🩺 Medical LLM¶

🤖 AAAI2026 · 12 paper notes

📌 Same area in other venues: 📷 CVPR2026 (1) · 🔬 ICLR2026 (20) · 💬 ACL2026 (47) · 🧪 ICML2026 (4) · 🧠 NeurIPS2025 (17) · 🧪 ICML2025 (4)

🔥 Top topics: Medical Imaging ×4 · LLM ×2 · RAG ×2 · Biomolecules ×2

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%.
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.
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.
GEM: Generative Entropy-Guided Preference Modeling for Few-shot Alignment of LLMs: GEM proposes a generative entropy-guided preference modeling approach that achieves efficient LLM alignment in low-resource settings (only 3,000 preference pairs) through cognitive filtering (entropy-based CoT scoring) and the SEGA algorithm (Self-Evaluated Group Advantage policy optimization).
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.
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.
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.
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.
Real-Time Trust Verification for Safe Agentic Actions Using TrustBench: This paper proposes TrustBench, a dual-mode framework: (1) Benchmark Mode — combines traditional metrics with LLM-as-a-Judge to evaluate 8 trust dimensions and learns a calibration mapping from agent confidence to actual accuracy; (2) Verification Mode — computes trust scores in real time after an agent formulates an action but before execution, blocking 87% of harmful actions with latency below 200ms, with specialization achieved through domain plugins (medical/financial/QA).
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.
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.