mlsanityData to 7 Oct 2026

Recourse under Model Multiplicity via Argumentative Ensembling (Technical Report)

arXiv 2312.15097 ↗First posted 22 Dec 2023cs.LG cs.AIcited by 15

Not on the map. Why?

Abstract (arXiv)

Model Multiplicity (MM) arises when multiple, equally performing machine learning models can be trained to solve the same prediction task. Recent studies show that models obtained under MM may produce inconsistent predictions for the same input. When this occurs, it becomes challenging to provide counterfactual explanations (CEs), a common means for offering recourse recommendations to individuals negatively affected by models' predictions. In this paper, we formalise this problem, which we name recourse-aware ensembling, and identify several desirable properties which methods for solving it should satisfy. We show that existing ensembling methods, naturally extended in different ways to provide CEs, fail to satisfy these properties. We then introduce argumentative ensembling, deploying computational argumentation to guarantee robustness of CEs to MM, while also accommodating customisable user preferences. We show theoretically and experimentally that argumentative ensembling satisfies properties which the existing methods lack, and that the trade-offs are minimal wrt accuracy.

Key facts. Cited by 15 papers in the map (data through 7 Oct 2026). Builds on 35 of its 65 references that are in the map. First posted on arXiv 22 Dec 2023.

Builds on

35 of 65 references.

References · 35

  1. 2017Counterfactual Explanations without Opening the Black Box: Automated Decisions and the GDPRcited by 1,086
  2. 2019Explaining Machine Learning Classifiers through Diverse Counterfactual Explanationscited by 612
  3. 2020Underspecification Presents Challenges for Credibility in Modern Machine Learningcited by 488
  4. 2018All Models are Wrong, but Many are Useful: Learning a Variable's Importance by Studying an Entire Class of Prediction Models Simultaneouslycited by 364
  5. 2018Actionable Recourse in Linear Classificationcited by 327
  6. 2018Explanations based on the Missing: Towards Contrastive Explanations with Pertinent Negativescited by 261
  7. 2019Predictive Multiplicity in Classificationcited by 162
  8. 2021Towards Robust and Reliable Algorithmic Recoursecited by 101
  9. 2019On the Existence of Simpler Machine Learning Modelscited by 87
  10. 2022Exploring the Whole Rashomon Set of Sparse Decision Treescited by 80
  11. 2017Interpretable Predictions of Tree-based Ensembles via Actionable Feature Tweakingcited by 76
  12. 2020On Counterfactual Explanations under Predictive Multiplicitycited by 73
23 more
  1. 2021Argumentative XAI: A Surveycited by 66
  2. 2021Characterizing Fairness Over the Set of Good Models Under Selective Labelscited by 65
  3. 2021NICE: An Algorithm for Nearest Instance Counterfactual Explanationscited by 62
  4. 2022Robust Counterfactual Explanations for Tree-Based Ensemblescited by 49
  5. 2022Rashomon Capacity: A Metric for Predictive Multiplicity in Classificationcited by 46
  6. 2023Robust Counterfactual Explanations for Neural Networks With Probabilistic Guaranteescited by 45
  7. 2020Empirical observation of negligible fairness-accuracy trade-offs in machine learning for public policycited by 44
  8. 2022Predictive Multiplicity in Probabilistic Classificationcited by 44
  9. 2020Interpreting Neural Networks as Gradual Argumentation Frameworks (Including Proof Appendix)cited by 38
  10. 2021A Survey on the Robustness of Feature Importance and Counterfactual Explanationscited by 38
  11. 2021Consistent Counterfactuals for Deep Modelscited by 37
  12. 2018An Interpretable Model with Globally Consistent Explanations for Credit Riskcited by 33
  13. 2022Formalising the Robustness of Counterfactual Explanations for Neural Networkscited by 24
  14. 2021Selective Ensembles for Consistent Predictionscited by 23
  15. 2023Provably Robust and Plausible Counterfactual Explanations for Neural Networks via Robust Optimisationcited by 20
  16. 2019Variable Importance Clouds: A Way to Explore Variable Importance for the Set of Good Modelscited by 19
  17. 2022Counterfactual Plans under Distributional Ambiguitycited by 19
  18. 2022Robust Bayesian Recoursecited by 18
  19. 2022Reconciling Individual Probability Forecastscited by 14
  20. 2022Explaining Random Forests using Bipolar Argumentation and Markov Networks (Technical Report)cited by 9
  21. 2022Forecasting Argumentation Frameworkscited by 5
  22. 2022Rediscovering Argumentation Principles Utilizing Collective Attackscited by 2
  23. 2023Consistent Explanations in the Face of Model Indeterminacy via Ensemblingcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 5; 2025: 7; 2026: 3.

15 papers.

Sort:

By area:

  1. 2024Robust Counterfactual Explanations in Machine Learning: A Surveycited by 29
  2. 2024Predictive Multiplicity of Knowledge Graph Embeddings in Link Predictioncited by 9
  3. 2024Interval Abstractions for Robust Counterfactual Explanationscited by 7
  4. 2024Rigorous Probabilistic Guarantees for Robust Counterfactual Explanationscited by 6
  5. 2025RobustX: Robust Counterfactual Explanations Made Easycited by 6
  6. 2025Representation Consistency for Accurate and Coherent LLM Answer Aggregationcited by 5
  7. 2024Learning-Augmented Robust Algorithmic Recoursecited by 4
  8. 2025Counterfactual Explanations for Model Ensembles Using Entropic Risk Measurescited by 4
  9. 2025Robust Counterfactual Explanations under Model Multiplicity Using Multi-Objective Optimizationcited by 3
  10. 2025Optimal Robust Recourse with Lp-Bounded Model Changecited by 3
  11. 2026ElliCE: Efficient and Provably Robust Algorithmic Recourse via the Rashomon Setscited by 2
  12. 2025Argumentative Ensembling for Robust Recourse under Model Multiplicitycited by 1
3 more
  1. 2025Synthesising Counterfactual Explanations via Label-Conditional Gaussian Mixture Variational Autoencoderscited by 1
  2. 2026Density-Guided Robust Counterfactual Explanations on Tabular Data under Model Multiplicitycited by 0
  3. 2026Target-confidence Recourse Using tSeTlin machines: TRUSTcited by 0