Research map / Label-efficient and robust learning / Recognition and long-tail
A Framework for Learning Ante-Hoc Explainable Models via Concepts
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 the proposed concept framework without concept-level supervision | ResNet18 classifier · CIFAR10 | accuracy: 84.50 -> 91.68 | Helped |
| Use the proposed concept framework without concept-level supervision | ResNet18 classifier · ImageNet | accuracy: 58.55 -> 65.09 | Helped |
| Use the proposed concept framework without concept-level supervision | ResNet18 classifier · AwA2 | accuracy: 76.41 -> 81.04 | Helped |
| Use the proposed concept framework without concept-level supervision | ResNet18 classifier · CUB-200 | accuracy: 58.81 -> 63.05 | Helped |
| Add concept-level supervision | Proposed model · AwA2 | accuracy: 81.04 -> 85.70 | Helped |
| Add concept-level supervision | Proposed model · CUB-200 | accuracy: 63.05 -> 65.28 | Helped |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · CIFAR10 | accuracy: 91.68 -> 91.28 | Hurt |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · ImageNet | accuracy: 65.09 -> 64.84 | Hurt |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · AwA2 | accuracy: 81.04 -> 79.89 | Hurt |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · CUB-200 | accuracy: 63.05 -> 61.93 | Hurt |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · CIFAR10 | faithfulness: 90.86 -> 90.93 | Helped |
| Add rotation-prediction self-supervision on concepts (ablation) | Proposed model · ImageNet | faithfulness: 58.73 -> 60.28 | Helped |
Closest papers by meaning
- Explainable Concept Generation through Vision-Language Preference Learning for Understanding Neural Networks' Internal Representations (ICML 2025 · similarity 0.86)
- Enhancing Performance of Explainable AI Models with Constrained Concept Refinement (ICML 2025 · similarity 0.86)
- FastCAV: Efficient Computation of Concept Activation Vectors for Explaining Deep Neural Networks (ICML 2025 · similarity 0.86)
- Understanding Inter-Concept Relationships in Concept-Based Models (ICML 2024 · similarity 0.86)
- How to Probe: Simple Yet Effective Techniques for Improving Post-hoc Explanations (ICLR 2025 · similarity 0.86)
- CoE: Chain-of-Explanation via Automatic Visual Concept Circuit Description and Polysemanticity Quantification (CVPR 2025 · similarity 0.86)
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.