mlsanityData to 7 Oct 2026

Reckoning with the Disagreement Problem: Explanation Consensus as a Training Objective

arXiv 2303.13299 ↗First posted 23 Mar 2023cs.LG cs.AIcited by 6

Not on the map. Why?

Abstract (arXiv)

As neural networks increasingly make critical decisions in high-stakes settings, monitoring and explaining their behavior in an understandable and trustworthy manner is a necessity. One commonly used type of explainer is post hoc feature attribution, a family of methods for giving each feature in an input a score corresponding to its influence on a model's output. A major limitation of this family of explainers in practice is that they can disagree on which features are more important than others. Our contribution in this paper is a method of training models with this disagreement problem in mind. We do this by introducing a Post hoc Explainer Agreement Regularization (PEAR) loss term alongside the standard term corresponding to accuracy, an additional term that measures the difference in feature attribution between a pair of explainers. We observe on three datasets that we can train a model with this loss term to improve explanation consensus on unseen data, and see improved consensus between explainers other than those used in the loss term. We examine the trade-off between improved consensus and model performance. And finally, we study the influence our method has on feature attribution explanations.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 31 of its 42 references that are in the map. First posted on arXiv 23 Mar 2023.

Builds on

31 of 42 references.

References · 31

  1. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  2. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  3. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  4. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  5. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  6. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,348
  7. 2017Learning Important Features Through Propagating Activation Differencescited by 1,454
  8. 2017Towards A Rigorous Science of Interpretable Machine Learningcited by 1,400
  9. 2017SmoothGrad: removing noise by adding noisecited by 1,097
  10. 2020Concept Bottleneck Modelscited by 1,022
  11. 2021Revisiting Deep Learning Models for Tabular Datacited by 725
  12. 2016The Mythos of Model Interpretabilitycited by 716
19 more
  1. 2019TabNet: Attentive Interpretable Tabular Learningcited by 640
  2. 2021Tabular Data: Deep Learning is Not All You Needcited by 530
  3. 2014OpenML: networked science in machine learningcited by 523
  4. 2019Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methodscited by 440
  5. 2021SAINT: Improved Neural Networks for Tabular Data via Row Attention and Contrastive Pre-Trainingcited by 310
  6. 2017Right for the Right Reasons: Training Differentiable Models by Constraining their Explanationscited by 293
  7. 2020Fast Differentiable Sorting and Rankingcited by 196
  8. 2020Evaluating and Aggregating Feature-based Model Explanationscited by 185
  9. 2022The Disagreement Problem in Explainable Machine Learning: A Practitioner's Perspectivecited by 156
  10. 2019Interpretations are useful: penalizing explanations to align neural networks with prior knowledgecited by 144
  11. 2019Analysis of Explainers of Black Box Deep Neural Networks for Computer Vision: A Surveycited by 93
  12. 2022Which Explanation Should I Choose? A Function Approximation Perspective to Characterizing Post Hoc Explanationscited by 77
  13. 2014Falling Rule Listscited by 58
  14. 2021How can I choose an explainer? An Application-grounded Evaluation of Post-hoc Explanationscited by 58
  15. 2022Can Neural Nets Learn the Same Model Twice? Investigating Reproducibility and Double Descent from the Decision Boundary Perspectivecited by 57
  16. 2022Transfer Learning with Deep Tabular Modelscited by 52
  17. 2022Revisiting Pretraining Objectives for Tabular Deep Learningcited by 43
  18. 2021Partial Order in Chaos: Consensus on Feature Attributions in the Rashomon Setcited by 12
  19. 2019Learning Deep Attribution Priors Based On Prior Knowledgecited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 1; 2024: 1; 2025: 1; 2026: 3.

6 papers.

  1. 2024Why do explanations fail? A typology and discussion on failures in XAIcited by 6
  2. 2025Beyond the Single-Best Model: Rashomon Partial Dependence Profile for Trustworthy Explanations in AutoMLcited by 6
  3. 2023Disagreement amongst counterfactual explanations: How transparency can be deceptivecited by 3
  4. 2026GRANITE: A Generalized Regional Framework for Identifying Agreement in Feature-Based Explanationscited by 1
  5. 2026AGRI-Fidelity: Evaluating the Reliability of Listenable Explanations for Poultry Disease Detectioncited by 0
  6. 2026Position: Explanation Stability Is a Property of the Model Method Pair, Not the Modelcited by 0