mlsanityData to 14 Oct 2026

Stochastic Concept Bottleneck Models

arXiv 2406.19272 ↗First posted 27 Jun 2024cs.LGcited by 51

Not on the map. Why?

Abstract (arXiv)

Concept Bottleneck Models (CBMs) have emerged as a promising interpretable method whose final prediction is based on intermediate, human-understandable concepts rather than the raw input. Through time-consuming manual interventions, a user can correct wrongly predicted concept values to enhance the model's downstream performance. We propose Stochastic Concept Bottleneck Models (SCBMs), a novel approach that models concept dependencies. In SCBMs, a single-concept intervention affects all correlated concepts, thereby improving intervention effectiveness. Unlike previous approaches that model the concept relations via an autoregressive structure, we introduce an explicit, distributional parameterization that allows SCBMs to retain the CBMs' efficient training and inference procedure. Additionally, we leverage the parameterization to derive an effective intervention strategy based on the confidence region. We show empirically on synthetic tabular and natural image datasets that our approach improves intervention effectiveness significantly. Notably, we showcase the versatility and usability of SCBMs by examining a setting with CLIP-inferred concepts, alleviating the need for manual concept annotations.

Key facts. Cited by 51 papers in the map (data through 14 Oct 2026). Builds on 29 of its 46 references that are in the map. First posted on arXiv 27 Jun 2024.

Builds on

29 of 46 references.

References · 29

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  4. 2020Language Models are Few-Shot Learnerscited by 32,419
  5. 2013Auto-Encoding Variational Bayescited by 4,053
  6. 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,739
  7. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,416
  8. 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,319
  9. 2016The Concrete Distribution: A Continuous Relaxation of Discrete Random Variablescited by 1,148
  10. 2020Concept Bottleneck Modelscited by 1,032
  11. 2016The Mythos of Model Interpretabilitycited by 721
  12. 2019Verified Uncertainty Calibrationcited by 309
17 more
  1. 2022Post-hoc Concept Bottleneck Modelscited by 293
  2. 2023Label-Free Concept Bottleneck Modelscited by 285
  3. 2021Promises and Pitfalls of Black-Box Concept Learning Modelscited by 131
  4. 2021Do Concept Bottleneck Models Learn as Intended?cited by 116
  5. 2020Stochastic Segmentation Networks: Modelling Spatially Correlated Aleatoric Uncertaintycited by 101
  6. 2023Probabilistic Concept Bottleneck Modelscited by 101
  7. 2022Interactive Concept Bottleneck Modelscited by 79
  8. 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Modelscited by 64
  9. 2023Learning to Receive Help: Intervention-Aware Concept Embedding Modelscited by 41
  10. 2023Human Uncertainty in Concept-Based AI Systemscited by 39
  11. 2018Influence-Directed Explanations for Deep Convolutional Networkscited by 31
  12. 2023Coarse-to-Fine Concept Bottleneck Modelscited by 31
  13. 2023Learning to Intervene on Concept Bottleneckscited by 29
  14. 2024Beyond Concept Bottleneck Models: How to Make Black Boxes Intervenable?cited by 28
  15. 2023Sparse Linear Concept Discovery Modelscited by 22
  16. 2023Interpretable and intervenable ultrasonography-based machine learning models for pediatric appendicitiscited by 11
  17. 2024Improving Intervention Efficacy via Concept Realignment in Concept Bottleneck Modelscited by 6

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 27; 2026: 23.

51 papers.

Sort:

By area:

  1. 2025A Comprehensive Survey on Self-Interpretable Neural Networkscited by 18
  2. 2025Causally Reliable Concept Bottleneck Modelscited by 17
  3. 2025If Concept Bottlenecks are the Question, are Foundation Models the Answer?cited by 15
  4. 2025Leakage and Interpretability in Concept-Based Modelscited by 13
  5. 2025Measuring Leakage in Concept-Based Methods: An Information Theoretic Approachcited by 9
  6. 2025Avoiding Leakage Poisoning: Concept Interventions Under Distribution Shiftscited by 9
  7. 2025Graph Concept Bottleneck Modelscited by 9
  8. 2025There Was Never a Bottleneck in Concept Bottleneck Modelscited by 8
  9. 2025Addressing Concept Mislabeling in Concept Bottleneck Models Through Preference Optimizationcited by 7
  10. 2025Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoningcited by 6
  11. 2025Enhancing Concept Localization in CLIP-based Concept Bottleneck Modelscited by 6
  12. 2025An Analysis of Concept Bottleneck Models: Measuring, Understanding, and Mitigating the Impact of Noisy Annotationscited by 5
38 more
  1. 2025LogicCBMs: Logic-Enhanced Concept-Based Learningcited by 5
  2. 2025Towards Reasonable Concept Bottleneck Modelscited by 3
  3. 2025Interpretable Hierarchical Concept Reasoning through Attention-Guided Graph Learningcited by 3
  4. 2025Interpretable Reward Modeling with Active Concept Bottleneckscited by 3
  5. 2025Quantifying the Accuracy-Interpretability Trade-Off in Concept-Based Sidechannel Modelscited by 3
  6. 2025A Geometric Unification of Concept Learning with Concept Conescited by 3
  7. 2026Hierarchical Concept-based Interpretable Modelscited by 3
  8. 2025From Pixels to Perception: Interpretable Predictions via Instance-wise Grouped Feature Selectioncited by 2
  9. 2026Process-Guided Concept Bottleneck Modelcited by 2
  10. 2026Mixture of Concept Bottleneck Expertscited by 2
  11. 2026Prototype-Grounded Concept Models for Verifiable Concept Alignmentcited by 2
  12. 2026Concepts Worth Having: Refining VLM-Guided Concept Bottleneck Models with Minimal Annotationscited by 2
  13. 2026Matryoshka Concept Bottleneck Modelscited by 2
  14. 2024Exploiting Interpretable Capabilities with Concept-Enhanced Diffusion and Prototype Networkscited by 1
  15. 2025Automated Genomic Interpretation via Concept Bottleneck Models for Medical Roboticscited by 1
  16. 2025Parameter-Efficient Construction of the Rashomon Slice for Concept Bottleneck Modelscited by 1
  17. 2025SALVE: Sparse Autoencoder-Latent Vector Editing for Mechanistic Control of Neural Networkscited by 1
  18. 2026Differentiable Logic Synthesis: Spectral Coefficient Selection via Sinkhorn-Constrained Compositioncited by 1
  19. 2026UrbanAlign: Post-hoc Semantic Calibration for VLM-Human Preference Alignmentcited by 1
  20. 2026SCoCCA: Multi-modal Sparse Concept Decomposition via Canonical Correlation Analysiscited by 1
  21. 2026Hoeffding Concept Bottleneck Models with Applications to Overhead Imagescited by 1
  22. 2026The Standard Interpretable Model: A general theory of interpretable machine learning to deductively design interpretable methods using Lagrangian mechanicscited by 1
  23. 2025Survival Concept-Based Learning Modelscited by 0
  24. 2025Signals, Concepts, and Laws: Toward Universal, Explainable Time-Series Forecastingcited by 0
  25. 2025Chat-CBM: Towards Interactive Concept Bottleneck Models with Frozen Large Language Modelscited by 0
  26. 2025Uncertainty-Aware Concept Bottleneck Models with Enhanced Interpretabilitycited by 0
  27. 2025Post-hoc Stochastic Concept Bottleneck Modelscited by 0
  28. 2026Interpretability in Deep Time Series Models Demands Semantic Alignmentcited by 0
  29. 2026Causal Neural Probabilistic Circuitscited by 0
  30. 2026Digging Deeper: Learning Multi-Level Concept Hierarchiescited by 0
  31. 2026Prism: Policy Reuse via Interpretable Strategy Mapping in Reinforcement Learningcited by 0
  32. 2026A Composite Activation Function for Learning Stable Binary Representationscited by 0
  33. 2026CB-SLICE: Concept-Based Interpretable Error Slice Discoverycited by 0
  34. 2026Formal Concept Lattices are Good Semantic Scaffolds for Concept-Based Learningcited by 0
  35. 2026In Defense of Information Leakage in Concept-based Modelscited by 0
  36. 2026Concept Flow Models: Anchoring Concept-Based Reasoning with Hierarchical Bottleneckscited by 0
  37. 2026Escaping Local Views: Discovering Latent Concepts for Interpretable Multi-Agent Reinforcement Learningcited by 0
  38. 2026Interpretable but Fragile? Robustness of Concept Bottlenecks under Geometric-Semantic Perturbationscited by 0

50 most cited of 51 All 51 (JSON)