🔬 Interpretability¶

📹 ICCV2025 · 11 paper notes

AIM: Amending Inherent Interpretability via Self-Supervised Masking: This paper proposes AIM, a top-down learnable binary masking mechanism for self-supervised spatial feature selection, built upon a feature pyramid architecture. Without requiring additional annotations, AIM guides CNNs to focus on genuinely discriminative features and suppress spurious correlations, simultaneously achieving inherent interpretability and improved OOD generalization.
ArgoTweak: Towards Self-Updating HD Maps through Structured Priors: This paper proposes ArgoTweak, the first HD map dataset providing complete triplets of "prior map + current sensor data + up-to-date ground-truth map." It decomposes large-scale map modifications into element-level atomic changes via a bijective change mapping framework, and introduces interpretable evaluation metrics (mAPC/mACC). Models trained on ArgoTweak reduce the sim2real gap by more than 10× compared to synthetic-prior baselines.
CAD-Recode: Reverse Engineering CAD Code from Point Clouds: This paper proposes CAD-Recode, which translates point clouds into executable Python CadQuery code to reconstruct CAD models. By leveraging a pretrained LLM (Qwen2-1.5B) as the decoder paired with a lightweight point cloud encoder, the method achieves more than 10× reduction in Chamfer Distance on three benchmarks: DeepCAD, Fusion360, and CC3D.
CAD-Recode: Reverse Engineering CAD Code from Point Clouds: CAD-Recode frames 3D CAD reverse engineering as a point-cloud-to-Python-code translation task. It leverages the Python code understanding capabilities of pretrained LLMs as the decoder, combined with a lightweight point cloud projector and a million-scale procedurally generated dataset, achieving significant improvements over existing methods on multiple CAD benchmarks while enabling LLM-driven CAD editing and question answering.
CE-FAM: Concept-Based Explanation via Fusion of Activation Maps: CE-FAM is a concept explanation method that trains a branch network sharing activation maps with an image classifier to simulate VLM embeddings, establishing a one-to-one correspondence among concept prediction → concept region (weighted sum of activation maps) → concept contribution (effect on classification score). The paper also introduces a novel NRA evaluation metric and surpasses existing methods on zero-shot concept reasoning.
Granular Concept Circuits: Toward a Fine-Grained Circuit Discovery for Concept Representations: This paper proposes Granular Concept Circuit (GCC), a method that automatically discovers fine-grained visual circuits encoding specific concepts in deep visual models by iteratively evaluating inter-neuron functional dependency (Neuron Sensitivity Score) and semantic consistency (Semantic Flow Score). GCC is the first method capable of discovering multiple concept-level circuits within a single query.
Learnable Fractional Reaction-Diffusion Dynamics for Under-Display ToF Imaging and Beyond: LFRD² proposes a hybrid framework that combines learnable time-fractional reaction-diffusion equations with neural networks for under-display ToF (UD-ToF) depth map restoration. The approach captures long-range memory dependencies across iterations via fractional calculus and introduces an efficient continuous convolution operator to replace discrete convolution, achieving state-of-the-art performance on UD-ToF depth restoration, ToF denoising, and depth super-resolution tasks.
Minerva: Evaluating Complex Video Reasoning: This paper introduces Minerva — a manually annotated benchmark of 1,515 complex video reasoning QA pairs, each with 5 answer choices and a detailed reasoning trace, designed to evaluate the video reasoning capabilities of multimodal large language models. It further establishes a video reasoning error taxonomy (Temporal / Perceptual / Logical / Completeness) and the MiRA automated evaluation framework.
"Principal Components" Enable A New Language of Images: This paper proposes Semanticist, a visual tokenization framework that embeds a provable PCA structure into the latent token space—where each subsequent token contributes decreasing, non-overlapping information—and employs a diffusion decoder to decouple the semantic-spectral entanglement effect, achieving state-of-the-art performance on both image reconstruction and autoregressive generation.
SVIP: Semantically Contextualized Visual Patches for Zero-Shot Learning: This paper proposes the SVIP framework, which addresses semantic misalignment in zero-shot learning at its source by identifying and replacing semantically irrelevant image patches at the input stage with learnable embeddings initialized from attribute-level word embeddings.
VITAL: More Understandable Feature Visualization through Distribution Alignment and Relevant Information Flow: This paper proposes VITAL, a feature visualization method that reframes the problem as aligning intermediate feature distributions with those of real images (rather than conventional activation maximization), and incorporates relevance scores to filter irrelevant features, producing neuron visualizations that are more interpretable to humans.