mlsanityData to 13 Oct 2026

Editable Concept Bottleneck Models

arXiv 2405.15476 ↗First posted 24 May 2024cs.LG cs.AI cs.CVcited by 22

Not on the map. Why?

Abstract (arXiv)

Concept Bottleneck Models (CBMs) have garnered much attention for their ability to elucidate the prediction process through a humanunderstandable concept layer. However, most previous studies focused on cases where the data, including concepts, are clean. In many scenarios, we often need to remove/insert some training data or new concepts from trained CBMs for reasons such as privacy concerns, data mislabelling, spurious concepts, and concept annotation errors. Thus, deriving efficient editable CBMs without retraining from scratch remains a challenge, particularly in large-scale applications. To address these challenges, we propose Editable Concept Bottleneck Models (ECBMs). Specifically, ECBMs support three different levels of data removal: concept-label-level, concept-level, and data-level. ECBMs enjoy mathematically rigorous closed-form approximations derived from influence functions that obviate the need for retraining. Experimental results demonstrate the efficiency and adaptability of our ECBMs, affirming their practical value in CBMs.

Key facts. Cited by 22 papers in the map (data through 13 Oct 2026). Builds on 42 of its 63 references that are in the map. First posted on arXiv 24 May 2024.

Builds on

42 of 63 references.

References · 42

  1. 2014Deep Learning Face Attributes in the Wildcited by 3,686
  2. 2023A Survey of Large Language Modelscited by 2,559
  3. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,273
  4. 2017Understanding Black-box Predictions via Influence Functionscited by 1,729
  5. 2019Machine Unlearningcited by 1,056
  6. 2020Concept Bottleneck Modelscited by 1,027
  7. 2023A Survey on Multimodal Large Language Modelscited by 905
  8. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 617
  9. 2019Certified Data Removal from Machine Learning Modelscited by 532
  10. 2022Post-hoc Concept Bottleneck Modelscited by 292
  11. 2023Label-Free Concept Bottleneck Modelscited by 284
  12. 2021Unrolling SGD: Understanding Factors Influencing Machine Unlearningcited by 277
30 more
  1. 2020Approximate Data Deletion from Machine Learning Modelscited by 255
  2. 2020Opportunities and Challenges in Explainable Artificial Intelligence (XAI): A Surveycited by 252
  3. 2020Influence Functions in Deep Learning Are Fragilecited by 249
  4. 2021Machine Unlearning of Features and Labelscited by 229
  5. 2020Forgetting Outside the Box: Scrubbing Deep Networks of Information Accessible from Input-Output Observationscited by 180
  6. 2020Mixed-Privacy Forgetting in Deep Networkscited by 163
  7. 2018Fast Approximate Natural Gradient Descent in a Kronecker-factored Eigenbasiscited by 144
  8. 2020Explaining Black Box Predictions and Unveiling Data Artifacts through Influence Functionscited by 144
  9. 2021Graph Unlearningcited by 138
  10. 2020Machine Unlearning for Random Forestscited by 131
  11. 2023DataInf: Efficiently Estimating Data Influence in LoRA-tuned LLMs and Diffusion Modelscited by 121
  12. 2020FastIF: Scalable Influence Functions for Efficient Model Interpretation and Debuggingcited by 120
  13. 2023An LLM can Fool Itself: A Prompt-Based Adversarial Attackcited by 114
  14. 2023Low-Cost High-Power Membership Inference Attackscited by 106
  15. 2023Probabilistic Concept Bottleneck Modelscited by 101
  16. 2018Understanding the Origins of Bias in Word Embeddingscited by 90
  17. 2022Deep Unlearning via Randomized Conditionally Independent Hessianscited by 84
  18. 2021AI in Finance: Challenges, Techniques and Opportunitiescited by 81
  19. 2022Recommendation Unlearningcited by 80
  20. 2022Concept Bottleneck Model with Additional Unsupervised Conceptscited by 79
  21. 2022Interactive Concept Bottleneck Modelscited by 79
  22. 2024MoRAL: MoE Augmented LoRA for LLMs' Lifelong Learningcited by 63
  23. 2019Repairing without Retraining: Avoiding Disparate Impact with Counterfactual Distributionscited by 48
  24. 2024Multi-hop Question Answering under Temporal Knowledge Editingcited by 36
  25. 2023Auxiliary Losses for Learning Generalizable Concept-based Modelscited by 35
  26. 2023Certified Minimax Unlearning with Generalization Rates and Deletion Capacitycited by 29
  27. 2022SEAT: Stable and Explainable Attentioncited by 25
  28. 2024Prompt-SAW: Leveraging Relation-Aware Graphs for Textual Prompt Compressioncited by 18
  29. 2020Multi-Stage Influence Functioncited by 16
  30. 2024Dialectical Alignment: Resolving the Tension of 3H and Security Threats of LLMscited by 11

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2024: 6; 2025: 11; 2026: 5.

22 papers.

Sort:

By area:

  1. 2024From Generation to Judgment: Opportunities and Challenges of LLM-as-a-judgecited by 483
  2. 2024FlipAttack: Jailbreak LLMs via Flippingcited by 79
  3. 2024Semi-supervised Concept Bottleneck Modelscited by 21
  4. 2024A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoningcited by 15
  5. 2025Mechanistic Unveiling of Transformer Circuits: Self-Influence as a Key to Model Reasoningcited by 11
  6. 2025Graph Concept Bottleneck Modelscited by 9
  7. 2025Who's Your Judge? On the Detectability of LLM-Generated Judgmentscited by 7
  8. 2024Towards Multi-dimensional Explanation Alignment for Medical Classificationcited by 6
  9. 2025A Comprehensive Survey on the Risks and Limitations of Concept-based Modelscited by 4
  10. 2024Faithful Interpretation for Graph Neural Networkscited by 3
  11. 2025Stable Vision Concept Transformers for Medical Diagnosiscited by 3
  12. 2026In-Context Learning Operates as Concept Subspace Learningcited by 3
10 more
  1. 2025Attributing Data for Sharpness-Aware Minimizationcited by 2
  2. 2025Debugging Concept Bottleneck Models through Removal and Retrainingcited by 2
  3. 2026Matryoshka Concept Bottleneck Modelscited by 1
  4. 2025Survival Concept-Based Learning Modelscited by 0
  5. 2025Efficient Text-Attributed Graph Learning through Selective Annotation and Graph Alignmentcited by 0
  6. 2025Towards Interpretable PolSAR Image Classification: Polarimetric Scattering Mechanism Informed Concept Bottleneck and Kolmogorov-Arnold Networkcited by 0
  7. 2025Chat-CBM: Towards Interactive Concept Bottleneck Models with Frozen Large Language Modelscited by 0
  8. 2026Concept Flow Models: Anchoring Concept-Based Reasoning with Hierarchical Bottleneckscited by 0
  9. 2026CleanVideo: Adaptive Concept Erasure for Text-to-Video Diffusion Modelscited by 0
  10. 2026Interpretable but Fragile? Robustness of Concept Bottlenecks under Geometric-Semantic Perturbationscited by 0