CoChat
Sign Up
Cited by 5Ethics and Social Impacts of AIExplainable Artificial Intelligence (XAI)Adversarial Robustness in Machine Learning

Systemizing Multiplicity: The Curious Case of Arbitrariness in Machine Learning

Open access
Ganesh, Prakhar, Taik, Afaf, Farnadi, Golnoosh·2025·Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
DOI 10.1609/aies.v8i2.36609
Open this paper in CoChat — free accountOpen PDF

Abstract

Algorithmic modeling relies on limited information in data to extrapolate outcomes for unseen scenarios, often embedding an element of arbitrariness in its decisions. A perspective on this arbitrariness that has recently gained interest is multiplicity—the study of arbitrariness across a set of "good models", i.e., those likely to be deployed in practice. In this work, we systemize the literature on multiplicity by: (a) formalizing the terminology around model design choices and their contribution to arbitrariness, (b) expanding the definition of multiplicity to incorporate underrepresented forms beyond just predictions and explanations, (c) clarifying the distinction between multiplicity and other lenses of arbitrariness, i.e., uncertainty and variance, and (d) distilling the benefits and potential risks of multiplicity into overarching trends, situating it within the broader landscape of responsible AI. We conclude by identifying open research questions and highlighting emerging trends in this young but rapidly growing area of research.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1609/aies.v8i2.36609

Cites

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Model selection is influenced by conventional choices such as learning rate and arbitrary choices such as random seeds.

    “In addition to such intentional choices, model selection is also influenced by conventional choices, such as the selection of a learning rate, or arbitrary choices, such as the selection of random seeds (Ganesh, Taik, and Farnadi 2025;Cavus, Woźnica, and Biecek 2025).”

    From Explainable AI Isn't Enough! Rethinking Algorithmic Contestability (2026) · cites this paper as Ganesh, Taik, and Farnadi 2025

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

No references could be resolved yet.

Research works printed but not linked: 162

  1. Adebayo, J.; Gilmer, J.; Muelly, M.; Goodfellow, I.; Hardt, M.; and Kim, B. 2018. Sanity checks for saliency maps. Advances in neural information processing systems, 31.
  2. Aïvodji, U.; Arai, H.; Gambs, S.; and Hara, S. 2021. Char- acterizing the risk of fairwashing. Advances in Neural Infor- mation Processing Systems, 34: 14822-14834.
  3. Ali, J.; Lahoti, P.; and Gummadi, K. P. 2021. Accounting for model uncertainty in algorithmic discrimination. In Pro- ceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, 336-345.
  4. Anderson, R. 2016. The Rashomon effect and communica- tion. Canadian Journal of Communication, 41(2): 249-270.
  5. Arnold, C.; Biedebach, L.; Küpfer, A.; and Neunhoeffer, M. 2024. The role of hyperparameters in machine learning models and how to tune them. Political Science Research and Methods, 12(4): 841-848.
  6. Baniecki, H.; Parzych, D.; and Biecek, P. 2024. The gram- mar of interactive explanatory model analysis. Data Mining and Knowledge Discovery, 38(5): 2596-2632.
  7. Barocas, S.; Hardt, M.; and Narayanan, A. 2023. Fairness and Machine Learning: Limitations and Opportunities. MIT Press.
  8. Bayes, T. 1763. An Essay Towards Solving a Problem in the Doctrine of Chances. Philosophical Transactions, 53: 370- 418.
  9. Bertsekas, D. P. 2014. Constrained optimization and La- grange multiplier methods. Academic press.
  10. Biecek, P.; Baniecki, H.; Krzyziński, M.; and Cook, D. 2024. Performance is not enough: the story told by a Rashomon quartet. Journal of Computational and Graphical Statistics, (just-accepted): 1-6.
  11. Biecek, P.; and Samek, W. 2024. Position: Explain to Ques- tion not to Justify. In Forty-first International Conference on Machine Learning.
  12. Birhane, A. 2022. Automating Ambiguity: Challenges and Pitfalls of Artificial Intelligence. Ph.D. thesis, University College Dublin. School of Computer Science.
  13. Black, E. 2022. (Un) Fairness Along the AI Pipeline Prob- lems and Solutions. Ph.D. thesis, Carnegie Mellon Univer- sity.
  14. Black, E.; Gillis, T.; and Hall, Z. Y. 2024. D-hacking. In The 2024 ACM Conference on Fairness, Accountability, and Transparency, 602-615.
  15. Black, E.; Koepke, J. L.; Kim, P.; Barocas, S.; and Hsu, M. 2024. Less discriminatory algorithms. Georgetown Law Journal, 113(1).
  16. Black, E.; Raghavan, M.; and Barocas, S. 2022. Model mul- tiplicity: Opportunities, concerns, and solutions. In Proceed- ings of the 2022 ACM Conference on Fairness, Accountabil- ity, and Transparency, 850-863.
  17. Bommasani, R.; Creel, K. A.; Kumar, A.; Jurafsky, D.; and Liang, P. S. 2022. Picking on the same person: Does al- gorithmic monoculture lead to outcome homogenization? Advances in Neural Information Processing Systems, 35: 3663-3678.
  18. Bonate, P. L. 2006. The art of modeling. Springer.
  19. Boner, Z.; Chen, H.; Semenova, L.; Parr, R.; and Rudin, C. 2024. Using Noise to Infer Aspects of Simplicity Without Learning. In The Thirty-eighth Annual Conference on Neu- ral Information Processing Systems.
  20. Bouthillier, X.; Delaunay, P.; Bronzi, M.; Trofimov, A.; Nichyporuk, B.; Szeto, J.; Mohammadi Sepahvand, N.; Raff, E.; Madan, K.; Voleti, V.; et al. 2021. Accounting for vari- ance in machine learning benchmarks. Proceedings of Ma- chine Learning and Systems, 3: 747-769.
  21. Box, G. E. P.; and Tiao, G. C. 1973. Bayesian inference in statistical analysis. International Statistical Review, 43: 242.
  22. Breiman, L. 2001. Statistical modeling: The two cultures (with comments and a rejoinder by the author). Statistical science, 16(3): 199-231.
  23. Buolamwini, J.; and Gebru, T. 2018. Gender shades: Inter- sectional accuracy disparities in commercial gender classifi- cation. In Conference on fairness, accountability and trans- parency, 77-91. PMLR.
  24. Carlini, N.; Athalye, A.; Papernot, N.; Brendel, W.; Rauber, J.; Tsipras, D.; Goodfellow, I.; Madry, A.; and Kurakin, A. 2019. On evaluating adversarial robustness. arXiv preprint arXiv:1902.06705. · arXiv
  25. Castillo, E.; Mínguez, R.; Castillo, C.; and Cofino, A. S. 2008. Dealing with the multiplicity of solutions of the l1 and l-inf regression models. European journal of operational re- search, 188(2): 460-484.
  26. Cavus, M.; and Biecek, P. 2024a. An Experimental Study on the Rashomon Effect of Balancing Methods in Imbalanced Classification. arXiv preprint arXiv:2405.01557. · arXiv
  27. Cavus, M.; and Biecek, P. 2024b. Investigating the Im- pact of Balancing, Filtering, and Complexity on Predictive Multiplicity: A Data-Centric Perspective. arXiv preprint arXiv:2412.09712. · arXiv
  28. Chen, Z.; Zhang, J. M.; Sarro, F.; and Harman, M. 2023. A comprehensive empirical study of bias mitigation methods for machine learning classifiers. ACM Transactions on Soft- ware Engineering and Methodology, 32(4): 1-30.
  29. Coker, B.; Rudin, C.; and King, G. 2021. A theory of statisti- cal inference for ensuring the robustness of scientific results. Management Science, 67(10): 6174-6197.
  30. Cooper, A. F.; Frankle, J.; and De Sa, C. 2022. Non- Determinism and the Lawlessness of Machine Learning Code. In Proceedings of the 2022 Symposium on Computer Science and Law, 1-8.
  31. Cooper, A. F.; Lee, K.; Choksi, M. Z.; Barocas, S.; De Sa, C.; Grimmelmann, J.; Kleinberg, J.; Sen, S.; and Zhang, B. 2024. Arbitrariness and social prediction: The confound- ing role of variance in fair classification. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 38, 22004-22012.
  32. Cooper, A. F.; Lu, Y.; Forde, J.; and De Sa, C. M. 2021. Hyperparameter optimization is deceiving us, and how to stop it. Advances in Neural Information Processing Systems, 34: 3081-3095.
  33. Coston, A.; Rambachan, A.; and Chouldechova, A. 2021. Characterizing fairness over the set of good models under selective labels. In International Conference on Machine Learning, 2144-2155.
  34. PMLR.
  35. Coupkova, E.; and Boutin, M. 2024. On the Rashomon ratio of infinite hypothesis sets. arXiv preprint arXiv:2404.17746. · arXiv
  36. Creel, K.; and Hellman, D. 2022. The algorithmic leviathan: Arbitrariness, fairness, and opportunity in algo- rithmic decision-making systems. Canadian Journal of Phi- losophy, 52(1): 26-43.
  37. D'Amour, A.; Heller, K.; Moldovan, D.; Adlam, B.; Ali- panahi, B.; Beutel, A.; Chen, C.; Deaton, J.; Eisenstein, J.; Hoffman, M. D.; et al. 2022. Underspecification presents challenges for credibility in modern machine learning. Jour- nal of Machine Learning Research, 23(226): 1-61.
  38. Davis, B.; Anderson, R.; and Walls, J. 2015. Rashomon Ef- fects: Kurosawa, Rashomon and their legacies. Routledge.
  39. De Toni, G.; Teso, S.; Lepri, B.; and Passerini, A. 2024. Time Can Invalidate Algorithmic Recourse. arXiv preprint arXiv:2410.08007. · arXiv
  40. Del Giudice, M. 2024. The prediction-explanation fallacy: a pervasive problem in scientific applications of machine learning. Methodology, 20(1): 22-46.
  41. Dietterich, T. G. 2000. Ensemble methods in machine learn- ing. In International workshop on multiple classifier sys- tems, 1-15. Springer.
  42. Domingos, P. 2000. A unified bias-variance decomposition. In Proceedings of 17th international conference on machine learning, 231-238.
  43. Morgan Kaufmann Stanford.
  44. Dong, J.; and Rudin, C. 2019. Variable importance clouds: A way to explore variable importance for the set of good models. arXiv preprint arXiv:1901.03209. · arXiv
  45. Dong, J.; and Rudin, C. 2020. Exploring the cloud of vari- able importance for the set of all good models. Nature Ma- chine Intelligence, 2(12): 810-824.
  46. Du, A. Y.; Ngo, D. D.; and Wu, Z. S. 2024. Reconcil- ing Model Multiplicity for Downstream Decision Making. arXiv preprint arXiv:2405.19667. · arXiv
  47. Dwork, C.; Hardt, M.; Pitassi, T.; Reingold, O.; and Zemel, R. 2012. Fairness through awareness. In Proceedings of the 3rd innovations in theoretical computer science conference, 214-226.
  48. Eerlings, G.; Vanbrabant, S.; Liesenborgs, J.; Ruiz, G. R.; Vanacken, D.; and Luyten, K. 2024. AI-Spectra: A Vi- sual Dashboard for Model Multiplicity to Enhance In- formed and Transparent Decision-Making. arXiv preprint arXiv:2411.10490. · arXiv
  49. Farnadi, G.; Havaei, M.; and Rostamzadeh, N. 2024. Po- sition: Cracking the Code of Cascading Disparity Towards Marginalized Communities. In Forty-first International Conference on Machine Learning.
  50. Fisher, A.; Rudin, C.; and Dominici, F. 2019. All models are wrong, but many are useful: Learning a variable's impor- tance by studying an entire class of prediction models simul- taneously. Journal of Machine Learning Research, 20(177): 1-81.
  51. Forde, J. Z.; Cooper, A. F.; Kwegyir-Aggrey, K.; De Sa, C.; and Littman, M. 2021. Model selection's disparate im- pact in real-world deep learning applications. arXiv preprint arXiv:2104.00606. · arXiv
  52. Gal, Y. 2016. Uncertainty in deep learning. Ph.D. thesis, University of Cambridge.
  53. Gallego-Posada, J. 2024. Constrained optimization for ma- chine learning: algorithms and applications. Ph.D. thesis, Université de Montréal.
  54. Ganesh, P. 2024. An Empirical Investigation into Bench- marking Model Multiplicity for Trustworthy Machine Learning: A Case Study on Image Classification. In Pro- ceedings of the IEEE/CVF Winter Conference on Applica- tions of Computer Vision, 4488-4497.
  55. Ganesh, P.; Chang, H.; Strobel, M.; and Shokri, R. 2023. On the impact of machine learning randomness on group fairness. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, 1789-1800.
  56. Ganesh, P.; Gohar, U.; Cheng, L.; and Farnadi, G. 2024. Dif- ferent Horses for Different Courses: Comparing Bias Miti- gation Algorithms in ML. arXiv preprint arXiv:2411.11101. · arXiv
  57. Ganesh, P.; Shokri, R.; and Farnadi, G. 2025. Rethinking Hallucinations: Correctness, Consistency, and Prompt Mul- tiplicity. In ICLR 2025 Workshop on Building Trust in Lan- guage Models and Applications.
  58. Geirhos, R.; Jacobsen, J.-H.; Michaelis, C.; Zemel, R.; Bren- del, W.; Bethge, M.; and Wichmann, F. A. 2020. Shortcut learning in deep neural networks. Nature Machine Intelli- gence, 2(11): 665-673.
  59. Gerchick, M.; Jegede, T.; Shah, T.; Gutierrez, A.; Beiers, S.; Shemtov, N.; Xu, K.; Samant, A.; and Horowitz, A. 2023. The devil is in the details: Interrogating values embedded in the allegheny family screening tool. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, 1292-1310.
  60. Gichoya, J. W.; Thomas, K.; Celi, L. A.; Safdar, N.; Baner- jee, I.; Banja, J. D.; Seyyed-Kalantari, L.; Trivedi, H.; and Purkayastha, S. 2023. AI pitfalls and what not to do: mitigat- ing bias in AI. The British Journal of Radiology, 96(1150): 20230023.
  61. Gillis, T. B.; Meursault, V.; and Ustun, B. 2024. Opera- tionalizing the Search for Less Discriminatory Alternatives in Fair Lending. In The 2024 ACM Conference on Fairness, Accountability, and Transparency, 377-387.
  62. Gomez, J. F.; Machado, C.; Paes, L. M.; and Calmon, F. 2024. Algorithmic Arbitrariness in Content Moderation. In The 2024 ACM Conference on Fairness, Accountability, and Transparency, 2234-2253.
  63. Goodfellow, I. J.; Shlens, J.; and Szegedy, C. 2014. Explain- ing and harnessing adversarial examples. arXiv preprint arXiv:1412.6572. · arXiv
  64. Gunasekaran, A.; Mistry, P.; and Chen, M. 2024. Which Explanation Should be Selected: A Method Agnostic Model Class Reliance Explanation for Model and Explanation Mul- tiplicity. SN Computer Science, 5(5): 1-20.
  65. Hamamoto, M.; and Egi, M. 2021. Model-agnostic ensemble-based explanation correction leveraging rashomon effect. In 2021 IEEE symposium series on computational intelligence (SSCI), 01-08. IEEE.
  66. Hamman, F.; Dissanayake, P.; Mishra, S.; Lecue, F.; and Dutta, S. 2024. Quantifying prediction consistency un- der model multiplicity in tabular LLMs. arXiv preprint arXiv:2407.04173. · arXiv
  67. Hamman, F.; Noorani, E.; Mishra, S.; Magazzeni, D.; and Dutta, S. 2023. Robust counterfactual explanations for neu- ral networks with probabilistic guarantees. In International Conference on Machine Learning, 12351-12367.
  68. PMLR.
  69. Hasan, M. G. M. M. 2022. Targeted Generation of Coun- terfactual Examples for Data Augmentation and to Mitigate the Rashomon Effect. Ph.D. thesis, Tennessee Technological University.
  70. Hasan, M. G. M. M.; and Talbert, D. 2022. Mitigating the rashomon effect in counterfactual explanation: A game- theoretic approach. In The International FLAIRS Confer- ence Proceedings, volume 35.
  71. He, Z.; and Shaposhnik, Y. 2023. Visualizing the Implicit Model Selection Tradeoff. Journal of Artificial Intelligence Research, 76: 829-881.
  72. Heider, K. G. 1988. The Rashomon effect: When ethnogra- phers disagree. American Anthropologist, 90(1): 73-81.
  73. Heljakka, A.; Trapp, M.; Kannala, J.; and Solin, A. 2022. Disentangling model multiplicity in deep learning. arXiv preprint arXiv:2206.08890. · arXiv
  74. Hort, M.; Chen, Z.; Zhang, J. M.; Harman, M.; and Sarro, F. 2024. Bias mitigation for machine learning classifiers: A comprehensive survey. ACM Journal on Responsible Com- puting, 1(2): 1-52.
  75. Hsu, H. 2023. Information-Theoretic Tools for Machine Learning Beyond Accuracy. Ph.D. thesis, Harvard Univer- sity Graduate School of Arts and Sciences.
  76. Hsu, H.; Brugere, I.; Sharma, S.; Lecue, F.; and Chen, C.-F. 2024a. RashomonGB: Analyzing the Rashomon Effect and Mitigating Predictive Multiplicity in Gradient Boosting. In The Thirty-eighth Annual Conference on Neural Information Processing Systems.
  77. Hsu, H.; and Calmon, F. 2022. Rashomon capacity: A met- ric for predictive multiplicity in classification. Advances in Neural Information Processing Systems, 35: 28988-29000.
  78. Hüllermeier, E.; and Waegeman, W. 2021. Aleatoric and epistemic uncertainty in machine learning: An introduction to concepts and methods. Machine learning, 110(3): 457- 506.
  79. Islam, R.; Pan, S.; and Foulds, J. R. 2021. Can we obtain fairness for free? In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, 586-596.
  80. Jacobs, A. Z.; and Wallach, H. 2021. Measurement and fair- ness. In Proceedings of the 2021 ACM conference on fair- ness, accountability, and transparency, 375-385.
  81. Jain, S.; Creel, K.; and Wilson, A. 2024. Scarce Resource Allocations That Rely On Machine Learning Should Be Randomized. arXiv preprint arXiv:2404.08592. · arXiv
  82. Jain, S.; Suriyakumar, V.; Creel, K.; and Wilson, A. 2024. Algorithmic Pluralism: A Structural Approach To Equal Op- portunity. In The 2024 ACM Conference on Fairness, Ac- countability, and Transparency, 197-206.
  83. Jain, S.; Wang, M.; Creel, K.; and Wilson, A. 2025. Alloca- tion Multiplicity: Evaluating the Promises of the Rashomon Set. arXiv preprint arXiv:2503.16621. · arXiv
  84. Jiang, J.; Leofante, F.; Rago, A.; and Toni, F. 2024a. Re- course under model multiplicity via argumentative ensem- bling. In Proceedings of the 23rd International Conference on Autonomous Agents and Multiagent Systems, 954-963.
  85. Jiang, J.; Leofante, F.; Rago, A.; and Toni, F. 2024b. Robust Counterfactual Explanations in Machine Learning: A Sur- vey. In Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, IJCAI-24, 8086-8094. International Joint Conferences on Artificial Intelligence Or- ganization. Survey Track.
  86. Kendall, A.; and Gal, Y. 2017. What uncertainties do we need in bayesian deep learning for computer vision? Ad- vances in neural information processing systems, 30.
  87. Kissel, N.; and Mentch, L. 2024. Forward stability and model path selection. Statistics and Computing, 34(2): 82.
  88. Kleinberg, J.; and Raghavan, M. 2021. Algorithmic mono- culture and social welfare. Proceedings of the National Academy of Sciences, 118(22): e2018340118.
  89. Kobylińska, K.; Krzyziński, M.; Machowicz, R.; Adamek, M.; and Biecek, P. 2024. Exploration of the Rashomon Set Assists Trustworthy Explanations for Medical Data. IEEE Journal of Biomedical and Health Informatics, 28(11): 6454-6465.
  90. Kong, E. B.; and Dietterich, T. G. 1995. Error-correcting output coding corrects bias and variance. In Proceedings of the Twelfth International Conference on International Con- ference on Machine Learning, ICML'95, 313-321.
  91. Morgan Kaufmann Publishers Inc. ISBN 1558603778.
  92. Kowal, D. R. 2022. Bayesian subset selection and vari- able importance for interpretable prediction and classifica- tion. Journal of Machine Learning Research, 23(108): 1-38.
  93. Kulynych, B.; Hsu, H.; Troncoso, C.; and Calmon, F. P. 2023. Arbitrary decisions are a hidden cost of differentially private training. In Proceedings of the 2023 ACM Confer- ence on Fairness, Accountability, and Transparency, 1609- 1623.
  94. Laberge, G.; Pequignot, Y.; Mathieu, A.; Khomh, F.; and Marchand, M. 2023. Partial order in chaos: consensus on feature attributions in the rashomon set. Journal of Machine Learning Research, 24(364): 1-50.
  95. Leofante, F.; Botoeva, E.; and Rajani, V. 2023. Counterfac- tual explanations and model multiplicity: a relational verifi- cation view. In Proceedings of the International Conference on Principles of Knowledge Representation and Reasoning, volume 19, 763-768.
  96. Leventi-Peetz, A.-M.; and Weber, K. 2022. Rashomon effect and consistency in explainable artificial intelligence (XAI). In Proceedings of the Future Technologies Conference, 796- 808. Springer.
  97. Li, S.; Barnard, A. S.; and Deng, Q. 2024. Practical attribution guidance for rashomon sets. arXiv preprint arXiv:2407.18482. · arXiv
  98. Li, S.; Wang, R.; Deng, Q.; and Barnard, A. S. 2024. Explor- ing the cloud of feature interaction scores in a Rashomon set. In The Twelfth International Conference on Learning Rep- resentations.
  99. Lipton, Z. C. 2018. The mythos of model interpretability: In machine learning, the concept of interpretability is both important and slippery. Queue, 16(3): 31-57.
  100. Liu, J.; Zhong, C.; Li, B.; Seltzer, M.; and Rudin, C. 2022. FasterRisk: Fast and accurate interpretable risk scores. Advances in Neural Information Processing Systems, 35: 17760-17773.
  101. Ma, S.; Schreiner, P. J.; Seaquist, E. R.; Ugurbil, M.; Zmora, R.; and Chow, L. S. 2020. Multiple predictively equivalent risk models for handling missing data at time of prediction: With an application in severe hypoglycemia risk prediction for type 2 diabetes. Journal of biomedical informatics, 103: 103379.
  102. Marx, C.; Calmon, F.; and Ustun, B. 2020. Predictive multi- plicity in classification. In International Conference on Ma- chine Learning, 6765-6774. PMLR.
  103. Mata, K.; Kanamori, K.; and Arimura, H. 2022. Comput- ing the Collection of Good Models for Rule Lists. In Proc. the 18th International Conference on Machine Learning and Data Mining (MLDM 2022).
  104. McCoy, R. T.; Min, J.; and Linzen, T. 2020. BERTs of a feather do not generalize together: Large variability in gen- eralization across models with similar test set performance. In Proceedings of the Third BlackboxNLP Workshop on Ana- lyzing and Interpreting Neural Networks for NLP, 217-227.
  105. Meyer, A. P.; Albarghouthi, A.; and D'Antoni, L. 2023. The dataset multiplicity problem: How unreliable data impacts predictions. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, 193-204.
  106. Meyer, A. P.; Kim, Y.-S.; Albarghouthi, A.; and D'Antoni, L. 2024. Perceptions of the Fairness Impacts of Multiplicity in Machine Learning. arXiv preprint arXiv:2409.12332. · arXiv
  107. Milani Fard, M.; Cormier, Q.; Canini, K.; and Gupta, M. 2016. Launch and iterate: Reducing prediction churn. Ad- vances in Neural Information Processing Systems, 29.
  108. Møllersen, K.; and Holsbø, E. 2023. Accounting for multi- plicity in machine learning benchmark performance. arXiv preprint arXiv:2303.07272. · arXiv
  109. Molnar, C.; König, G.; Herbinger, J.; Freiesleben, T.; Dandl, S.; Scholbeck, C. A.; Casalicchio, G.; Grosse-Wentrup, M.; and Bischl, B. 2020. General pitfalls of model-agnostic in- terpretation methods for machine learning models. In In- ternational Workshop on Extending Explainable AI Beyond Deep Models and Classifiers, 39-68. Springer.
  110. Müller, S.; Toborek, V.; Beckh, K.; Jakobs, M.; Bauckhage, C.; and Welke, P. 2023. An empirical evaluation of the Rashomon effect in explainable machine learning. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases, 462-478. Springer.
  111. Nguyen, V.-L.; Shaker, M. H.; and Hüllermeier, E. 2022. How to measure uncertainty in uncertainty sampling for ac- tive learning. Machine Learning, 111(1): 89-122.
  112. Niculescu-Mizil, A.; and Caruana, R. 2005. Predicting good probabilities with supervised learning. In Proceedings of the 22nd international conference on Machine learning, 625- 632.
  113. Okazaki, K.; Ribeiro, T.; Lo, K.-Y.; Sakuma, J.; and In- oue, K. 2024. Data-driven Analysis of Domain Specificity for Explainable Session-based Recommendation System. In Proceedings of the National Conference of the Japanese So- ciety for Artificial Intelligence, No. 38 (2024), 2Q5IS104- 2Q5IS104. Japan Society for Artificial Intelligence.
  114. Paes, L. M.; Cruz, R.; Calmon, F. P.; and Diaz, M. 2023. On the inevitability of the Rashomon effect. In 2023 IEEE In- ternational Symposium on Information Theory (ISIT), 549- 554. IEEE.
  115. Pawelczyk, M.; Broelemann, K.; and Kasneci, G. 2020. On counterfactual explanations under predictive multiplicity. In Conference on Uncertainty in Artificial Intelligence, 809- 818.
  116. PMLR.
  117. Pecher, B.; Srba, I.; and Bielikova, M. 2024. A survey on stability of learning with limited labelled data and its sen- sitivity to the effects of randomness. ACM Computing Sur- veys, 57(1): 1-40.
  118. Perry, R.; and Zarsky, T. 2015. 'May the Odds Be Ever in Your Favor': Lotteries in Law. Alabama Law Review, 66: 1035-1098.
  119. Picard, D. 2021. Torch. manual seed (3407) is all you need: On the influence of random seeds in deep learning architec- tures for computer vision. arXiv preprint arXiv:2109.08203. · arXiv
  120. Poiret, C.; Grigis, A.; Thomas, J.; and Noulhiane, M. 2023. Can we Agree? On the Rash\= omon Effect and the Reliability of Post-Hoc Explainable AI. arXiv preprint arXiv:2308.07247. · arXiv
  121. Ponomareva, N.; Hazimeh, H.; Kurakin, A.; Xu, Z.; Deni- son, C.; McMahan, H. B.; Vassilvitskii, S.; Chien, S.; and Thakurta, A. G. 2023. How to dp-fy ml: A practical guide to machine learning with differential privacy. Journal of Ar- tificial Intelligence Research, 77: 1113-1201.
  122. Rawal, K.; Kamar, E.; and Lakkaraju, H. 2021. Algorithmic Recourse in the Wild: Understanding the Impact of Data and Model Shifts. arxiv.
  123. Rebuffi, S.-A.; Gowal, S.; Calian, D. A.; Stimberg, F.; Wiles, O.; and Mann, T. A. 2021. Data augmentation can improve robustness. Advances in Neural Information Processing Sys- tems, 34: 29935-29948.
  124. Roth, A.; Tolbert, A.; and Weinstein, S. 2023. Reconciling Individual Probability Forecasts. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Trans- parency, 101-110.
  125. Rudin, C. 2019. Stop explaining black box machine learning models for high stakes decisions and use interpretable mod- els instead. Nature machine intelligence, 1(5): 206-215.
  126. Rudin, C.; Zhong, C.; Semenova, L.; Seltzer, M.; Parr, R.; Liu, J.; Katta, S.; Donnelly, J.; Chen, H.; and Boner, Z. 2024. Amazing things come from having many good mod- els. arXiv preprint arXiv:2407.04846. · arXiv
  127. Sellam, T.; Yadlowsky, S.; Tenney, I.; Wei, J.; Saphra, N.; D'Amour, A.; Linzen, T.; Bastings, J.; Turc, I. R.; Eisen- stein, J.; et al. 2022. The MultiBERTs: BERT Reproduc- tions for Robustness Analysis. In International Conference on Learning Representations.
  128. Semenova, L. 2024. In Pursuit of Simplicity: The Role of the Rashomon Effect for Informed Decision Making. Ph.D. thesis, Duke University.
  129. Semenova, L.; Chen, H.; Parr, R.; and Rudin, C. 2024. A path to simpler models starts with noise. Advances in neural information processing systems, 36.
  130. Semenova, L.; Rudin, C.; and Parr, R. 2022. On the exis- tence of simpler machine learning models. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, 1827-1858.
  131. Shahin Shamsabadi, A.; Yaghini, M.; Dullerud, N.; Wyllie, S.; Aïvodji, U.; Alaagib, A.; Gambs, S.; and Papernot, N. 2022. Washing the unwashable: On the (im) possibility of fairwashing detection. Advances in Neural Information Pro- cessing Systems, 35: 14170-14182.
  132. Sharma, R.; Redyuk, S.; Mukherjee, S.; Sipka, A.; Vollmer, S.; and Selby, D. 2024. X Hacking: The Threat of Misguided AutoML. arXiv preprint arXiv:2401.08513. · arXiv
  133. Simson, J.; Pfisterer, F.; and Kern, C. 2024. One model many scores: Using multiverse analysis to prevent fairness hacking and evaluate the influence of model design decisions. In The 2024 ACM Conference on Fairness, Accountability, and Transparency, 1305-1320.
  134. Sinha, A.; Malo, P.; and Deb, K. 2017. A review on bilevel optimization: From classical to evolutionary approaches and applications. IEEE transactions on evolutionary computa- tion, 22(2): 276-295.
  135. Smith, F. B.; Kossen, J.; Trollope, E.; van der Wilk, M.; Foster, A.; and Rainforth, T. 2024. Rethinking Aleatoric and Epistemic Uncertainty. In NeurIPS 2024 Workshop on Bayesian Decision-making and Uncertainty.
  136. Sokol, K.; Kull, M.; Chan, J.; and Salim, F. 2024. Cross- model Fairness: Empirical Study of Fairness and Ethics Un- der Model Multiplicity. ACM Journal on Responsible Com- puting, 1(3): 1-27.
  137. Visual Communi- cation in Cultural Media. The Rashomon Effect in the Im- age Globalization Paradigm. Review of Artistic Education, 26(1): 177-185.
  138. Sreedharan, S.; Kambhampati, S.; et al. 2018. Han- dling model uncertainty and multiplicity in explanations via model reconciliation. In Proceedings of the International Conference on Automated Planning and Scheduling, vol- ume 28, 518-526.
  139. Sun, Y.; Wang, T.; and Rudin, C. 2024. Improving Decision Sparsity. In The Thirty-eighth Annual Conference on Neural Information Processing Systems.
  140. Tan, H. 2024. Evaluating Explanation Robustness to Model Pruning. In 2024 International Joint Conference on Neural Networks (IJCNN), 1-8. IEEE.
  141. Teney, D.; Peyrard, M.; and Abbasnejad, E. 2022. Predict- ing is not understanding: Recognizing and addressing under- specification in machine learning. In European Conference on Computer Vision, 458-476. Springer.
  142. Tindale, C. 2016. Collateral beliefs and the Rashomon ef- fect. Cogency, 8(2): 125-145.
  143. Trabasso, T. 2018. Whose memory is it? The social context of remembering. In Memory for everyday and emotional events, 429-443. Routledge.
  144. Tramer, F.; Carlini, N.; Brendel, W.; and Madry, A. 2020. On adaptive attacks to adversarial example defenses. Advances in neural information processing systems, 33: 1633-1645.
  145. Tran, D.; Snoek, J.; and Lakshminarayanan, B. 2020. Practi- cal uncertainty estimation and out-of-distribution robustness in deep learning. NeurIPS Tutorial, Google Brain.
  146. Tulabandhula, T.; and Rudin, C. 2014. Robust Optimization using Machine Learning for Uncertainty Sets. ECML/PKDD 2014, 121.
  147. Vaicenavicius, J.; Widmann, D.; Andersson, C.; Lindsten, F.; Roll, J.; and Schön, T. 2019. Evaluating model calibration in classification. In The 22nd international conference on artificial intelligence and statistics, 3459-3467.
  148. PMLR.
  149. Wang, A.; Kapoor, S.; Barocas, S.; and Narayanan, A. 2024. Against predictive optimization: On the legitimacy of decision-making algorithms that optimize predictive accu- racy. ACM Journal on Responsible Computing, 1(1): 1-45.
  150. Wang, A.; and Russakovsky, O. 2021. Directional bias am- plification. In International Conference on Machine Learn- ing, 10882-10893. PMLR.
  151. Watson-Daniels, J. 2024. The Roads Not Taken: Model Mul- tiplicity in Machine Learning. Ph.D. thesis, Harvard Univer- sity Graduate School of Arts and Sciences.
  152. Watson-Daniels, J.; Barocas, S.; Hofman, J. M.; and Chouldechova, A. 2023. Multi-target multiplicity: Flexibil- ity and fairness in target specification under resource con- straints. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, 297-311.
  153. Watson-Daniels, J.; Calmon, F. d. P.; D'Amour, A.; Long, C.; Parkes, D. C.; and Ustun, B. 2024. Predictive churn with the set of good models. arXiv preprint arXiv:2402.07745. · arXiv
  154. Wu, B.; Chen, J.; Cai, D.; He, X.; and Gu, Q. 2021. Do wider neural networks really help adversarial robustness? Advances in Neural Information Processing Systems, 34: 7054-7067.
  155. Wu, F.; Black, E.; and Chandrasekaran, V. 2024. Genera- tive monoculture in large language models. arXiv preprint arXiv:2407.02209. · arXiv
  156. Xin, R.; Zhong, C.; Chen, Z.; Takagi, T.; Seltzer, M.; and Rudin, C. 2022. Exploring the whole rashomon set of sparse decision trees. Advances in neural information processing systems, 35: 14071-14084.
  157. Yang, Y.; and Loog, M. 2016. Active learning using uncer- tainty information. In 2016 23rd International Conference on Pattern Recognition (ICPR), 2646-2651. IEEE.
  158. Zhang, P.; Wang, H.; Naik, N.; Xiong, C.; et al. 2020. DIME: An information-theoretic difficulty measure for AI datasets. In NeurIPS 2020 Workshop: Deep Learning through Infor- mation Geometry.
  159. Zhang, Y.; Khanduri, P.; Tsaknakis, I.; Yao, Y.; Hong, M.; and Liu, S. 2024. An Introduction to Bilevel Optimization: Foundations and applications in signal processing and ma- chine learning. IEEE Signal Processing Magazine, 41(1): 38-59.
  160. Zhong, C. 2024. Interpretability and Multiplicity: a Path to Trustworthy Machine Learning. Ph.D. thesis, Duke Univer- sity.
  161. Zhong, C.; Chen, Z.; Liu, J.; Seltzer, M.; and Rudin, C. 2024. Exploring and interacting with the set of good sparse gen- eralized additive models. Advances in neural information processing systems, 36.
  162. Zuin, G.; Marcelino, F.; Borges, L.; Couto, J.; Jorge, V.; Laurindo, M.; Barcelos, G.; Cunha, M.; Alvarenga, V.; Ro- drigues, H.; et al. 2021. Predicting Heating Sliver in Duplex Stainless Steels Manufacturing through Rashomon Sets. In 2021 International Joint Conference on Neural Networks (IJCNN), 1-8. IEEE.

Cited by

Sorted by citations

Authors
Ganesh, P., Taik, A., & Farnadi, G.
Affiliations
Mila - Quebec Artificial Intelligence Institute (CA)
ORCID
Venue
Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society
Publisher
Association for the Advancement of Artificial Intelligence (AAAI)
Published
2025-10-15
Volume
8
Issue
2
Pages
1032-1048
ISSN
3065-8365 (ISSN-L)
DOI
10.1609/aies.v8i2.36609
OpenAlex
W4415230897
Cited by
5
References
0
Citations by year
2026: 5
FWCI
3.68
Type
conference paper
Language
en
Open access
Locations
  • Proceedings of the AAAI/ACM Conference on AI Ethics and Society (publishedVersion)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • arXiv (Cornell University) (cc-by)
Funders
  • Canadian Institute for Advanced Research
  • Compute Canada
  • Natural Sciences and Engineering Research Council of Canada
  • Fonds de recherche du Québec – Nature et technologies
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
  • Physical Sciences › Computer Science › Artificial Intelligence › Adversarial Robustness in Machine Learning
Keywords
multiplicity, arbitrariness, extrapolation, uncertainty, variance, responsible AI, explainability, algorithmic fairness, model robustness

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review