mlsanityData to 14 Oct 2026

CUBE: Contrastive Understanding by Balanced Experiments

arXiv 2509.10825 ↗First posted 13 Sep 2025cs.LG cs.AI stat.MLcited by 0

Not on the map. Why?

Abstract (arXiv)

We introduce CUBE, Contrastive Understanding by Balanced Experiments, to estimate main effects and pairwise Banzhaf interactions over a fixed two-state probe space using shared balanced queries. Each design cancels constant and odd-order contamination from pairwise estimates, while randomization ensures unbiasedness and explicit alias probabilities yield exact mean squared errors without sparsity assumptions or order truncation. Matching risk bounds establish asymptotic minimax optimality within a specified class of nonadaptive, equal-weight, reversal-symmetric balanced contrast estimators, as admissible budgets grow while remaining small relative to the full probe space. Experiments reveal different interaction-specific risks despite identical output distributions and low-order effects, and different sign risks despite equal mean squared errors. Across evaluated tabular models, CUBE generally outperforms several sampling-based estimators, while Faith-Banzhaf-2 achieves the lowest mean error at the largest tested budgets.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 76 of its 181 references that are in the map. First posted on arXiv 13 Sep 2025.

Builds on

76 of 181 references.

References · 76

  1. 2016Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localizationcited by 6,733
  2. 2017A Unified Approach to Interpreting Model Predictionscited by 6,555
  3. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,070
  4. 2017Axiomatic Attribution for Deep Networkscited by 3,384
  5. 2015Learning Deep Features for Discriminative Localizationcited by 2,436
  6. 2017Understanding Black-box Predictions via Influence Functionscited by 1,741
  7. 2017Learning Important Features Through Propagating Activation Differencescited by 1,458
  8. 2017Interpretability Beyond Feature Attribution: Quantitative Testing with Concept Activation Vectors (TCAV)cited by 1,319
  9. 2018Sanity Checks for Saliency Mapscited by 1,156
  10. 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,089
  11. 2018This Looks Like That: Deep Learning for Interpretable Image Recognitioncited by 737
  12. 2017Network Dissection: Quantifying Interpretability of Deep Visual Representationscited by 708
38 more
  1. 2019Explaining Machine Learning Classifiers through Diverse Counterfactual Explanationscited by 614
  2. 2017Interpretable Explanations of Black Boxes by Meaningful Perturbationcited by 612
  3. 2018A Benchmark for Interpretability Methods in Deep Neural Networkscited by 476
  4. 2019The many Shapley values for model explanationcited by 344
  5. 2017The (Un)reliability of saliency methodscited by 333
  6. 2018Actionable Recourse in Linear Classificationcited by 328
  7. 2020Neural Additive Models: Interpretable Machine Learning with Neural Netscited by 319
  8. 2019On the (In)fidelity and Sensitivity for Explanationscited by 310
  9. 2019Towards Efficient Data Valuation Based on the Shapley Valuecited by 281
  10. 2018Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negativescited by 262
  11. 2019Understanding Deep Networks via Extremal Perturbations and Smooth Maskscited by 259
  12. 2018Learning to Explain: An Information-Theoretic Perspective on Model Interpretationcited by 248
  13. 2019FACE: Feasible and Actionable Counterfactual Explanationscited by 233
  14. 2020Explaining by Removing: A Unified Framework for Model Explanationcited by 218
  15. 2016Visualizing the Effects of Predictor Variables in Black Box Supervised Learning Modelscited by 213
  16. 2019Explaining individual predictions when features are dependent: More accurate approximations to Shapley valuescited by 208
  17. 2019Full-Gradient Representation for Neural Network Visualizationcited by 191
  18. 2018Representer Point Selection for Explaining Deep Neural Networkscited by 187
  19. 2019Feature relevance quantification in explainable AI: A causal problemcited by 167
  20. 2020Improving KernelSHAP: Practical Shapley Value Estimation via Linear Regressioncited by 154
  21. 2018Explaining Image Classifiers by Counterfactual Generationcited by 144
  22. 2019The Shapley Taylor Interaction Indexcited by 143
  23. 2022Data Banzhaf: A Robust Data Valuation Framework for Machine Learningcited by 132
  24. 2018Net2Vec: Quantifying and Explaining how Concepts are Encoded by Filters in Deep Neural Networkscited by 129
  25. 2019Asymmetric Shapley values: incorporating causal knowledge into model-agnostic explainabilitycited by 114
  26. 2020Causal Shapley Values: Exploiting Causal Knowledge to Explain Individual Predictions of Complex Modelscited by 107
  27. 2020Explaining Explanations: Axiomatic Feature Interactions for Deep Networkscited by 103
  28. 2021Sampling Permutations for Shapley Value Estimationcited by 102
  29. 2021NODE-GAM: Neural Generalized Additive Model for Interpretable Deep Learningcited by 88
  30. 2020GAMI-Net: An Explainable Neural Network based on Generalized Additive Models with Structured Interactionscited by 82
  31. 2017Detecting Statistical Interactions from Neural Network Weightscited by 80
  32. 2020How does this interaction affect me? Interpretable attribution for feature interactionscited by 73
  33. 2022Faith-Shap: The Faithful Shapley Interaction Indexcited by 69
  34. 2018Beyond Word Importance: Contextual Decomposition to Extract Interactions from LSTMscited by 67
  35. 2018Hierarchical interpretations for neural network predictionscited by 66
  36. 2022From Shapley Values to Generalized Additive Models and backcited by 60
  37. 2013Bounding the Estimation Error of Sampling-based Shapley Value Approximationcited by 58
  38. 2023SHAP-IQ: Unified Approximation of any-order Shapley Interactionscited by 51

50 most cited of 76 All 76 (JSON)

Built on by

Not cited yet.