrastive learning, and BCE-based classifier uniformity learning. By decoupling per-class logits via Sigmoid, it suppresses the imbalance effects inherent to long-tailed distributions, achieving state-of-the-art performance on CIFAR10/100-LT, ImageNet-LT, and iNaturalist2018.
CATFormer: When Continual Learning Meets Spiking Transformers With Dynamic Thresholds: This paper proposes CATFormer, a data-replay-free continual learning framework built upon a spiking Vision Transformer, which achieves task-specific neuronal excitability modulation via context-adaptive dynamic firing thresholds. Over sequences of up to 100 tasks, the model not only avoids forgetting but actually improves in accuracy — a phenomenon the authors term "reverse forgetting."
Explanation-Preserving Augmentation for Semi-Supervised Graph Representation Learning: This paper proposes EPA-GRL (Explanation-Preserving Augmentation for Graph Representation Learning), which employs a GNN explainer trained with a small number of labels to identify semantic subgraphs (explanation subgraphs). During augmentation, only the non-semantic portions (marginal subgraphs) are perturbed, achieving semantics-preserving graph augmentation. EPA-GRL significantly outperforms semantics-agnostic random augmentation methods across 6 benchmarks.

Browse all 16 Self-Supervised Learning papers →

📐 Optimization & Theory (21)¶

Pareto-Grid-Guided Large Language Models for Fast and High-Quality Heuristics Design in Multi-Objective Combinatorial Optimization: This paper proposes MPaGE, a framework that integrates LLMs with a Pareto Front Grid mechanism and semantic clustering to automatically generate heuristics for multi-objective combinatorial optimization problems (MOCOPs), jointly optimizing solution quality and runtime efficiency. MPaGE achieves significant improvements in HV and IGD over baselines such as EoH and MEoH on Bi-TSP, Tri-TSP, Bi-CVRP, and Bi-KP benchmarks.
Instance Generation for Meta-Black-Box Optimization through Latent Space Reverse Engineering: This paper proposes the LSRE framework, which constructs a two-dimensional latent instance space for BBO problems via an autoencoder, and employs enhanced genetic programming to reverse-engineer diverse synthetic optimization problem instances from this space, forming the Diverse-BBO benchmark that substantially improves the generalization performance of MetaBBO methods.
Beyond the Mean: Fisher-Orthogonal Projection for Natural Gradient Descent in Large Batch Training: This paper proposes Fisher-Orthogonal Projection (FOP), which supplements variance information by orthogonally projecting sub-batch gradient differences under the Fisher metric, enabling the second-order optimizer KFAC to remain effective in ultra-large batch training and achieving up to ×7.5 speedup.
Bridging Synthetic and Real Routing Problems via LLM-Guided Instance Generation and Progressive Adaptation: This paper proposes EvoReal, a framework that employs LLM-driven evolutionary search to generate synthetic VRP instances structurally aligned with real-world distributions, and then adapts pretrained neural solvers to real benchmarks via a two-stage progressive fine-tuning strategy. EvoReal substantially outperforms existing neural solvers on TSPLib (1.05% gap) and CVRPLib (2.71% gap).
Efficient and Reliable Hitting-Set Computations for the Implicit Hitting Set Approach: To address numerical instability arising from the reliance on commercial IP solvers in the hitting-set component of the implicit hitting set (IHS) framework, this paper proposes alternative approaches based on pseudo-Boolean (PB) reasoning and stochastic local search (SLS), as well as hybrid strategies. The work realizes the first certifiable IHS computation and demonstrates effective trade-offs between efficiency and reliability across 1,786 benchmark instances.
MOTIF: Multi-strategy Optimization via Turn-based Interactive Framework: This paper proposes MOTIF, a framework that models solver design as a multi-strategy optimization problem. Through a turn-based competitive mechanism involving two LLM agents guided by Monte Carlo Tree Search (MCTS), MOTIF jointly optimizes multiple interdependent algorithmic components within combinatorial optimization solvers, consistently outperforming existing methods across TSP, CVRP, BPP, and other combinatorial optimization domains.
Data Heterogeneity and Forgotten Labels in Split Federated Learning: This paper systematically investigates catastrophic forgetting (CF) caused by data heterogeneity in Split Federated Learning — with particular focus on intra-round forgetting induced by the server-side processing order — and proposes Hydra, a multi-head method that partitions and trains the final layers of part-2 in groups before aggregation, significantly reducing the performance gap (PG) across labels by up to 75.4%.
Cost-Minimized Label-Flipping Poisoning Attack to LLM Alignment: This work provides the first theoretical analysis of the minimum cost required to steer an LLM policy toward an attacker's target by flipping preference labels during RLHF/DPO alignment. The problem is formalized as a convex optimization problem, upper and lower bounds on the cost are derived, and a post-processing method called PCM (Poisoning Cost Minimization) is proposed to substantially reduce the number of label flips while preserving the poisoning effect.
GHOST: Solving the Traveling Salesman Problem on Graphs of Convex Sets: This paper proposes GHOST, a hierarchical optimal search algorithm for solving the Traveling Salesman Problem on Graphs of Convex Sets (GCS). By combining combinatorial path search with convex trajectory optimization, and employing a novel abstract path expansion algorithm to compute admissible lower bounds that guide best-first search, GHOST achieves optimality guarantees while outperforming the monolithic mixed-integer convex programming (MICP) baseline by orders of magnitude in runtime.
ECPv2: Fast, Efficient, and Scalable Global Optimization of Lipschitz Functions: This paper proposes ECPv2, which introduces three innovations—adaptive lower bound, Worst-$m$ memory, and fixed random projection—to reduce the per-run complexity of Lipschitz global optimization from $\Omega(n^2 d)$ to $\Omega(n(m+d)\log n)$, while maintaining an $O(n^{-1/d})$ regret convergence rate that matches the minimax lower bound.

Browse all 21 Optimization & Theory papers →

📐 Learning Theory (3)¶

Streaming Generated Gaussian Process Experts for Online Learning and Control: Extended Version: This paper proposes SkyGP (Streaming Kernel-induced Progressively Generated Expert GP), which handles streaming data via kernel-distance-driven progressive expert generation and time-aware configurable aggregation, inheriting the learning guarantees of exact GP while maintaining bounded computational complexity. SkyGP comprehensively outperforms state-of-the-art methods on both benchmark regression tasks and real-time control experiments.
Generalizing Analogical Inference from Boolean to Continuous Domains: This paper revisits the theoretical foundations of analogical inference: it first constructs a counterexample demonstrating the failure of classical generalization bounds in the Boolean domain, then proposes a unified analogical inference framework based on parameterized generalized means, extending discrete classification to continuous regression domains.
A Switching Framework for Online Interval Scheduling with Predictions: For the irrevocable online interval scheduling problem, this paper proposes the SemiTrust-and-Switch framework and the SmoothMerge randomized algorithm. By switching between or blending a prediction-trusting strategy and a classical greedy algorithm, the approach achieves near-optimal performance when predictions are accurate (consistency) and degrades gracefully when predictions are erroneous (robustness and smoothness). Tightness of the framework on specific instances is also established.

🔗 Causal Inference (7)¶

I-CAM-UV: Integrating Causal Graphs over Non-Identical Variable Sets Using Causal Additive Models with Unobserved Variables: This paper proposes I-CAM-UV, a method that enumerates consistent DAGs satisfying structural constraints derived from multiple CAM-UV outputs over non-identical variable sets, recovering causal relations lost due to unobserved variables, and introduces an optimal-first search algorithm exploiting cost monotonicity for efficient combinatorial search.
KTCF: Actionable Recourse in Knowledge Tracing via Counterfactual Explanations for Education: This paper proposes KTCF, a counterfactual explanation generation method for Knowledge Tracing (KT) that leverages inter-concept relationships to produce sparse and actionable counterfactuals, subsequently post-processed into sequentially ordered instructional recommendations. KTCF comprehensively outperforms baseline methods across validity, sparsity, and actionability metrics.
Causal Inference Under Threshold Manipulation: Bayesian Mixture Modeling and Heterogeneous Treatment Effects: This paper proposes the BMTM/HBMTM Bayesian mixture model framework. In scenarios where consumers strategically manipulate spending to reach reward thresholds, the framework decomposes the observed distribution into bunching and non-bunching sub-distributions to accurately estimate threshold causal effects and heterogeneous treatment effects across subgroups.
Learning Subgroups with Maximum Treatment Effects without Causal Heuristics: Under the SCM framework, the paper proves that the subgroup with maximum treatment effect must exhibit homogeneous pointwise effects (Theorem 1); under the partition model assumption, it proves that optimal subgroup discovery reduces to standard supervised learning (Theorem 2), achievable via CART with the Gini index. On 77 ACIC-2016 semi-synthetic datasets, the proposed method achieves a mean treatment effect of 10.54 (vs. 7.84 for the runner-up), ranking first on 51.9% of datasets.
CaDyT: Causal Structure Learning for Dynamical Systems with Theoretical Score Analysis: This paper proposes CaDyT, which combines Gaussian process-based continuous-time dynamics modeling (via Adams-Bashforth integrators for exact inference) with the Minimum Description Length (MDL) principle for structure search. The method simultaneously addresses irregular sampling and causal structure identification, substantially outperforming all baselines on double-mass spring, diamond graph, and Rössler oscillator benchmarks (AUPRC 0.79 vs. runner-up 0.39).
From Theory of Mind to Theory of Environment: Counterfactual Simulation of Latent Environmental Dynamics: This paper proposes the concept of "Theory of Environment" (ToE), arguing that humans may infer latent environmental dynamics through computational mechanisms shared with Theory of Mind (ToM), thereby expanding the dimensionality of motor exploration and facilitating behavioral innovation.
Sparse Additive Model Pruning for Order-Based Causal Structure Learning: This paper proposes SARTRE, a framework that employs randomized tree embeddings and group-sparse regression to learn sparse additive models, replacing the hypothesis-testing-based redundant edge pruning in CAM-pruning for order-based causal structure learning. SARTRE achieves significant speedups without sacrificing accuracy.

🔬 Interpretability (37)¶

Distribution-Based Feature Attribution for Explaining the Predictions of Any Classifier: This paper proposes DFAX, the first distribution-based feature attribution method, which quantifies feature importance by comparing the conditional probability density of a target instance under the target class versus non-target classes. It provides the first formal definition of feature attribution, and demonstrates significant improvements over SHAP/LIME and other baselines across 10 datasets while being orders of magnitude faster.
LLM Circuit Analyses Are Consistent Across Training and Scale: This paper presents the first systematic tracking of internal circuits in decoder-only LLMs across 300 billion tokens of training and model scales ranging from 70M to 2.8B parameters. It finds that while specific attention heads may be replaced over the course of training, the underlying algorithms remain stable and consistent across scales, suggesting that circuit analyses conducted on smaller models generalize to larger models and longer training runs.
SOM Directions are Better than One: Multi-Directional Refusal Suppression in Language Models: This paper demonstrates that refusal behavior in LLMs is not encoded by a single direction but rather forms a low-dimensional manifold. It employs self-organizing maps (SOM) to extract multiple refusal directions and applies Bayesian optimization to search for the optimal ablation combination, surpassing single-direction baselines and dedicated jailbreak algorithms across multiple models.
Partially Shared Concept Bottleneck Models: This paper proposes PS-CBM, a framework that integrates multimodal concept generation (combining LLM semantics with visual cues from exemplar images), a partially shared concept strategy (merging concepts based on activation patterns), and a Concept-Efficient Accuracy (CEA) evaluation metric. PS-CBM achieves higher classification accuracy and interpretability with fewer concepts across 11 datasets.
SparK: Query-Aware Unstructured Sparsity with Recoverable KV Cache Channel Pruning: This paper proposes SparK — a training-free, token-wise unstructured channel pruning method for KV cache. It selects salient channels via query-aware saliency scoring and recovers the contribution of pruned channels through a recovery mechanism. At an 80% pruning ratio, performance degradation remains below 5%. The method is orthogonal to token eviction approaches and can reduce KV cache storage by an additional 30%+.
SparseRM: A Lightweight Preference Modeling with Sparse Autoencoder: SparseRM leverages sparse autoencoders (SAE) to extract preference-relevant directions from LLM intermediate representations, constructing a lightweight reward model via projection vectors. With fewer than 1% trainable parameters, it surpasses most mainstream reward models and demonstrates stronger generalization in online iterative alignment frameworks.
Share Your Attention: Transformer Weight Sharing via Matrix-Based Dictionary Learning: Inspired by dictionary learning, this paper proposes the MASA framework, which decomposes the attention projection matrices (Q/K/V/O) across Transformer layers into linear combinations of shared matrix atoms, achieving performance on par with or superior to the original Transformer at a 66.7% attention parameter compression ratio.
Adaptive Evidential Learning for Temporal-Semantic Robustness in Moment Retrieval: This paper proposes DEMR, a framework that introduces Deep Evidential Regression (DER) into video moment retrieval. It mitigates modal imbalance via a Reflective Flipped Fusion (RFF) module and corrects the counter-intuitive uncertainty estimation bias in vanilla DER via a Geom-regularizer, achieving significant improvements on both standard and debiased benchmarks.
DR.Experts: Differential Refinement of Distortion-Aware Experts for Blind Image Quality Assessment: This paper proposes the DR.Experts framework, which leverages DA-CLIP to obtain distortion-type priors, employs a Distortion Saliency Differential Module (DSDM) to disentangle distortion attention from semantic attention and thereby purify distortion features, and then applies a Dynamic Distortion Weighting Module (DDWM) to adaptively weight each distortion type's features according to its perceptual impact. The method achieves state-of-the-art performance on five BIQA benchmarks.
Hypothesis Generation via LLM-Automated Language Bias for ILP: This paper proposes the first end-to-end framework in which a multi-agent LLM system (Actor/Critic) automatically constructs ILP language bias (predicate system, type declarations, and mode constraints) from raw text. A Translator agent converts text into Prolog facts, and the MAXSYNTH solver then induces a globally optimal rule set based on the MDL principle. The framework achieves 88.3% and 81.3% accuracy on the SHOES and ZENDO tasks, respectively, with variance below 5% across four LLMs.

Browse all 37 Interpretability papers →

📦 Model Compression (60)¶

Efficient Reasoning for Large Reasoning Language Models via Certainty-Guided Reflection Suppression: This paper proposes CGRS (Certainty-Guided Reflection Suppression), a training-free efficient reasoning method that dynamically suppresses reflection trigger tokens (e.g., "Wait", "But") when the model exhibits high confidence, reducing token consumption of large reasoning language models by 18.5%–41.9% while maintaining reasoning accuracy.
Put the Space of LoRA Initialization to the Extreme to Preserve Pre-trained Knowledge: This paper proposes LoRA-Null, which initializes LoRA within the null space of pre-trained input activations (rather than the null space of weights). From an information-theoretic perspective, the effective rank of activations is much lower than that of weights, meaning their null space encodes less pre-trained knowledge, thereby substantially mitigating catastrophic forgetting during fine-tuning.
Don't Start Over: A Cost-Effective Framework for Migrating Personalized Prompts Between LLMs: This paper proposes PUMA, a framework that leverages lightweight adapters and a grouped user selection strategy to efficiently migrate personalized soft prompts from a source LLM to a target LLM with a different architecture. PUMA matches or surpasses from-scratch training on three large-scale datasets while reducing computational cost by up to 98%.
InfoCom: Kilobyte-Scale Communication-Efficient Collaborative Perception with Information-Aware Feature Compression: This paper proposes InfoCom, a framework that applies an extended information bottleneck (IB) principle to compress the communication payload of collaborative perception from the MB scale to the KB scale—a 440× reduction compared to Where2comm—while maintaining near-lossless perception performance. The framework consists of three core modules: information-aware encoding, sparse mask generation, and multi-scale decoding.
Distilling Cross-Modal Knowledge via Feature Disentanglement: This paper proposes Frequency-Decoupled Cross-Modal Knowledge Distillation (FD-CMKD), which decomposes teacher and student features into low-frequency (modality-shared semantics) and high-frequency (modality-specific details) components via Fourier transform, applies strong-consistency MSE and weak-consistency logMSE losses respectively, and introduces scale normalization along with shared classifier alignment to bridge the feature space. FD-CMKD consistently outperforms existing distillation methods across multiple cross-modal scenarios including audio–visual, image–text, and semantic segmentation.
First-Order Error Matters: Accurate Compensation for Quantized Large Language Models: This paper identifies a critical yet overlooked issue in LLM post-training quantization: the column-wise compensation process renders first-order gradient terms non-negligible. The proposed FOEM method incorporates first-order terms into the error compensation formula, reducing the perplexity of Llama3-8B under 3-bit quantization by 17.3% with virtually no additional computational overhead.
SpecQuant: Spectral Decomposition and Adaptive Truncation for Ultra-Low-Bit LLMs Quantization: SpecQuant proposes a two-stage quantization framework based on adaptive Fourier-domain decomposition: it first smoothly migrates activation outliers into weights, then suppresses high-frequency noise in the weights via channel-wise low-frequency Fourier truncation. On LLaMA-3 8B, W4A4 quantization achieves only 1.5% accuracy degradation, while delivering 2× speedup and 3× memory savings.
Earth-Adapter: Bridge Geospatial Domain Gaps with Mixture of Frequency Adaptation: This paper proposes Earth-Adapter, the first parameter-efficient fine-tuning (PEFT) method specifically designed to address artifact problems in remote sensing imagery. Through a frequency-guided Mixture of Adapters (MoA), features are decomposed into high- and low-frequency subspaces, independently optimized, and then dynamically aggregated. The method outperforms the baseline Rein across three settings: remote sensing semantic segmentation (SS), domain adaptation (DA), and domain generalization (DG).
Failures to Surface Harmful Contents in Video Large Language Models: This paper presents the first systematic security analysis of VideoLLMs, identifying three structural design flaws — sparse temporal sampling, spatial token downsampling, and modality fusion imbalance — that cause clearly visible harmful content in videos to be omitted from model-generated textual summaries (omission rate exceeding 90%). Three zero-query black-box attacks are designed to empirically validate the severity of these vulnerabilities.
XLinear: A Lightweight and Accurate MLP-Based Model for Long-Term Time Series Forecasting with Exogenous Inputs: This paper proposes XLinear, a lightweight time series forecasting model based on MLP with sigmoid gating. Through a global token mechanism, it efficiently integrates endogenous and exogenous variable information, achieving an optimal accuracy–efficiency trade-off across 12 datasets.

Browse all 60 Model Compression papers →

🕸️ Graph Learning (37)¶

S-DAG: A Subject-Based Directed Acyclic Graph for Multi-Agent Heterogeneous Reasoning: This paper proposes S-DAG, which uses a GNN to identify relevant subjects and their dependencies from a given question, constructing a directed acyclic graph. Subject nodes are matched to the most capable expert LLMs (14 domain-specific models of 7–13B parameters), and collaborative reasoning proceeds in DAG topological order (supporting subjects → dominant subject). The resulting small-model pool surpasses GPT-4o-mini (59.73 vs. 58.52) and approaches the performance of a 72B model.
Logical Characterizations of GNNs with Mean Aggregation: This paper provides a systematic logical characterization of GNNs using mean aggregation: under the non-uniform setting, mean-GNNs are equivalent to Ratio Modal Logic (RML); under the uniform setting (relative to MSO), they are equivalent to Modal Logic (ML); when the combination function is additionally required to be continuous and the classification function is a threshold, the expressiveness drops significantly to Alternation-Free Modal Logic (AFML).
Relink: Constructing Query-Driven Evidence Graph On-the-Fly for GraphRAG: This paper proposes a paradigm shift in GraphRAG from "build-then-reason" to "reason-and-construct," introducing the Relink framework that dynamically constructs query-specific evidence graphs—combining a high-precision KG backbone with a high-recall latent relation pool, unified via a query-driven ranker to assess relevance on demand, complete missing paths, and filter distractor facts—achieving average gains of 5.4% EM and 5.2% F1 across 5 multi-hop QA benchmarks.
Beyond Fact Retrieval: Episodic Memory for RAG with Generative Semantic Workspaces: This paper proposes the Generative Semantic Workspace (GSW), a neuroscience-inspired generative memory framework that constructs structured episodic memory representations for LLMs, achieving an F1 of 0.85 on EpBench while reducing query-time context tokens by 51%.
Beyond Fixed Depth: Adaptive Graph Neural Networks for Node Classification Under Varying Homophily: This paper proposes AD-GNN, which theoretically analyzes node-level homophily/heterophily characteristics and adaptively assigns different aggregation depths to individual nodes, enabling unified handling of node classification on both homophilic and heterophilic graphs within a single framework.
MyGram: Modality-aware Graph Transformer with Global Distribution for Multi-modal Entity Alignment: This paper proposes MyGram, which captures deep structural contextual information within each modality via a Modality-aware Graph Diffusion (MGD) module, and introduces a global distribution alignment loss (Gram Loss) based on the determinant of the Gram matrix to enforce cross-modal semantic consistency in high-dimensional space, achieving more robust multi-modal entity alignment.
Are Graph Transformers Necessary? Efficient Long-Range Message Passing with Fractal Nodes in MPNNs: This paper proposes Fractal Nodes (FN) to enhance long-range message passing in MPNNs. Subgraph-level aggregation nodes are generated via METIS graph partitioning, combined with low-pass and high-pass filters (LPF+HPF) and a learnable frequency parameter $\omega$. MLP-Mixer is adopted for cross-subgraph communication. The approach achieves $O(L(|V|+|E|))$ linear complexity while matching or surpassing Graph Transformer performance, earning an AAAI Oral.
MUG: Meta-path-aware Universal Heterogeneous Graph Pre-Training: MUG is the first universal heterogeneous graph pre-training method that requires no LLM. It unifies heterogeneous node/relation types via contextual structural encoding, aligns representation spaces across graphs with a dimension-aware encoder, and achieves transferable encoding and aggregation through a shared GNN encoder over meta-path views combined with global scatter regularization. MUG substantially outperforms existing methods on cross-domain and few-shot node classification.
UniHR: Hierarchical Representation Learning for Unified Knowledge Graph Link Prediction: This paper proposes UniHR, a unified framework that converts hyper-relational, temporal, and nested KGs into a triple-based representation via Hierarchical Data Representation (HiDR), and designs a Hierarchical Structure Learning (HiSL) module for two-stage intra-fact and inter-fact message passing. UniHR achieves state-of-the-art or competitive link prediction results across 9 datasets spanning 5 KG types.
Adaptive Initial Residual Connections for GNNs with Theoretical Guarantees: This paper proposes Adaptive Initial Residual Connections (Adaptive IRC), which allows each node to have a personalized residual strength learned from its initial features. It provides the first theoretical proof of a positive lower bound on the Dirichlet energy of initial residual connections with activation functions (guaranteeing the absence of over-smoothing), and introduces a PageRank-based heuristic variant that achieves comparable or superior performance without learning additional parameters.

Browse all 37 Graph Learning papers →

📈 Time Series (31)¶

Mitigating Error Accumulation in Co-Speech Motion Generation via Global Rotation Diffusion and Multi-Level Constraints: This paper proposes GlobalDiff, a framework that, for the first time, performs diffusion-based generation in the global joint rotation space, fundamentally eliminating error accumulation in hierarchical forward kinematics. A three-level joint–bone–motion constraint scheme compensates for the structural priors lost under global representation. GlobalDiff achieves state-of-the-art performance on multi-speaker co-speech motion generation benchmarks, improving FGD by 46% over the previous best method.
Towards Non-Stationary Time Series Forecasting with Temporal Stabilization and Frequency Differencing: This paper proposes DTAF, a dual-branch framework that extracts and removes heterogeneous non-stationary patterns via a non-stationary MoE filter in the temporal domain, tracks frequency drift via spectral differencing in the frequency domain, and fuses complementary information from both domains through dual-branch attention for robust non-stationary time series forecasting.
Detecting the Future: All-at-Once Event Sequence Forecasting with Horizon Matching: This paper proposes DEF (Detection-based Event Forecasting), which draws on the set-matching idea from DETR in object detection and employs the Hungarian algorithm to align predicted and ground-truth event sequences, achieving high-accuracy and high-diversity long-horizon event forecasting with state-of-the-art results on five datasets.
Mask the Redundancy: Evolving Masking Representation Learning for Multivariate Time-Series Clustering: This paper proposes EMTC, a framework that dynamically masks redundant timestamps via Importance-aware Variate-wise Masking (IVM), combined with Multi-Endogenous Views (MEV) generation and cluster-guided contrastive learning, achieving an average F1 improvement of 4.85% across 15 MTS clustering benchmarks.
Revitalizing Canonical Pre-Alignment for Irregular Multivariate Time Series Forecasting: This paper is the first to argue that Canonical Pre-Alignment (CPA) should not be abandoned for Irregular Multivariate Time Series (IMTS) forecasting. It proposes KAFNet, which addresses the efficiency bottleneck of CPA via three modules—Pre-Convolution smoothing, Temporal Kernel Aggregation (TKA), and Frequency-domain Linear Attention (FLA)—achieving state-of-the-art accuracy on 4 IMTS benchmarks while reducing parameters by 7.2× and accelerating training/inference by 8.4×.
A Unified Shape-Aware Foundation Model for Time Series Classification: This paper proposes UniShape — a foundation model for time series classification that adaptively aggregates multi-scale discriminative subsequences (shapelets) via a shape-aware adapter, and learns transferable shapelet representations at both instance and shape levels through prototype-based contrastive pretraining. With only 3.1M parameters, UniShape achieves state-of-the-art performance on 128 UCR datasets (average accuracy 87.08%) while providing strong classification interpretability.
C3RL: Rethinking the Combination of Channel-independence and Channel-mixing from Representation Learning: This paper proposes C3RL, a SimSiam-based contrastive learning framework that treats channel-independence (CI) and channel-mixing (CM) strategies as two transposed views of the same data to construct positive pairs. By jointly optimizing representation learning and forecasting through a Siamese network, C3RL improves the best-performance rate of CI models from 43.6% to 81.4% and CM models from 23.8% to 76.3%.
DeepBooTS: Dual-Stream Residual Boosting for Drift-Resilient Time-Series Forecasting: This paper proposes DeepBooTS, which leverages bias-variance decomposition theory to demonstrate that weighted ensembling reduces variance and thereby mitigates concept drift. The method introduces a dual-stream residual-decreasing boosting architecture in which each block corrects the residual of the preceding block, achieving an average improvement of 15.8% across multiple datasets.
Scaling LLM Speculative Decoding: Non-Autoregressive Forecasting in Large-Batch Scenarios: This paper proposes SpecFormer, a non-autoregressive draft model architecture that integrates unidirectional and bidirectional attention. By reducing reliance on large prefix trees and minimizing position-dependent parameters, SpecFormer achieves consistent LLM inference acceleration in large-batch scenarios.
Sonnet: Spectral Operator Neural Network for Multivariable Time Series Forecasting: This paper proposes Sonnet, which maps inputs to the time-frequency domain via learnable wavelet transforms, introduces multivariate coherence-based attention (MVCA) to model inter-variable dependencies, and employs a Koopman operator for stable temporal evolution forecasting. Sonnet achieves state-of-the-art performance on 34 out of 47 forecasting tasks, reducing average MAE by 2.2%.

Browse all 31 Time Series papers →

🏥 Medical Imaging (76)¶

Towards Ultrasound-based Reliable Disease Diagnosis Using Causal Inference: Given a B-mode ultrasound image and a contrast-enhanced ultrasound video from the same case, the method uses reference feature dictionaries and causal-adjustment-inspired attention when two network branches exchange morphology and perfusion information, aiming to reduce reliance on incidental associations and reporting 89.80% benign/malignant classification accuracy on its dataset of 486 breast cases.
MedEyes: Learning Dynamic Visual Focus for Medical Progressive Diagnosis: MedEyes is a hybrid-policy reinforcement learning framework that introduces a Gaze-guided Reasoning Navigator (GRN) to simulate the "scan-and-drill" visual search pattern of clinical physicians. Combined with a Confidence Value Sampler (CVS) and dual-stream GRPO optimization, the framework enables dynamic visual focus for progressive medical diagnostic reasoning, achieving an average improvement of 8.5 pp across five medical VQA benchmarks.
MAISI-v2: Accelerated 3D High-Resolution Medical Image Synthesis with Rectified Flow and Region-specific Contrastive Loss: This paper presents MAISI-v2, the first framework to introduce Rectified Flow into 3D medical image synthesis. By replacing DDPM with Rectified Flow, it achieves a 33× speedup, and a novel region-specific contrastive loss is designed to improve conditioning fidelity for small regions such as tumors. The utility of synthesized data is validated on downstream tumor segmentation tasks.
Decoding with Structured Awareness: Integrating Directional, Frequency-Spatial, and Structural Attention for Medical Image Segmentation: This paper proposes a novel decoder framework for medical image segmentation comprising three modules: Adaptive Cross-Fusion Attention (ACFA) for directional awareness, Triple Feature Fusion Attention (TFFA) for spatial-frequency-wavelet fusion, and Structural-aware Multi-scale Masking Module (SMMM), achieving state-of-the-art performance across multiple benchmark datasets.
NeuroBridge: Bio-Inspired Self-Supervised EEG-to-Image Decoding via Cognitive Priors and Bidirectional Semantic Alignment: This paper proposes NeuroBridge, a framework that employs Cognitive Prior Augmentation (CPA, asymmetric augmentation to simulate perceptual variability) and a Shared Semantic Projector (SSP, bidirectional alignment into a unified semantic space). On the THINGS-EEG dataset under a 200-class zero-shot EEG-to-image retrieval task, the method achieves 63.2% Top-1 (+12.3%) and 89.9% Top-5 (+10.2%), substantially surpassing the existing state of the art.
PriorRG: Prior-Guided Contrastive Pre-training and Coarse-to-Fine Decoding for Chest X-ray Report Generation: PriorRG proposes a two-stage chest X-ray report generation framework that aligns clinical context with spatiotemporal visual features via prior-guided contrastive pre-training, then progressively integrates clinical context, disease progression, and multi-level visual cues through prior-aware coarse-to-fine decoding, achieving a 3.6% improvement in BLEU-4 and a 3.8% improvement in F1 on MIMIC-CXR.
DeepGB-TB: A Risk-Balanced Cross-Attention Gradient-Boosted Convolutional Network for Rapid, Interpretable Tuberculosis Screening: This paper proposes DeepGB-TB, a multimodal TB screening system combining a lightweight 1D-CNN (for cough audio) and gradient-boosted decision trees (for demographic features). A bidirectional cross-attention module (CM-BCA) fuses heterogeneous data by mimicking clinical reasoning, while a tuberculosis risk-balanced loss (TRBL) minimizes missed diagnoses. The system achieves AUROC 0.903 on a 7-country dataset and supports offline real-time inference on mobile devices.
FIA-Edit: Frequency-Interactive Attention for Efficient and High-Fidelity Inversion-Free Text-Guided Image Editing: This paper proposes FIA-Edit, an inversion-free text-guided image editing framework based on frequency-interactive attention. It introduces a Frequency Representation Interaction (FRI) module that performs frequency-domain fusion of source/target features within self-attention, and a Feature Injection (FIJ) module that explicitly incorporates source image features into cross-attention. The framework achieves precise semantic editing while maintaining high background fidelity, and for the first time applies a general image editing method to clinical surgical bleeding image augmentation.
Advancing Safe Mechanical Ventilation Using Offline RL With Hybrid Actions and Clinically Aligned Rewards: This paper addresses the problem of optimizing mechanical ventilation (MV) settings in the ICU via offline RL. A hybrid action space approach (HybridIQL/HybridEDAC) is proposed to avoid distributional shift caused by conventional discretization. Clinically aligned reward functions are introduced based on ventilator-free days (VFD) and physiological safety ranges, with multi-objective optimization used to select the optimal reward. The number of optimizable ventilation parameters is scaled from 2–3 to 6, and HybridIQL achieves the best balance between performance and policy coverage.
SEMC: Structure-Enhanced Mixture-of-Experts Contrastive Learning for Ultrasound Standard Plane Recognition: This paper proposes the SEMC framework, which aligns shallow structural cues with deep semantic representations via a Semantic-Structure Fusion Module (SSFM), and performs hierarchical contrastive learning over multi-level features through a Mixture-of-Experts Contrastive Recognition Module (MCRM), thereby enhancing fine-grained discriminability for ultrasound standard plane recognition. A new liver ultrasound dataset, LP2025, is also introduced.

Browse all 76 Medical Imaging papers →

🩺 Medical LLM (12)¶

BiCA: Effective Biomedical Dense Retrieval with Citation-Aware Hard Negatives: This paper proposes a hard negative mining method that constructs a multi-hop semantic graph from PubMed citation chains and performs random walks thereon. Using only 20k training samples and minimal fine-tuning steps, 33M/110M small models surpass retrieval baselines with billions of parameters on BEIR and LoTTE.
MIRAGE: Scaling Test-Time Inference with Parallel Graph-Retrieval-Augmented Reasoning Chains: This paper proposes MIRAGE, a framework that extends conventional linear reasoning chains into a parallel multi-chain reasoning paradigm. It combines adaptive retrieval from structured medical knowledge graphs (via neighborhood expansion and multi-hop traversal) with cross-chain verification to resolve contradictions, consistently outperforming GPT-4o, ToT, and Search-o1 on three medical QA benchmarks.
CliCARE: Grounding Large Language Models in Clinical Guidelines for Decision Support over Longitudinal Cancer Electronic Health Records: This paper proposes CliCARE, a framework that transforms unstructured longitudinal cancer EHRs into temporal knowledge graphs (TKGs), aligns them with clinical practice guideline (CPG) knowledge graphs, and provides evidence-grounded clinical decision support for LLMs. An LLM-as-a-Judge evaluation protocol highly correlated with expert assessments is also introduced.
Expert-Guided Prompting and Retrieval-Augmented Generation for Emergency Medical Service Question Answering: This paper constructs EMSQA, the first multiple-choice QA dataset for the emergency medical services domain (24.3K questions, 10 clinical topics, 4 certification levels), and proposes the Expert-CoT and ExpertRAG frameworks to inject domain expertise into LLM reasoning and retrieval, achieving up to 4.59% accuracy improvement over standard RAG.
A Principle-Driven Adaptive Policy for Group Cognitive Stimulation Dialogue for Elderly with Cognitive Impairment: This paper proposes the GCSD system for group Cognitive Stimulation Therapy (CST) targeting elderly individuals with cognitive impairment. The system integrates four modules — multi-speaker context control, dynamic participant state modeling (soft prompt), a cognitive stimulation attention loss, and a multi-dimensional reward policy optimization — built on a fine-tuned Qwen-2.5-3B backbone. Training is conducted on 500+ hours of real Cantonese CST dialogues and 10,000+ simulated conversations. The system achieves a BLEU-4 of 27.93, surpassing GPT-4o and other large models, with an A/B test win rate of 50% versus GPT-4o's 39%.
LUCID: Learning-Enabled Uncertainty-Aware Certification of Stochastic Dynamical Systems: This paper proposes LUCID, the first verification engine capable of providing quantified safety guarantees for black-box stochastic dynamical systems. By combining data-driven control barrier certificates, conditional mean embeddings, and finite Fourier kernel expansions, LUCID reformulates a semi-infinite non-convex optimization problem into a tractable linear program.
ShortageSim: Simulating Drug Shortages under Information Asymmetry: This paper proposes ShortageSim, the first LLM-based multi-agent simulation framework for drug shortages. It models strategic decision-making among FDA regulators, manufacturers, and buyers under information asymmetry, achieving an 84% improvement in predicting resolution lag time on historical shortage data, and provides a controlled testbed for evaluating regulatory strategies.
Learning Cell-Aware Hierarchical Multi-Modal Representations for Robust Molecular Modeling: This paper proposes the CHMR framework, which addresses missing biological modalities via structure-aware propagation, and introduces Tree-VQ to model hierarchical dependencies among molecules, cells, and genes. Evaluated on 728 tasks across 9 benchmarks, CHMR achieves a 3.6% improvement in classification and 17.2% in regression, enabling robust cell-aware molecular representation learning.
Voices, Faces, and Feelings: Multi-modal Emotion-Cognition Captioning for Mental Health Understanding: This paper proposes the Emotion-Cognition cooperative Multi-modal Captioning (ECMC) task and framework. A dual-stream BridgeNet extracts emotion and cognition features from video, audio, and text, and a LLaMA decoder generates natural language descriptions. The system provides interpretable emotion-cognition profiles for mental health assessment, substantially improving both diagnostic accuracy and explainability.
Measuring Stability Beyond Accuracy in Small Open-Source Medical Large Language Models for Pediatric Endocrinology: This paper systematically evaluates six small open-source medical LLMs (<10B parameters) in pediatric endocrinology, demonstrating that accuracy alone is insufficient to characterize model reliability: semantically neutral prompt variations lead to significant output shifts (Stuart-Maxwell $p < 10^{-4}$), high consistency does not imply correctness, and even differences in CUDA versions can induce statistically significant output distribution changes.

Browse all 12 Medical LLM papers →

🧬 Computational Biology (20)¶

CellStream: Dynamical Optimal Transport Informed Embeddings for Reconstructing Cellular Trajectories from Snapshots Data: This paper proposes CellStream, a deep learning framework that jointly learns an autoencoder and unbalanced dynamical optimal transport (OT) to simultaneously obtain low-dimensional embeddings and continuous cellular dynamics from discrete-time single-cell snapshot data, achieving significant improvements over existing methods in temporal consistency and velocity consistency.
MergeDNA: Context-aware Genome Modeling with Dynamic Tokenization through Token Merging: MergeDNA is proposed to achieve context-aware dynamic DNA tokenization via differentiable Token Merging, combined with a hierarchical autoencoder and adaptive masked token modeling for pretraining. With 380M parameters, it surpasses GENERator at 1.3B.
ProtSAE: Disentangling and Interpreting Protein Language Models via Semantically-Guided Sparse Autoencoders: This paper proposes ProtSAE, which incorporates semantic annotations and domain ontology knowledge as guidance signals during sparse autoencoder training to address the semantic entanglement problem of conventional SAEs. The method aligns latent features of protein language models with biological concepts (molecular function, biological process, ion binding sites, etc.) with high precision, while maintaining high reconstruction fidelity and supporting concept-level generation steering.
TrinityDNA: A Bio-Inspired Foundational Model for Efficient Long-Sequence DNA Modeling: TrinityDNA is a bio-inspired DNA foundation model integrating three innovations: a Groove Fusion module for capturing major/minor groove structural features, a Gated Reverse Complement mechanism for handling double-strand complementary symmetry, and Sliding Multi-Window Attention for multi-scale long-range dependency modeling. Combined with an Evolutionary Training Strategy (ETS) progressing from prokaryotes to eukaryotes, TrinityDNA achieves an average MCC of 0.708 across 15 GUE benchmark tasks (surpassing NT with 2.5B parameters), leads on both prokaryotic and eukaryotic zero-shot tasks across 19 benchmarks, and introduces a new CDS annotation benchmark for long-sequence inference evaluation.
ConSurv: Multimodal Continual Learning for Survival Analysis: This paper proposes ConSurv, the first multimodal continual learning framework for survival analysis. Through two core components — Multi-Stage Mixture-of-Experts (MS-MoE) and Feature-Constrained Replay (FCR) — ConSurv effectively mitigates catastrophic forgetting in settings that integrate whole slide pathology images and genomic data, comprehensively outperforming existing methods on the newly constructed MSAIL benchmark.
GP-MoLFormer-Sim: Test Time Molecular Optimization through Contextual Similarity Guidance: This paper proposes GP-MoLFormer-Sim, a training-free test-time molecular generation guidance method that leverages the contextual embeddings of a chemical language model (GP-MoLFormer) to estimate similarity to target molecules, dynamically adjusting logits during autoregressive decoding. Combined with a genetic algorithm (GP-MoLFormer-Sim+GA), the method achieves an average rank of 2nd across 23 tasks on the PMO benchmark and outperforms MOLLEO—which relies on GPT-4—under a strict black-box oracle setting.
S2Drug: Bridging Protein Sequence and 3D Structure in Contrastive Representation Learning for Virtual Screening: This paper proposes S2Drug, a two-stage contrastive learning framework. Stage 1 performs large-scale protein sequence–ligand contrastive pre-training on ChemBL with a bilateral data sampling strategy to reduce noise and redundancy. Stage 2 fine-tunes on PDBBind by fusing sequence and 3D structural information via a residue-level gating module and incorporating a binding site prediction auxiliary task. S2Drug substantially outperforms existing methods on the DUD-E and LIT-PCBA virtual screening benchmarks.
Dual-Path Knowledge-Augmented Contrastive Alignment Network for Spatially Resolved Transcriptomics: This paper proposes DKAN, a Dual-path Knowledge-Augmented contrastive Alignment Network that integrates semantic information from external gene databases as a cross-modal coordinator. Combined with a unified one-stage contrastive learning paradigm and an adaptive weighting mechanism, DKAN predicts spatially resolved gene expression from H&E-stained whole slide images (WSI), achieving state-of-the-art performance across three public ST datasets.
EPO: Diverse and Realistic Protein Ensemble Generation via Energy Preference Optimization: This paper proposes EPO (Energy Preference Optimization), which combines reverse SDE sampling with listwise energy-ranked preference optimization to align a pretrained protein generator with the target Boltzmann distribution using only energy signals. EPO achieves state-of-the-art performance across 9 metrics on three benchmarks (Tetrapeptides, ATLAS, and Fast-Folding), entirely eliminating the need for expensive molecular dynamics (MD) simulations.
HiFusion: Hierarchical Intra-Spot Alignment and Regional Context Fusion for Spatial Gene Expression Prediction from Histopathology: This paper proposes HiFusion, a framework comprising two complementary modules — Hierarchical Intra-Spot Modeling (HISM) and Context-Aware Cross-Scale Fusion (CCF) — to accurately predict spatial gene expression from H&E-stained whole-slide images, achieving state-of-the-art performance on two benchmark datasets under both 2D cross-validation and 3D sample-specific evaluation settings.

Browse all 20 Computational Biology papers →

⚛️ Physics & Scientific Computing (15)¶

PhysicsCorrect: A Training-Free Approach for Stable Neural PDE Simulations: This paper proposes PhysicsCorrect, a training-free correction framework that models PDE residual correction as a linearized inverse problem and precomputes a cached pseudoinverse. At inference time, it achieves up to 100× error reduction with less than 5% computational overhead, and is applicable to arbitrary pretrained neural operators including FNO, UNet, and ViT.
Phys-Liquid: A Physics-Informed Dataset for Estimating 3D Geometry and Volume of Transparent Deformable Liquids: This work introduces the Phys-Liquid dataset (97,200 physics-simulated images with 3D meshes), which models dynamic deformation of liquids inside transparent containers based on the Navier-Stokes equations, and proposes a four-stage reconstruction pipeline (segmentation → multi-view mask generation → 3D reconstruction → scaling) to achieve high-accuracy liquid geometry and volume estimation in both simulated and real-world scenes.
Towards a Foundation Model for Partial Differential Equations Across Physics Domains: This paper proposes PDE-FM, a modular PDE foundation model combining spatial-spectral dual-modal tokenization, FiLM-based physics modulation, and a Mamba state-space backbone. It achieves an average 46% reduction in VRMSE across 12 heterogeneous physics-domain datasets from The Well benchmark.
SAOT: An Enhanced Locality-Aware Spectral Transformer for Solving PDEs: This paper proposes SAOT (Spectral Attention Operator Transformer), which captures high-frequency local details via linear-complexity Wavelet Attention (WA) and complements it with the global receptive field of Fourier Attention (FA) through a gated fusion mechanism. SAOT achieves state-of-the-art performance on 6 operator learning benchmarks, reducing the relative error on Navier-Stokes by 22.3% compared to Transolver.
Knowledge-Guided Masked Autoencoder with Linear Spectral Mixing and Spectral-Angle-Aware Reconstruction: This paper proposes KARMA, a framework that embeds the Linear Spectral Mixing Model (LSMM) as a physics constraint within the ViT-MAE decoder, combined with a Spectral Angle Mapper (SAM) loss, to improve reconstruction fidelity and downstream transfer performance for hyperspectral remote sensing imagery.
SVD-NO: Learning PDE Solution Operators with SVD Integral Kernels: This paper proposes SVD-NO, a neural operator that explicitly parameterizes the SVD decomposition of integral kernels, achieving $O(ndL)$ linear computational complexity while maintaining high expressiveness, and attaining new state-of-the-art performance on 5 PDE benchmarks.
Catastrophic Forgetting in Kolmogorov-Arnold Networks: The first systematic study of catastrophic forgetting in Kolmogorov-Arnold Networks (KANs): establishes a theoretical framework linking forgetting to activation support overlap and intrinsic data dimensionality, and proposes KAN-LoRA for continual fine-tuning knowledge editing in language models.
FlashKAT: Understanding and Addressing Performance Bottlenecks in the Kolmogorov-Arnold Transformer: This paper provides an in-depth analysis of the root cause behind KAT (Kolmogorov-Arnold Transformer) training being 123× slower than ViT. The bottleneck is identified not as FLOPs but as memory stalls caused by gradient accumulation during backpropagation (global memory contention from atomic add operations). The proposed FlashKAT restructures GPU kernels to achieve an 86.5× training speedup and reduces gradient rounding errors by nearly an order of magnitude.
Fast 3D Surrogate Modeling for Data Center Thermal Management: This paper develops a vision-based 3D surrogate modeling framework for data centers. Server workloads, fan speeds, and air-conditioning temperature setpoints are encoded as 3D voxel representations, and architectures including 3D CNN U-Net, 3D Fourier Neural Operator, and 3D Vision Transformer are employed for real-time temperature field prediction. The proposed framework achieves inference speeds up to 20,000× faster than traditional CFD solvers while enabling a 7% reduction in energy consumption.
PIMRL: Physics-Informed Multi-Scale Recurrent Learning for Burst-Sampled Spatiotemporal Dynamics: This paper proposes PIMRL, a framework for learning from burst-sampled (short high-frequency bursts followed by long intervals) sparse spatiotemporal data. It features a dual-module architecture combining macro-scale latent-space reasoning and micro-scale physics correction, integrated via cross-scale message passing, achieving up to 80% error reduction across 5 PDE benchmarks.

Browse all 15 Physics & Scientific Computing papers →

🌍 Earth Science (2)¶

MdaIF: Robust One-Stop Multi-Degradation-Aware Image Fusion with Language-Driven Semantics: This paper proposes MdaIF, a framework that leverages a vision-language model (VLM) to extract degradation-aware semantic priors for guiding mixture-of-experts (MoE) routing and channel attention modulation, enabling one-stop infrared-visible image fusion across multiple degradation scenarios without requiring degradation-type annotations.
RENEW: Risk- and Energy-Aware Navigation in Dynamic Waterways: This paper proposes RENEW, a global path planner for autonomous surface vessels (ASVs) operating in dynamic water current (ocean current) environments. It introduces a unified risk- and energy-aware strategy via adaptive no-go zone identification, best-effort contingency planning, and a hierarchical architecture based on Constrained Delaunay Triangulation (CDT), achieving zero collisions in emergency maneuver tests.

📡 Signal & Communications (3)¶

Text-Guided Channel Perturbation and Pretrained Knowledge Integration for Unified Multi-Modality Image Fusion: This paper proposes UP-Fusion, a unified multi-modality image fusion framework comprising three modules — Semantic-aware Channel Pruning Module (SCPM), Geometric Affine Modulation (GAM), and CLIP Text-guided Channel Perturbation Module (TCPM) — that employs a single set of weights (trained solely on infrared-visible data) to simultaneously handle both IVIF and medical image fusion tasks, achieving state-of-the-art performance on both.
Balancing Multimodal Domain Generalization via Gradient Modulation and Projection: This paper proposes a Gradient Modulation Projection (GMP) strategy that addresses inter-modality optimization imbalance and inter-task gradient conflicts in multimodal domain generalization (MMDG) through two components: Inter-modality Gradient Decoupled Modulation (IGDM) and Conflict-Adaptive Gradient Projection (CAGP), achieving state-of-the-art performance on multiple benchmarks.
Task Aware Modulation Using Representation Learning for Upscaling of Terrestrial Carbon Fluxes: This paper proposes TAM-RL, a framework that formulates terrestrial carbon flux upscaling as a zero-shot regression transfer learning problem. By combining a BiLSTM task encoder with FiLM modulation and a knowledge-guided loss derived from the carbon balance equation, the method achieves a 9.6% reduction in GPP RMSE and a 43.8% improvement in NEE R² over FLUXCOM-X-BASE across 150+ flux tower sites.

👥 Social Computing (10)¶

Cross-modal Prompting for Balanced Incomplete Multi-modal Emotion Recognition: This paper proposes Cross-modal Prompting (ComP), which addresses the modality imbalance problem in incomplete multi-modal emotion recognition (IMER) via progressive prompt generation, cross-modal knowledge propagation, and a dynamic scheduler, achieving state-of-the-art performance across 4 datasets and 7 missing rates.
Fact2Fiction: Targeted Poisoning Attack to Agentic Fact-checking System: This paper proposes Fact2Fiction, the first poisoning attack framework targeting agentic fact-checking systems (e.g., DEFAME, InFact). It employs a Planner Agent to simulate claim decomposition and generate sub-questions, reverse-engineers key reasoning points from system justifications to craft targeted malicious evidence, and allocates the poisoning budget according to sub-claim importance. At a poisoning rate of only 1%, Fact2Fiction achieves 8.9%–21.2% higher attack success rate (ASR) than the state-of-the-art PoisonedRAG.
FactGuard: Event-Centric and Commonsense-Guided Fake News Detection: This paper proposes FactGuard, a framework that leverages LLMs to extract event-centric content (with style removed) and generate commonsense rationales. A Rationale Usability Evaluator dynamically assesses the reliability of LLM suggestions. Knowledge distillation yields a lightweight variant, FactGuard-D, that operates without LLM inference, achieving both robustness and efficiency in fake news detection.
Bias Association Discovery Framework for Open-Ended LLM Generations: This paper proposes the Bias Association Discovery Framework (BADF), which systematically extracts both known and unknown bias associations between demographic identities and descriptive concepts from LLM open-ended story generation, overcoming the limitation of prior methods that rely on predefined bias concepts.
Reasoning About the Unsaid: Misinformation Detection with Omission-Aware Graph Inference: This paper proposes OmiGraph, the first omission-aware misinformation detection framework. By constructing omission-aware graphs, leveraging LLMs to reason about omission intent, and employing omission-guided message passing and aggregation mechanisms, OmiGraph extracts deception patterns from "what is unsaid," achieving average gains of +5.4% F1 and +5.3% ACC on bilingual datasets.
SceneJailEval: A Scenario-Adaptive Multi-Dimensional Framework for Jailbreak Evaluation: This paper proposes SceneJailEval, a scenario-adaptive multi-dimensional jailbreak evaluation framework that defines 14 jailbreak scenarios and 10 evaluation dimensions. Through a pipeline of scenario classification → dynamic dimension selection → multi-dimensional detection → weighted harm scoring, it achieves F1 of 0.917 on a self-constructed dataset (surpassing SOTA by 6%) and 0.995 on JBB (surpassing SOTA by 3%), while supporting harm severity quantification beyond binary classification.
T2Agent: A Tool-augmented Multimodal Misinformation Detection Agent with Monte Carlo Tree Search: This paper proposes T2Agent, a misinformation detection agent integrating an extensible toolset with Monte Carlo Tree Search (MCTS). By decomposing detection into sub-tasks targeting distinct forgery sources via a multi-source verification mechanism, T2Agent achieves a new state of the art on MMfakebench, improving the accuracy of the baseline MMDAgent by 28.7% using GPT-4o as the backbone.
Argumentative Debates for Transparent Bias Detection: This paper proposes ABIDE (Argumentative BIas Detection by DEbate), which constructs Quantitative Bipolar Argumentation Frameworks (QBAFs) via neighborhood-based argument schemes, models the bias detection process as a structured debate, enables transparent bias reasoning from individual neighborhoods to the global level, and formally proves the correspondence between QBAF semantics and the expected behavior of bias detection.
From Imitation to Discrimination: Toward A Generalized Curriculum Advantage Mechanism Enhancing Cross-Domain Reasoning Tasks: This paper proposes CAPO (Curriculum Advantage Policy Optimization), an adaptive curriculum mechanism based on advantage signals. Through a two-phase strategy — first imitation (using only positive-advantage samples) then discrimination (introducing negative signals) — CAPO stably and significantly improves LLM performance on mathematical reasoning and multimodal GUI reasoning tasks.
Multi-modal Dynamic Proxy Learning for Personalized Multiple Clustering: This paper proposes the Multi-DProxy framework, which leverages learnable textual proxies for personalized multiple clustering through three key innovations: gated cross-modal fusion, dual-constraint proxy optimization, and dynamic candidate management, achieving state-of-the-art performance on all public benchmarks.

🛡️ AI Safety (45)¶

Detect All-Type Deepfake Audio: Wavelet Prompt Tuning for Enhanced Auditory Perception: This work establishes the first all-type (speech/sound/singing/music) audio deepfake detection benchmark and proposes Wavelet Prompt Tuning (WPT) to enhance the full-frequency perception of SSL features using Discrete Wavelet Transform. Without adding training parameters, WPT outperforms full fine-tuning, achieving an average EER of only 3.58% after co-training.
An Information Theoretic Evaluation Metric for Strong Unlearning: This work reveals a fundamental flaw in existing black-box unlearning evaluation metrics (such as MIA and JSD)—modifying only the final layer can satisfy all black-box metrics while the intermediate layers completely retain information about the forgotten data. To address this, the authors propose IDI, a white-box metric that quantifies unlearning effectiveness by estimating the difference in mutual information between each layer and the forgotten labels via InfoNCE. They also propose the COLA method, which achieves IDI scores close to retraining (Retrain) on CIFAR-10/100 and ImageNet-1K.
Yours or Mine? Overwriting Attacks Against Neural Audio Watermarking: This work presents the first systematic study on overwriting attacks against neural audio watermarking, proposing a three-tier attack framework (white-box, gray-box, and black-box). It achieves near 100% attack success rates across three SOTA methods (AudioSeal, Timbre, and WavMark), exposing severe security vulnerabilities in existing audio watermarking systems.
RegionMarker: A Region-Triggered Semantic Watermarking Framework for Embedding-as-a-Service: This paper proposes RegionMarker, a watermarking framework triggered by semantic regions. By defining trigger regions in a low-dimensional space and injecting semantic watermarks, it represents the first EaaS copyright protection method capable of simultaneously defending against CSE attacks, paraphrasing attacks, and dimensional perturbation attacks.
An Improved Privacy and Utility Analysis of Differentially Private SGD with Bounded Domain and Smooth Losses: Under the assumption that the loss function is only $L$-smooth (convexity is not required), this paper derives tighter closed-form RDP privacy bounds for DPSGD and, for the first time, provides a complete convergence/utility analysis in bounded domain scenarios, revealing that a smaller parameter domain diameter can simultaneously improve both privacy and utility.
Credal Ensemble Distillation for Uncertainty Quantification: Propounds the Credal Ensemble Distillation (CED) framework, which distills a deep ensemble teacher into a single model, CREDIT. Instead of a single softmax distribution, this model predicts interval-valued class probabilities that define a credal set. It achieves superior or comparable uncertainty estimation on OOD detection tasks while substantially reducing inference overhead (reducing inference time from $5\times$ to $1\times$).
FairGSE: Fairness-Aware Graph Neural Network without High False Positive Rates: This work is the first to reveal the "FPR shortcut" issue in fairness-aware GNNs, where existing methods achieve fairness metrics by misclassifying a large number of negative samples as positive. It proposes the FairGSE framework, which reweights graph edges by maximizing 2D structural entropy to simultaneously improve fairness and reduce false positive rates, achieving a 39% reduction in FPR.
MPD-SGR: Robust Spiking Neural Networks with Membrane Potential Distribution-Driven Surrogate Gradient Regularization: A theoretical connection is established between SNN robustness error and surrogate gradient (SG) magnitude. This reveals that reducing the overlap ratio between the membrane potential distribution (MPD) and the effective gradient interval of the SG can effectively decrease sensitivity to adversarial perturbations. Based on this, the MPD-SGR regularization method is proposed, which significantly outperforms existing SNN defense methods under both vanilla and adversarial training settings.
Towards Effective, Stealthy, and Persistent Backdoor Attacks Targeting Graph Foundation Models: Proposes GFM-BA, the first systematic backdoor attack method targeting the pre-training phase of Graph Foundation Models (GFMs). By employing three modules—label-free trigger association, node-adaptive trigger generation, and persistent backdoor anchoring—it simultaneously addresses the three major challenges of effectiveness, stealthiness, and persistence.
Breaking the Dyadic Barrier: Rethinking Fairness in Link Prediction Beyond Demographic Parity: This paper reveals three fundamental flaws of dyadic fairness and Demographic Parity ($\Delta\text{DP}$) in link prediction—insufficient GNN expressiveness, obscured subgroup bias, and ranking insensitivity. It proposes a ranking-aware fairness metric based on NDKL and a post-processing algorithm, MORAL, achieving state-of-the-art fairness-utility trade-offs across six datasets.

Browse all 45 AI Safety papers →

📂 Others (117)¶

LeanRAG: Knowledge-Graph-Based Generation with Semantic Aggregation and Hierarchical Retrieval: This paper proposes LeanRAG, a framework that employs a semantic aggregation algorithm to automatically construct explicit relations among summary nodes in a hierarchical knowledge graph, thereby breaking "semantic islands." Combined with a bottom-up retrieval strategy based on the Lowest Common Ancestor (LCA), LeanRAG efficiently navigates the hierarchical structure, achieving state-of-the-art performance on four QA benchmarks while reducing retrieval redundancy by 46%.
OR-R1: Automating Modeling and Solving of Operations Research Optimization Problems: OR-R1 proposes a data-efficient two-stage training framework (SFT + TGRPO) that achieves an average solving accuracy of 67.7% using only 1/10 of the synthetic data required by ORLM, surpassing existing SOTA methods. Additionally, test-time reinforcement learning reduces the performance gap between single-sample generation (Pass@1) and multi-sample generation (Pass@8) from 13% to 7%.
On the Edge of Core (Non-)Emptiness: An Automated Reasoning Approach to Approval-Based Multi-Winner Voting: This paper proposes an automated reasoning framework based on Mixed Integer Linear Programming (MILP) to investigate the major open problem of whether core stability always exists in approval-based multi-winner voting. The framework establishes new existence results, uncovers previously unknown relationships between core stability and other axioms (e.g., Lindahl pricability), and refutes an existing conjecture.
CAE: Hierarchical Semantic Alignment for Image Clustering: By combining two complementary semantic sources — noun-level (WordNet) and caption-level (Flickr image captions) — and constructing a semantic space via optimal transport alignment followed by adaptive fusion, this work achieves training-free image clustering with a 4.2% accuracy improvement on ImageNet-1K.
DFDT: Dynamic Fast Decision Tree for IoT Data Stream Mining on Edge Devices: This paper proposes DFDT (Dynamic Fast Decision Tree), a memory-constrained data stream mining algorithm for IoT edge devices. Through three coordinated mechanisms — activity-aware pre-pruning, dynamic grace period, and adaptive tie threshold — DFDT achieves an optimal trade-off among accuracy, memory usage, and runtime.
Controllable Financial Market Generation with Diffusion Guided Meta Agent: This paper proposes the Diffusion Guided Meta Agent (DigMA), which formalizes controllable financial market generation as a conditional generation task. A conditional diffusion model captures the dynamics of market states (time-varying distribution parameters of mid-price returns and order arrival rates), while a Meta Agent with financial economics priors generates order flow under the guidance of the controller. DigMA outperforms existing methods in both controllability and generation fidelity.
Designing Incident Reporting Systems for Harms from General-Purpose AI: Through a literature review and case studies of nine safety-critical industries (nuclear energy, aviation, healthcare, etc.), this paper proposes a seven-dimensional institutional design framework for AI incident reporting systems, providing systematic guidance for policy design of general-purpose AI incident reporting in the United States.
Finding Diverse Solutions Parameterized by Cliquewidth: This paper extends the parameterized framework for finding diverse solutions from treewidth to the strictly more powerful graph parameter cliquewidth, proving that any monotone dynamic programming algorithm parameterized by a cliquewidth decomposition can be converted into an algorithm for the diverse version with minimal overhead. A new family of Venn diversity measures is also proposed.
Improved Runtime Guarantees for the SPEA2 Multi-Objective Optimizer: By rigorously analyzing the more complex selection mechanism of SPEA2, this paper demonstrates that its population dynamics are fundamentally different from those of NSGA-II — the σ-criterion induces a uniform distribution of objective values across the population — yielding runtime upper bounds with a substantially weaker dependence on population size, indicating that SPEA2 is more robust to parameter choices.
DiffMM: Efficient Method for Accurate Noisy and Sparse Trajectory Map Matching via One Step Diffusion: This paper proposes DiffMM, the first approach to introduce diffusion models into map matching. By combining a road-segment-aware trajectory encoder with a one-step Shortcut diffusion process, DiffMM achieves simultaneous improvements in accuracy and efficiency on sparse trajectories and complex road networks, with inference speed approximately 17× faster than the second-best method.

Browse all 117 Others papers →