mlsanityData to 14 Oct 2026

Efficient Computation of Shap Explanation Scores for Neural Network Classifiers via Knowledge Compilation

arXiv 2303.06516 ↗First posted 11 Mar 2023cs.AI cs.DB cs.LGcited by 3

Not on the map. Why?

Abstract (arXiv)

The use of Shap scores has become widespread in Explainable AI. However, their computation is in general intractable, in particular when done with a black-box classifier, such as neural network. Recent research has unveiled classes of open-box Boolean Circuit classifiers for which Shap can be computed efficiently. We show how to transform binary neural networks into those circuits for efficient Shap computation.We use logic-based knowledge compilation techniques. The performance gain is huge, as we show in the light of our experiments.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 17 of its 39 references that are in the map. First posted on arXiv 11 Mar 2023.

Builds on

17 of 39 references.

References · 17

  1. 2017A Unified Approach to Interpreting Model Predictionscited by 6,555
  2. 2018A Survey Of Methods For Explaining Black Box Modelscited by 973
  3. 2020Binary Neural Networks: A Surveycited by 190
  4. 2020Explainable Deep Learning: A Field Guide for the Uninitiatedcited by 127
  5. 2020On The Reasons Behind Decisionscited by 100
  6. 2020On the Tractability of SHAP Explanationscited by 92
  7. 2017Verifying Properties of Binarized Deep Neural Networkscited by 79
  8. 2021A comprehensive review of Binary Neural Networkcited by 45
  9. 2020On Tractable Representations of Binary Neural Networkscited by 34
  10. 2021On the Complexity of SHAP-Score-Based Explanations: Tractability via Knowledge Compilation and Non-Approximability Resultscited by 32
  11. 2020Causality-based Explanation of Classification Outcomescited by 25
  12. 2016SDDs are Exponentially More Succinct than OBDDscited by 19
5 more
  1. 2019Smoothing Structured Decomposable Circuitscited by 14
  2. 2020Declarative Approaches to Counterfactual Explanations for Classificationcited by 14
  3. 2018On the Relative Succinctness of Sentential Decision Diagramscited by 6
  4. 2017On Compiling DNNFs without Determinismcited by 4
  5. 2020Variable Shift SDD: A More Succinct Sentential Decision Diagramcited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2023: 1; 2024: 1; 2025: 0; 2026: 1.

3 papers.

  1. 2023Attribution-Scores in Data Management and Explainable Machine Learningcited by 2
  2. 2024The Distributional Uncertainty of the SHAP score in Explainable Machine Learningcited by 1
  3. 2026Reconciling Consistency-Based Diagnosis with Actual-Causality-Based Explanationscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026xDNN(ASP): Explanation Generation System for Deep Neural Networks powered by Answer Set Programming
  2. 2025SHAP Meets Tensor Networks: Provably Tractable Explanations with Parallelism
  3. 2025Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoning
  4. 2026Scaling the Explanation of Multi-Class Bayesian Network Classifiers
  5. 2023COOL: A Constraint Object-Oriented Logic Programming Language and its Neural-Symbolic Compilation System
  6. 2026NeuroRule: Making Black-Box Neural Networks Explainable through Rule-set Evolution
  7. 2026To Neuro-Symbolic Classification and Beyond by Compiling Description Logic Ontologies to Probabilistic Circuits
  8. 2024The Computational Complexity of Circuit Discovery for Inner Interpretability