Research map / Label-efficient and robust learning / Recognition and long-tail
Probabilistic Concept Bottleneck Models
Eunji Kim, Dahuin Jung, Sangha Park, Siwon Kim, Sungroh Yoon
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 |
|---|---|---|---|
| introduce probabilistic rather than deterministic embeddings (ablation) | ProbCBM · CUB, 224 × 224 | Class accuracy: 0.677$\pm$0.004 -> 0.680$\pm$0.004 | No clear effect |
| use embedding means instead of Monte Carlo sampling at inference (ablation) | ProbCBM · CUB, 224 × 224 | Class accuracy: 0.679$\pm$0.003 -> 0.680$\pm$0.004 | No clear effect |
| use probabilistic embeddings instead of deterministic embeddings (ablation) | ProbCBM · AwA2 | Class accuracy: 0.880$\pm$0.002 -> 0.880$\pm$0.002 | No clear effect |
| use embedding means instead of Monte Carlo sampling at inference (ablation) | ProbCBM · AwA2 | Class accuracy: 0.880$\pm$0.001 -> 0.880$\pm$0.002 | No clear effect |
| increase class-training concept replacement probability to 1 (ablation) | ProbCBM · CUB | Classification performance | Hurt |
| vary the concept embedding dimension (ablation) | ProbCBM · CUB | Concept and class prediction performance | No clear effect |
| vary the class embedding dimension (ablation) | ProbCBM · CUB | Concept and class prediction performance | No clear effect |
| use a small concept embedding dimension (Dc=4) (ablation) | ProbCBM · CUB occlusion uncertainty detection | Proportion of samples with increased uncertainty: Dc=4 | Hurt |
| add ground-truth concept replacement during class-predictor training | ProbCBM · CUB and AwA2 | Class accuracy | Improved |
| rank concepts by uncertainty for intervention rather than intervening in random order | ProbCBM · CUB | Class accuracy | Improved |
| occlude image regions | ProbCBM · CUB and AwA2 | Uncertainty detection | Improved |
Closest papers by meaning
- Editable Concept Bottleneck Models (ICML 2025 · similarity 0.89)
- Stochastic Concept Bottleneck Models (NeurIPS 2024 · similarity 0.89)
- Energy-Based Concept Bottleneck Models: Unifying Prediction, Concept Intervention, and Probabilistic Interpretations (ICLR 2024 · similarity 0.89)
- Bayesian Concept Bottleneck Models with LLM Priors (NeurIPS 2025 · similarity 0.88)
- A theoretical design of concept sets: improving the predictability of concept bottleneck models (NeurIPS 2024 · similarity 0.88)
- An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations (NeurIPS 2025 · similarity 0.88)
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.
