Research map / Vision-language and multimodal / Human and embodied modeling
Rethinking Concept Bottleneck Models: From Pitfalls to Solutions
Merve Tapli, Quentin Bouniot, Wolfgang Stammer, Zeynep Akata, Emre Akbas
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 knowledge distillation from linear probe teacher (ablation) | non-linear CBM · ImageNet100 | accuracy: 82.27 -> 85.47 | Improved |
| add knowledge distillation from linear probe teacher (ablation) | non-linear CBM · Places365 | accuracy: 33.64 -> 35.44 | Improved |
| add knowledge distillation (ablation) | linear CBM · ImageNet100 | accuracy: 87.96 -> 90.37 | Improved |
| add knowledge distillation (ablation) | non-linear CBM · ImageNet100 | accuracy: 82.27 -> 85.56 | Improved |
| add non-linearity (ReLU) to concept encoder (ablation) | CBM · ImageNet100 | accuracy: 87.96 -> 82.27 | Hurt |
| replace relevant concepts with random words | non-linear CBM · ImageNet100 | accuracy: drops by roughly 25 points | Hurt |
| replace relevant concepts with random words | linear CBM · ImageNet100 | accuracy | No clear effect |
| use relevant instead of random concept set | SAIL activations · ImageNet100 task-agnostic | entropy: 4.565 -> 3.660 | Improved |
| use relevant instead of random concept set | SAIL activations · Places365 task-specific | entropy: 4.578 -> 2.889 | Improved |
| use SAIL instead of CLIP for goodness metric | concept activations · Places365 task-agnostic | entropy: 4.482 -> 3.077 | Improved |
| use stronger backbone Perception Encoder instead of ResNet | CBM-Suite with SigLIP · ImageNet100 | accuracy: 90.12 -> 96.14 | Improved |
| use SigLIP instead of SAIL VLM | CBM-Suite with Perception Encoder · ImageNet100 | accuracy: 93.56 -> 96.14 | Improved |
Closest papers by meaning
- VLG-CBM: Training Concept Bottleneck Models with Vision-Language Guidance (NeurIPS 2024 · similarity 0.93)
- Energy-Based Concept Bottleneck Models: Unifying Prediction, Concept Intervention, and Probabilistic Interpretations (ICLR 2024 · similarity 0.92)
- Concepts Worth Having: Refining VLM-Guided Concept Bottleneck Models with Minimal Annotations (NeurIPS 2026 · similarity 0.92)
- Hybrid Concept Bottleneck Models (CVPR 2025 · similarity 0.92)
- V2C-CBM: Building Concept Bottlenecks with Vision-to-Concept Tokenizer (AAAI 2025 · similarity 0.92)
- Object-Centric Concept-Bottlenecks (NeurIPS 2025 · similarity 0.91)
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.
