mlsanityData to 7 Oct 2026

Understanding Inter-Concept Relationships in Concept-Based Models

arXiv 2405.18217 ↗First posted 28 May 2024cs.LGcited by 13

Not on the map. Why?

Abstract (arXiv)

Concept-based explainability methods provide insight into deep learning systems by constructing explanations using human-understandable concepts. While the literature on human reasoning demonstrates that we exploit relationships between concepts when solving tasks, it is unclear whether concept-based methods incorporate the rich structure of inter-concept relationships. We analyse the concept representations learnt by concept-based models to understand whether these models correctly capture inter-concept relationships. First, we empirically demonstrate that state-of-the-art concept-based models produce representations that lack stability and robustness, and such methods fail to capture inter-concept relationships. Then, we develop a novel algorithm which leverages inter-concept relationships to improve concept intervention accuracy, demonstrating how correctly capturing inter-concept relationships can improve downstream tasks.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 30 of its 50 references that are in the map. First posted on arXiv 28 May 2024.

Builds on

30 of 50 references.

References · 30

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
  3. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  4. 2013Efficient Estimation of Word Representations in Vector Spacecited by 4,346
  5. 2019Invariant Risk Minimizationcited by 1,832
  6. 2016CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoningcited by 1,445
  7. 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,410
  8. 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,309
  9. 2020Concept Bottleneck Modelscited by 1,022
  10. 2020Knowledge Graphscited by 619
  11. 2020Does the Whole Exceed its Parts? The Effect of AI Explanations on Complementary Team Performancecited by 396
  12. 2019The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervisioncited by 369
18 more
  1. 2021Interpretable Machine Learning: Fundamental Principles and 10 Grand Challengescited by 364
  2. 2022Post-hoc Concept Bottleneck Modelscited by 289
  3. 2023Label-Free Concept Bottleneck Modelscited by 282
  4. 2019Explanations can be manipulated and geometry is to blamecited by 215
  5. 2022Quantus: An Explainable AI Toolkit for Responsible Evaluation of Neural Network Explanations and Beyondcited by 171
  6. 2021Promises and Pitfalls of Black-Box Concept Learning Modelscited by 128
  7. 2021What Do We Want From Explainable Artificial Intelligence (XAI)? -- A Stakeholder Perspective on XAI and a Conceptual Model Guiding Interdisciplinary XAI Researchcited by 119
  8. 2023Probabilistic Concept Bottleneck Modelscited by 101
  9. 2022Interactive Concept Bottleneck Modelscited by 79
  10. 2022GlanceNets: Interpretabile, Leak-proof Concept-based Modelscited by 73
  11. 2022Concept Embedding Models: Beyond the Accuracy-Explainability Trade-Offcited by 61
  12. 2022Concept-level Debugging of Part-Prototype Networkscited by 52
  13. 2020Now You See Me (CME): Concept-based Model Extractioncited by 51
  14. 2023Learning to Receive Help: Intervention-Aware Concept Embedding Modelscited by 41
  15. 2023Human Uncertainty in Concept-Based AI Systemscited by 39
  16. 2023Towards Robust Metrics for Concept Representation Evaluationcited by 29
  17. 2022Trust in AI: Interpretability is not necessary or sufficient, while black-box interaction is necessary and sufficientcited by 13
  18. 2024Do Concept Bottleneck Models Respect Localities?cited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 8; 2026: 5.

13 papers.

Sort:

By area:

  1. 2025A Comprehensive Survey on Self-Interpretable Neural Networkscited by 18
  2. 2025Addressing Concept Mislabeling in Concept Bottleneck Models Through Preference Optimizationcited by 7
  3. 2025SUB: Benchmarking CBM Generalization via Synthetic Attribute Substitutionscited by 6
  4. 2025LogicCBMs: Logic-Enhanced Concept-Based Learningcited by 5
  5. 2025B-cos LM: Efficiently Transforming Pre-trained Language Models for Improved Explainabilitycited by 4
  6. 2025A Comprehensive Survey on the Risks and Limitations of Concept-based Modelscited by 4
  7. 2025On The Variability of Concept Activation Vectorscited by 3
  8. 2026Hierarchical Concept-based Interpretable Modelscited by 3
  9. 2025Bridging Fairness and Explainability: Can Input-Based Explanations Promote Fairness in Hate Speech Detection?cited by 0
  10. 2026Digging Deeper: Learning Multi-Level Concept Hierarchiescited by 0
  11. 2026Metonymy in vision models undermines attention-based interpretabilitycited by 0
  12. 2026Formal Concept Lattices are Good Semantic Scaffolds for Concept-Based Learningcited by 0
1 more
  1. 2026ICON Decomposition: Auditing deep neural networks for shortcuts by decomposing layer-wise representations using conceptscited by 0