Research map / Vision-language and multimodal / Multimodal LLMs
A Survey on Interpretability in Visual Recognition
Qiyang Wan, Chengzhi Gao, Ruiping Wang, Xilin Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use FinerCAM instead of GradCAM for concise saliency | CAM methods · CUB localization | L0 norm / Perturb.: L0 norm 0.2295 -> 0.1038; Perturb. 0.1324 -> 0.1772 | Mixed |
| use KPCA CAM instead of GradCAM | CAM methods · CUB localization | L0 norm / Deletion: L0 norm 0.2295 -> 0.1328; Deletion 0.1769 -> 0.1272 | Mixed |
| use ScoreCAM or AblationCAM instead of GradCAM | CAM methods · CUB localization | Runtime: 108 ms -> 4814 ms (ScoreCAM); 108 ms -> 3844 ms (AblationCAM) | Hurt |
| use FEM instead of GradCAM | CAM methods · CUB localization | Runtime / Deletion: Runtime 108 ms -> 33 ms; Deletion 0.1769 -> 0.1196 | Mixed |
| use VLG-CBM instead of CBM | concept bottleneck models · CUB semantic | IoC / Perturb.: IoC 0.0154 -> 0.8445; Perturb. 0.0158 -> 0.0021 | Improved |
| use VLG-CBM instead of CBM | concept bottleneck models · CUB semantic | Concept%: 0.2134 -> 0.4339 | Hurt |
| use LaBo instead of CBM | concept bottleneck models · CUB semantic | IoC: 0.0154 -> 0.0000 | Hurt |
| use CSS VL-CBM instead of LFCBM | concept bottleneck models · CUB semantic | Rel. Cost: +73.06% -> +1.28% | Improved |
Closest papers by meaning
- A Multimodal Automated Interpretability Agent (ICML 2024 · similarity 0.87)
- Navigating the Maze of Explainable AI: A Systematic Approach to Evaluating Methods and Metrics (NeurIPS 2024 · similarity 0.86)
- Enhancing Pre-trained Representation Classifiability can Boost its Interpretability (ICLR 2025 · similarity 0.86)
- Visual Graph Arena: Evaluating Visual Conceptualization of Vision and Multimodal Large Language Models (ICML 2025 · similarity 0.86)
- Enhancing Cognition and Explainability of Multimodal Foundation Models with Self-Synthesized Data (ICLR 2025 · similarity 0.86)
- MMT-Bench: A Comprehensive Multimodal Benchmark for Evaluating Large Vision-Language Models Towards Multitask AGI (ICML 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
