mlsanityData to 7 Oct 2026

Towards Rigorous Explainability by Feature Attribution

arXiv 2604.15898 ↗First posted 17 Apr 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

For around a decade, non-symbolic methods have been the option of choice when explaining complex machine learning (ML) models. Unfortunately, such methods lack rigor and can mislead human decision-makers. In high-stakes uses of ML, the lack of rigor is especially problematic. One prime example of provable lack of rigor is the adoption of Shapley values in explainable artificial intelligence (XAI), with the tool SHAP being a ubiquitous example. This paper overviews the ongoing efforts towards using rigorous symbolic methods of XAI as an alternative to non-rigorous non-symbolic approaches, concretely for assigning relative feature importance.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 28 of its 67 references that are in the map. First posted on arXiv 17 Apr 2026.

Builds on

28 of 67 references.

References · 28

  1. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  3. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  4. 2013Intriguing properties of neural networkscited by 5,114
  5. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  6. 2018Abduction-Based Explanations for Machine Learning Modelscited by 159
  7. 2017PMLB: A Large Benchmark Suite for Machine Learning Evaluation and Comparisoncited by 139
  8. 2018A Symbolic Approach to Explaining Bayesian Network Classifierscited by 130
  9. 2021On Explaining Random Forests with SATcited by 57
  10. 2020Explaining Naive Bayes and Other Linear Classifiers with Polynomial Time and Delaycited by 56
  11. 2022Logic-Based Explainability in Machine Learningcited by 47
  12. 2021SAT-Based Rigorous Explanations for Decision Listscited by 40
16 more
  1. 2021Explanations for Monotonic Classifierscited by 39
  2. 2021On Efficiently Explaining Graph-Based Classifierscited by 37
  3. 2021On the Complexity of SHAP-Score-Based Explanations: Tractability via Knowledge Compilation and Non-Approximability Resultscited by 32
  4. 2022VeriX: Towards Verified Explainability of Deep Neural Networkscited by 30
  5. 2023The Inadequacy of Shapley Values for Explainabilitycited by 30
  6. 2023Logic for Explainable AIcited by 23
  7. 2023Explainability is NOT a Gamecited by 23
  8. 2025Explaining k-Nearest Neighbors: Abductive and Counterfactual Explanationscited by 10
  9. 2024Rethinking Explainable Machine Learning as Applied Statisticscited by 8
  10. 2024Abductive explanations of classifiers under constraints: Complexity and propertiescited by 7
  11. 2025Space Explanations of Neural Network Classificationcited by 5
  12. 2023A Refutation of Shapley Values for Explainabilitycited by 4
  13. 2023Refutation of Shapley Values for XAI -- Additional Evidencecited by 4
  14. 2024Axiomatic Characterisations of Sample-based Explainerscited by 3
  15. 2024SHAP scores fail pervasively even when Lipschitz succeedscited by 2
  16. 2025The Explanation Game -- Rekindled (Extended Version)cited by 1

Built on by

Not cited yet.