Research map / Detection and segmentation / Open-set and few-shot detection
Explain Any Concept: Segment Anything Meets Concept-Based Explanation
Ao Sun, Pingchuan Ma, Yuanyuan Yuan, Shuai Wang
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 |
|---|---|---|---|
| add parameter sharing (frozen target FC layer) in surrogate (ablation) | PIE surrogate · ImageNet/Insertion | AUC: 50.40 -> 81.78 | Improved |
| add parameter sharing (ablation) | PIE surrogate · COCO/Insertion | AUC: 42.86 -> 87.08 | Improved |
| add parameter sharing (ablation) | PIE surrogate · ImageNet/Deletion | AUC (lower better): 32.87 -> 12.47 | Improved |
| use PIE surrogate instead of linear surrogate model (ablation) | Shapley estimation · ImageNet/Insertion | AUC: 78.11 -> 81.78 | Improved |
| use PIE surrogate instead of linear surrogate model (ablation) | Shapley estimation · COCO/Deletion | AUC (lower better): 14.36 -> 13.71 | Improved |
| use PIE instead of linear surrogate (ablation) | Shapley estimation · ImageNet/Insertion | processing time (sec.): 36 -> 245 | Hurt |
| use SAM concepts with PIE instead of superpixel baselines | EAC vs DeepLIFT · ImageNet/Insertion | AUC (%): 75.235 -> 83.400 | Improved |
| use SAM concepts instead of morphological-operation clustering (ablation) | DeepLIFT · ImageNet/Insertion | AUC (%): 14.707 -> 83.400 | Improved |
| use EAC instead of baseline explanations | EAC · human preference ImageNet | images preferred (of 100): 70 vs 6 | Improved |
Closest papers by meaning
- Towards Human-Understandable Multi-Dimensional Concept Discovery (CVPR 2025 · similarity 0.87)
- Show and Tell: Visually Explainable Deep Neural Nets via Spatially-Aware Concept Bottleneck Models (CVPR 2025 · similarity 0.87)
- I-CEE: Tailoring Explanations of Image Classification Models to User Expertise (AAAI 2024 · similarity 0.86)
- Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks' Internal Representations (ICML 2025 · similarity 0.86)
- Perceive Anything: Recognize, Explain, Caption, and Segment Anything in Images and Videos (NeurIPS 2025 · similarity 0.86)
- Q-SENN: Quantized Self-Explaining Neural Networks (AAAI 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.
