Research map / Label-efficient and robust learning / Recognition and long-tail
Concept Embedding Models: Beyond the Accuracy-Explainability Trade-Off
Read the paperCheck 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 concept embeddings instead of Boolean or Fuzzy scalar concepts | Concept bottleneck models · Dot | task accuracy: up to +45% | Helped |
| Add concept supervision through MixCEM | CelebA task model · CelebA | task accuracy: +5% | Helped |
| Use a MixCEM bottleneck instead of a Hybrid-CBM bottleneck | Logistic linear model predicting unannotated CUB concepts · CUB, remaining 75% of concept annotations | mean concept accuracy: 91.83% ± 0.51% -> 94.33% ± 0.88% | Helped |
| Use Hybrid-CBM unsupervised bottleneck activations | Concept bottleneck representations · Dot | concept alignment score: up to 25% lower | Hurt |
| Use Hybrid-CBM bottleneck interventions | Hybrid-CBM · CUB and CelebA | task accuracy response to concept interventions | No clear effect |
| Add RandInt training with randomly applied concept interventions (ablation) | MixCEM · CUB and CelebA | task accuracy after concept interventions | Helped |
| Increase the RandInt intervention probability (ablation) | MixCEM · CUB | task accuracy without interventions | Hurt |
| Increase the RandInt intervention probability (ablation) | MixCEM · CUB | task accuracy with interventions | Helped |
Closest papers by meaning
- Hyperbolic Concept Bottleneck Models (NeurIPS 2026 · similarity 0.90)
- Sample-efficient Learning of Concepts with Theoretical Guarantees: from Data to Concepts without Interventions (NeurIPS 2025 · similarity 0.90)
- An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotations (NeurIPS 2025 · similarity 0.89)
- Causally Reliable Concept Bottleneck Models (NeurIPS 2025 · similarity 0.89)
- Interpretable but Fragile? Robustness of Concept Bottlenecks under Geometric-Semantic Perturbations (NeurIPS 2026 · similarity 0.89)
- Learning to Intervene on Concept Bottlenecks (ICML 2024 · 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.