mlsanityData to 14 Oct 2026

Graph Concept Bottleneck Models

arXiv 2508.14255 ↗First posted 19 Aug 2025cs.LGcited by 9

Not on the map. Why?

Abstract (arXiv)

Concept Bottleneck Models (CBMs) provide explicit interpretations for deep neural networks through concepts and allow intervention with concepts to adjust final predictions. Existing CBMs assume concepts are conditionally independent given labels and isolated from each other, ignoring the hidden relationships among concepts. However, the set of concepts in CBMs often has an intrinsic structure where concepts are generally correlated: changing one concept will inherently impact its related concepts. To mitigate this limitation, we propose GraphCBMs: a new variant of CBM that facilitates concept relationships by constructing latent concept graphs, which can be combined with CBMs to enhance model performance while retaining their interpretability. Our experiment results on real-world image classification tasks demonstrate Graph CBMs offer the following benefits: (1) superior in image classification tasks while providing more concept structure information for interpretability; (2) able to utilize latent concept graphs for more effective interventions; and (3) robust in performance across different training and architecture settings.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Builds on 46 of its 59 references that are in the map. First posted on arXiv 19 Aug 2025.

Builds on

46 of 59 references.

References · 46

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  4. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  5. 2020Language Models are Few-Shot Learnerscited by 32,419
  6. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  7. 2022Training language models to follow instructions with human feedbackcited by 15,971
  8. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  9. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  10. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  11. 2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,446
  12. 2019CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparisoncited by 1,423
34 more
  1. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,405
  2. 2020Graph Contrastive Learning with Augmentationscited by 1,208
  3. 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,167
  4. 2020Concept Bottleneck Modelscited by 1,032
  5. 2017Zero-Shot Learning -- A Comprehensive Evaluation of the Good, the Bad and the Uglycited by 538
  6. 2019InfoGraph: Unsupervised and Semi-supervised Graph-Level Representation Learning via Mutual Information Maximizationcited by 441
  7. 2018Invariant and Equivariant Graph Networkscited by 346
  8. 2018Neural Relational Inference for Interacting Systemscited by 339
  9. 2019DAG-GNN: DAG Structure Learning with Graph Neural Networkscited by 331
  10. 2022Language in a Bottle: Language Model Guided Concept Bottlenecks for Interpretable Image Classificationcited by 312
  11. 2022Post-hoc Concept Bottleneck Modelscited by 293
  12. 2023Label-Free Concept Bottleneck Modelscited by 285
  13. 2019Learning Discrete Structures for Graph Neural Networkscited by 231
  14. 2017struc2vec: Learning Node Representations from Structural Identitycited by 215
  15. 2023Learning to Exploit Temporal Structure for Biomedical Vision-Language Processingcited by 211
  16. 2021Discrete Graph Structure Learning for Forecasting Multiple Time Seriescited by 171
  17. 2018Learning graphs from data: A signal representation perspectivecited by 119
  18. 2023Probabilistic Concept Bottleneck Modelscited by 101
  19. 2020Differentiable Graph Module (DGM) for Graph Convolutional Networkscited by 86
  20. 2023Learning Concise and Descriptive Attributes for Visual Recognitioncited by 73
  21. 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Modelscited by 64
  22. 2024Stochastic Concept Bottleneck Modelscited by 51
  23. 2023Learning Bottleneck Concepts in Image Classificationcited by 49
  24. 2019Semantic-Aware Scene Recognitioncited by 43
  25. 2024Incremental Residual Concept Bottleneck Modelscited by 43
  26. 2024Energy-Based Concept Bottleneck Models: Unifying Prediction, Concept Intervention, and Probabilistic Interpretationscited by 33
  27. 2022Augmentations in Hypergraph Contrastive Learning: Fabricated and Generativecited by 29
  28. 2024Editable Concept Bottleneck Modelscited by 23
  29. 2023Sparse Linear Concept Discovery Modelscited by 22
  30. 2024Linear Explanations for Individual Neuronscited by 19
  31. 2025Causally Reliable Concept Bottleneck Modelscited by 17
  32. 2024DCBM: Data-Efficient Visual Concept Bottleneck Modelscited by 16
  33. 2022Multi-level Contrastive Learning Framework for Sequential Recommendationcited by 8
  34. 2023Federated Learning of Models Pre-Trained on Different Features with Consensus Graphscited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 7.

9 papers.

  1. 2025Interpretable Neuropsychiatric Diagnosis via Concept-Guided Graph Neural Networkscited by 1
  2. 2026Revealing Combinatorial Reasoning of GNNs via Graph Concept Bottleneck Layercited by 1
  3. 2025A Tool Bottleneck Framework for Clinically-Informed and Interpretable Medical Image Understandingcited by 0
  4. 2026Distilling LLM Reasoning into Graph of Concept Predictorscited by 0
  5. 2026DCG-Net: Dual Cross-Attention with Concept-Value Graph Reasoning for Interpretable Medical Diagnosiscited by 0
  6. 2026Exploring Concept Subspace for Self-explainable Text-Attributed Graph Learningcited by 0
  7. 2026Hyperbolic Concept Bottleneck Modelscited by 0
  8. 2026ReCBM: Uncertainty-Gated Relational Reasoning for Concept Bottleneck Modelscited by 0
  9. 2026MolLedger: An Additive Graph Neural Network with Chemically Grounded ADME Attributionscited by 0