mlsanityData to 7 Oct 2026

Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lens

arXiv 2502.11245 ↗First posted 16 Feb 2025cs.LG cs.AIcited by 13

Not on the map. Why?

Abstract (arXiv)

Concept-based Models are neural networks that learn a concept extractor to map inputs to high-level concepts and an inference layer to translate these into predictions. Ensuring these modules produce interpretable concepts and behave reliably in out-of-distribution is crucial, yet the conditions for achieving this remain unclear. We study this problem by establishing a novel connection between Concept-based Models and reasoning shortcuts (RSs), a common issue where models achieve high accuracy by learning low-quality concepts, even when the inference layer is fixed and provided upfront. Specifically, we extend RSs to the more complex setting of Concept-based Models and derive theoretical conditions for identifying both the concepts and the inference layer. Our empirical results highlight the impact of RSs and show that existing methods, even combined with multiple natural mitigation strategies, often fail to meet these conditions in practice.

Key facts. Cited by 13 papers in the map (data through 7 Oct 2026). Builds on 77 of its 109 references that are in the map. First posted on arXiv 16 Feb 2025.

Builds on

77 of 109 references.

References · 77

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2015Distilling the Knowledge in a Neural Networkcited by 10,076
  4. 2015Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networkscited by 9,617
  5. 2013Auto-Encoding Variational Bayescited by 3,980
  6. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  7. 2020Shortcut Learning in Deep Neural Networkscited by 1,928
  8. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  9. 2016CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoningcited by 1,445
  10. 2020Concept Bottleneck Modelscited by 1,022
  11. 2018This Looks Like That: Deep Learning for Interpretable Image Recognitioncited by 731
  12. 2018Towards Robust Interpretability with Self-Explaining Neural Networkscited by 505
38 more
  1. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 494
  2. 2019Neural Probabilistic Logic Programming in DeepProbLogcited by 389
  3. 2022Language in a Bottle: Language Model Guided Concept Bottlenecks for Interpretable Image Classificationcited by 310
  4. 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 308
  5. 2022Post-hoc Concept Bottleneck Modelscited by 289
  6. 2023Label-Free Concept Bottleneck Modelscited by 282
  7. 2017A Semantic Loss Function for Deep Learning with Symbolic Knowledgecited by 267
  8. 2017Deep Learning for Case-Based Reasoning through Prototypes: A Neural Network that Explains Its Predictionscited by 234
  9. 2020Concept Whitening for Interpretable Image Recognitioncited by 225
  10. 2019SATNet: Bridging deep learning and logical reasoning using a differentiable satisfiability solvercited by 173
  11. 2021Promises and Pitfalls of Black-Box Concept Learning Modelscited by 128
  12. 2021Do Concept Bottleneck Models Learn as Intended?cited by 114
  13. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 105
  14. 2020On Linear Identifiability of Learned Representationscited by 103
  15. 2022Semantic Probabilistic Layers for Neuro-Symbolic Learningcited by 103
  16. 2020ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICAcited by 102
  17. 2023Probabilistic Concept Bottleneck Modelscited by 101
  18. 2023Concept-based Explainable Artificial Intelligence: A Surveycited by 97
  19. 2021Independent mechanism analysis, a new concept?cited by 96
  20. 2020Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanationscited by 94
  21. 2017Logic Tensor Networks for Semantic Image Interpretationcited by 92
  22. 2023NeurASP: Embracing Neural Networks into Answer Set Programmingcited by 92
  23. 2020Explainable Object-induced Action Decision for Autonomous Vehiclescited by 88
  24. 2019The Incomplete Rosetta Stone Problem: Identifiability Results for Multi-View Nonlinear ICAcited by 86
  25. 2022Concept Bottleneck Model with Additional Unsupervised Conceptscited by 79
  26. 2022GlanceNets: Interpretabile, Leak-proof Concept-based Modelscited by 73
  27. 2020Coherent Hierarchical Multi-Label Classification Networkscited by 65
  28. 2024VLG-CBM: Training Concept Bottleneck Models with Vision-Language Guidancecited by 65
  29. 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Modelscited by 63
  30. 2023Not All Neuro-Symbolic Concepts Are Created Equal: Analysis and Mitigation of Reasoning Shortcutscited by 63
  31. 2022Concept Embedding Models: Beyond the Accuracy-Explainability Trade-Offcited by 61
  32. 2022Function Classes for Identifiable Nonlinear Independent Component Analysiscited by 53
  33. 2022Leveraging Explanations in Interactive Machine Learning: An Overviewcited by 52
  34. 2022A-NeSI: A Scalable Approximate Method for Probabilistic Neurosymbolic Inferencecited by 50
  35. 2021MultiplexNet: Towards Fully Satisfied Logical Constraints in Neural Networkscited by 44
  36. 2022Synergies between Disentanglement and Sparsity: Generalization and Identifiability in Multi-Task Learningcited by 44
  37. 2023Identifying Spurious Biases Early in Training through the Lens of Simplicity Biascited by 44
  38. 2023Learning to Receive Help: Intervention-Aware Concept Embedding Modelscited by 41

50 most cited of 77 All 77 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 6; 2026: 7.

13 papers.

Sort:

By area:

  1. 2025If Concept Bottlenecks are the Question, are Foundation Models the Answer?cited by 15
  2. 2025Object Centric Concept Bottleneckscited by 10
  3. 2025When Does Closeness in Distribution Imply Representational Similarity? An Identifiability Perspectivecited by 4
  4. 2025Quantifying the Accuracy-Interpretability Trade-Off in Concept-Based Sidechannel Modelscited by 3
  5. 2026GNN Explanations that do not Explain and How to find Themcited by 3
  6. 2025Towards Reasonable Concept Bottleneck Modelscited by 2
  7. 2025Neural Concept Verifier: Scaling Prover-Verifier Games via Concept Encodingscited by 0
  8. 2026Logit Distance Bounds Representational Similaritycited by 0
  9. 2026Constraint-Based Analysis of Reasoning Shortcuts in Neurosymbolic Learningcited by 0
  10. 2026Concepts Worth Having: Refining VLM-Guided Concept Bottleneck Models with Minimal Annotationscited by 0
  11. 2026Concise and Logically Consistent Conformal Sets for Neuro-Symbolic Concept-Based Modelscited by 0
  12. 2026A homotopy-type-theoretic generalization of neurosymbolic inferencecited by 0
1 more
  1. 2026Loss Invariance Determines What Concept Layers Encode: Volume Grounding in Echocardiographycited by 0