ConEx Bridges Saliency Maps and Concept-Based Reasoning for Interpretable Vision
Yehonatan Elisha · hf · 2026-10-07
Researchers introduce ConEx (Concept-based Explanations), a framework linking pixel-level saliency visualization with concept-based reasoning so visual explanations are both faithful and interpretable.
ConEx automatically discovers class-specific concepts represented as concept activation vectors (CAVs), learned without manual supervision via an architecture-specific masking mechanism that reduces segmentation-mask noise. Its saliency maps reveal where each concept appears and how it contributes to the prediction. Two complementary metrics, Vector-Concept Match (VCM) and Concept-Class Match (CCM), quantify concept alignment and enable comparison with prior methods.
Experiments show state-of-the-art performance on faithfulness, segmentation, and concept-quality benchmarks.
More from Research
- Slides released: 'Ultimate guide to multi-harness RL' talk from Kernel Panic Madrid — SergioPaniego · 2026-10-07
- AI Math Skeptics Pushed Back On—'Give It 2 More Months,' Researcher Quips — AvivTamar1 · 2026-10-07
- New proof bounds π's irrationality exponent at 6.0446, fully formalized in Lean 4 — Michael_D_Moor · 2026-10-07
- After 3-sum hits n^1.999, complexity theorists debate whether math is 'all square packing' — thomasahle · 2026-10-07
- NVIDIA's UNREAL: one model unifies corpus retrieval and long-context at 128K+ — nvidia · 2026-10-07
- gpt-4.1-nano as RAG judge barely beats chance: AUROC 0.603 on RAGBench — Giulio Zeloni · 2026-10-07