Research map / Label-efficient and robust learning / Recognition and long-tail
Concepts' Information Bottleneck Models
Karim Galliamov, Syed M Ahsan Kazmi, Adil Khan, Adín Ramírez Rivera
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 IB regularizer IB_E on the concept layer | CBM (HI) · CUB | class accuracy: 0.644 -> 0.691 | Improved |
| add IB regularizer on the concept layer | CBM (SJ) · aPY | class accuracy: 0.797 -> 0.860 | Improved |
| add IB regularizers | CBM (HJ) · AwA2 | class accuracy: 0.853 -> 0.845 (IB_B), 0.856 (IB_E) | No clear effect |
| add IB regularizers | IntCEM and AR-CBM · CUB | class accuracy: 0.759 -> 0.778 (IntCEM); 0.761 -> 0.777 (AR-CBM) | Improved |
| add IB regularizers | CBM variants · AwA2 | concept accuracy: 0.979 -> 0.968 (CBM HJ + IB_B) | Hurt |
| add IB regularizers | CBM (SJ) · CUB | OIS (concept leakage): 4.69 -> 2.00 | Improved |
| add IB regularizers | CBM (HI) · CUB | OIS (concept leakage): 3.77 -> 3.69 | No clear effect |
| add IB_E regularizer under intervention | CBM (SJ) · CUB | AUC of test-time intervention: 54.374 -> 64.6338 | Improved |
| stop gradient from H(C) into the encoder (ablation) | IB_B on CBM (SJ) · CUB | class accuracy: 0.608 -> 0.726 | Improved |
| clip gradient norm to 1.0 (ablation) | IB_B on CBM (SJ) · CUB | class accuracy: 0.608 -> 0.663 | Improved |
| vary beta between values 0.10 and 0.25 (ablation) | IB_B and IB_E on CBM (SJ) · CUB | class accuracy: 0.725 to 0.728 | No clear effect |
| add IB regularizers when training 1000 epochs (ablation) | CBM · CUB | class accuracy: 0.795 -> 0.810 | Improved |
Closest papers by meaning
- An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations (NeurIPS 2025 · similarity 0.91)
- Stochastic Concept Bottleneck Models (NeurIPS 2024 · similarity 0.91)
- Learning to Intervene on Concept Bottlenecks (ICML 2024 · similarity 0.90)
- Energy-Based Concept Bottleneck Models: Unifying Prediction, Concept Intervention, and Probabilistic Interpretations (ICLR 2024 · similarity 0.90)
- Hybrid Concept Bottleneck Models (CVPR 2025 · similarity 0.90)
- A theoretical design of concept sets: improving the predictability of concept bottleneck models (NeurIPS 2024 · similarity 0.90)
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.
