CoChat
Sign Up
PreprintCited by 61Machine Learning and AlgorithmsMachine Learning and Data ClassificationDomain Adaptation and Few-Shot Learning

Consistent Estimators for Learning to Defer to an Expert

Open access
Mozannar, Hussein, Sontag, David·2020·arXiv
DOI 10.48550/arxiv.2006.01862arXiv 2006.01862
Open this paper in CoChat — free accountOpen PDF

Abstract

Learning algorithms are often used in conjunction with expert decision makers in practical scenarios, however this fact is largely ignored when designing these algorithms. In this paper we explore how to learn predictors that can either predict or choose to defer the decision to a downstream expert. Given only samples of the expert's decisions, we give a procedure based on learning a classifier and a rejector and analyze it theoretically. Our approach is based on a novel reduction to cost sensitive learning where we give a consistent surrogate loss for cost sensitive learning that generalizes the cross entropy loss. We show the effectiveness of our approach on a variety of experimental tasks.

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

DataCite · 10.48550/arxiv.2006.01862

Cites

6 of 6 references verified0 citation errors · 0 not checkable

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?

  • The optimization problem in (5) is amenable to procedures inspired by learning to defer that can outperform the post-hoc procedure proposed.

    “Finally, the optimization problem in (5) is amenable to procedures inspired by learning to defer (Mozannar and Sontag 2020) that can outperform the post-hoc procedure proposed.”

    From When to Show a Suggestion? Integrating Human Feedback in AI-Assisted Programming (2024) · cites this paper as Mozannar and Sontag 2020

    The Problem Formulation (p. 4) presents learning to defer as a cost-sensitive optimization problem, and the Conclusion (p. 21) reports better accuracy than confidence-score baselines. The source does not assess the citing paper’s optimization problem (5) or its particular proposed post-hoc procedure, so it supports the general approach and possibility of improvement, not that specific comparison. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • The state-of-the-art approach to managing assignments in human-AI collaboration is learning to defer.

    “The state-of-the-art approach to managing assignments in HAIC is learning to defer (L2D)4–13, in which algorithms choose whether to assign a task to a human expert or an ML model, leveraging their complementary strengths.”

    From A benchmarking framework and dataset for learning to defer in human-AI decision-making (2025) · cites this paper as [8]

    The Introduction (p. 1) describes learning a classifier and a rejector that decides whether the model or a human expert should predict, using their complementary strengths. The Related Work section (p. 2) calls expert deferral an area of recent work, but the paper does not establish learning to defer as the state-of-the-art approach to assignments across human–AI collaboration. (LLM, conf=0.82)

    Experimental — accuracy not yet measured

  • Learning to Defer algorithms improve human-AI collaboration by deferring decisions to human experts when they are likely to be more accurate than the AI model.

    “Learning to Defer (L2D) algorithms improve human-AI collaboration by deferring decisions to human experts when they are likely to be more accurate than the AI model.”

    From A benchmarking framework and dataset for learning to defer in human-AI decision-making (2025) · cites this paper as [8]

  • For α=0, the probability of error is the same for all instances within a class, which is equivalent to class-dependent label noise.

    “For α=0, the probability of error is the same for all instances within a class, which is equivalent to class-dependent label noise, as used in previous L2D testing4,8,11.”

    From A benchmarking framework and dataset for learning to defer in human-AI decision-making (2025) · cites this paper as [8]

  • The paper chose to create a total of 50 experts because this far surpasses the numbers previously considered in learning-to-defer research.

    “We chose to create a total of 50 experts as this far surpasses the numbers previously considered in L2D research4–13.”

    From A benchmarking framework and dataset for learning to defer in human-AI decision-making (2025) · cites this paper as [8]

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

Flagged first, then by citations

Research works printed but not linked: 47

  1. [BCG + 19] David Berthelot, Nicholas Carlini, Ian Goodfellow, Nicolas Papernot, Avital Oliver, and Colin Raffel. Mixmatch: A holistic approach to semi-supervised learning. arXiv preprint arXiv:1905.02249 , 2019. · arXiv
  2. [BDLSS12] Shai Ben-David, David Loker, Nathan Srebro, and Karthik Sridharan. Minimizing the misclassification error rate using a surrogate convex loss. arXiv preprint arXiv:1206.6442 , 2012. · arXiv
  3. [BJM06] Peter L Bartlett, Michael I Jordan, and Jon D McAuliffe. Convexity, classification, and risk bounds. Journal of the American Statistical Association , 101(473):138-156, 2006.
  4. [BNK + 20] Gagan Bansal, Besmira Nushi, Ece Kamar, Eric Horvitz, and Daniel S Weld. Optimizing ai for teamwork. arXiv preprint arXiv:2004.13102 , 2020. · arXiv
  5. [BW08] Peter L Bartlett and Marten H Wegkamp. Classification with a reject option using a hinge loss. Journal of Machine Learning Research , 9(Aug):1823-1840, 2008.
  6. [CCD + 19] Ran Canetti, Aloni Cohen, Nishanth Dikkala, Govind Ramnarayan, Sarah Scheffler, and Adam Smith. From soft classifiers to hard decisions: How fair can we be? In Proceedings of the Conference on Fairness, Accountability, and Transparency , pages 309-318. ACM, 2019.
  7. [CDM16a] Corinna Cortes, Giulia DeSalvo, and Mehryar Mohri. Boosting with abstention. In Advances in Neural Information Processing Systems , pages 1660-1668, 2016.
  8. [CDM16b] Corinna Cortes, Giulia DeSalvo, and Mehryar Mohri. Learning with rejection. In International Conference on Algorithmic Learning Theory , pages 67-82. Springer, 2016.
  9. [CGHS19] Minmin Chen, Ramki Gummadi, Chris Harris, and Dale Schuurmans. Surrogate objectives for batch policy optimization in one-step decision making. In Advances in Neural Information Processing Systems , pages 8825-8835, 2019.
  10. [Cho70] C Chow. On optimum recognition error and reject tradeoff. IEEE Transactions on information theory , 16(1):41-46, 1970.
  11. [DBW19] Thomas Davidson, Debasmita Bhattacharya, and Ingmar Weber. Racial bias in hate speech and abusive language detection datasets. arXiv preprint arXiv:1905.12516 , 2019. · arXiv
  12. [DDS + 09] Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition , pages 248-255. Ieee, 2009.
  13. [DI18] Cynthia Dwork and Christina Ilvento. Fairness under composition. arXiv preprint arXiv:1806.06122 , 2018. · arXiv
  14. [DKGGR19] Abir De, Paramita Koley, Niloy Ganguly, and Manuel Gomez-Rodriguez. Regression under human assistance. arXiv preprint arXiv:1909.02963 , 2019. · arXiv
  15. [DMS15] Giulia DeSalvo, Mehryar Mohri, and Umar Syed. Learning with deep cascades. In International Conference on Algorithmic Learning Theory , pages 254-269. Springer, 2015.
  16. [DWMW17] Thomas Davidson, Dana Warmsley, Michael Macy, and Ingmar Weber. Automated hate speech detection and the problem of offensive language. In Eleventh international aaai conference on web and social media , 2017.
  17. [EYW10] Ran El-Yaniv and Yair Wiener. On the foundations of noise-free selective classification. Journal of Machine Learning Research , 11(May):1605-1641, 2010.
  18. [FHH + 17] Joseph Futoma, Sanjay Hariharan, Katherine Heller, Mark Sendak, Nathan Brajer, Meredith Clement, Armando Bedoya, and Cara O'Brien. An improved multi-output gaussian process rnn with real-time validation for early sepsis detection. In Machine Learning for Healthcare Conference , pages 243-254, 2017.
  19. [FHT01] Jerome Friedman, Trevor Hastie, and Robert Tibshirani. The elements of statistical learning , volume 1. Springer series in statistics New York, 2001.
  20. [GC19a] Ben Green and Yiling Chen. Disparate interactions: An algorithm-in-the-loop analysis of fairness in risk assessments. In Proceedings of the Conference on Fairness, Accountability, and Transparency , pages 90-99. ACM, 2019.
  21. [GEY17] Yonatan Geifman and Ran El-Yaniv. Selective classification for deep neural networks. In Advances in neural information processing systems , pages 4878-4887, 2017.
  22. [GEY19] Yonatan Geifman and Ran El-Yaniv. Selectivenet: A deep neural network with an integrated reject option. arXiv preprint arXiv:1901.09192 , 2019. · arXiv
  23. [HAA + 17] Kanza Hamid, Amina Asif, Wajid Abbasi, Durre Sabih, et al. Machine learning with abstention for automated liver disease diagnosis. In 2017 International Conference on Frontiers of Information Technology (FIT) , pages 356-361. IEEE, 2017.
  24. [IRK + 19] Jeremy Irvin, Pranav Rajpurkar, Michael Ko, Yifan Yu, Silviana Ciurea-Ilcus, Chris Chute, Henrik Marklund, Behzad Haghgoo, Robyn Ball, Katie Shpanskaya, et al. Chexpert: A large chest radiograph dataset with uncertainty labels and expert comparison. In Proceedings of the AAAI Conference on Artificial Intelligence , volume 33, pages 590-597, 2019.
  25. [JBGB19] Shagun Jhaver, Iris Birman, Eric Gilbert, and Amy Bruckman. Human-machine collaboration for content regulation: The case of reddit automoderator. ACM Transactions on Computer-Human Interaction (TOCHI) , 26(5):1-35, 2019.
  26. [JGP16] Eric Jang, Shixiang Gu, and Ben Poole. Categorical reparameterization with gumbelsoftmax. arXiv preprint arXiv:1611.01144 , 2016. · arXiv
  27. [JJ94] Michael I Jordan and Robert A Jacobs. Hierarchical mixtures of experts and the em algorithm. Neural computation , 6(2):181-214, 1994.
  28. [JKGG18] Heinrich Jiang, Been Kim, Melody Guan, and Maya Gupta. To trust or not to trust a classifier. In Advances in neural information processing systems , pages 5541-5552, 2018.
  29. [KBZ + 19] Alexander Kolesnikov, Lucas Beyer, Xiaohua Zhai, Joan Puigcerver, Jessica Yung, Sylvain Gelly, and Neil Houlsby. Large scale learning of general visual representations for transfer. arXiv preprint arXiv:1912.11370 , 2019. · arXiv
  30. [Kim14] Yoon Kim. Convolutional neural networks for sentence classification. arXiv preprint arXiv:1408.5882 , 2014. · arXiv
  31. [LH16] Ilya Loshchilov and Frank Hutter. Sgdr: Stochastic gradient descent with warm restarts. arXiv preprint arXiv:1608.03983 , 2016. · arXiv
  32. [LS13] Phil Long and Rocco Servedio. Consistency versus realizable h-consistency for multiclass classification. In International Conference on Machine Learning , pages 801-809, 2013.
  33. [MDSGR20] Vahid Balazadeh Meresht, Abir De, Adish Singla, and Manuel Gomez-Rodriguez. Learning to switch between machines and humans. arXiv preprint arXiv:2002.04258 , 2020. · arXiv
  34. [MMT16] Chris J Maddison, Andriy Mnih, and Yee Whye Teh. The concrete distribution: A continuous relaxation of discrete random variables. arXiv preprint arXiv:1611.00712 , 2016. · arXiv
  35. [MPZ18] David Madras, Toni Pitassi, and Richard Zemel. Predict responsibly: Improving fairness and accuracy by learning to defer. In Advances in Neural Information Processing Systems , pages 6150-6160, 2018.
  36. [MRT18] Mehryar Mohri, Afshin Rostamizadeh, and Ameet Talwalkar. Foundations of machine learning . MIT press, 2018.
  37. [NCHS19] Chenri Ni, Nontawat Charoenphakdee, Junya Honda, and Masashi Sugiyama. On possibility and impossibility of multiclass classification with rejection. arXiv preprint arXiv:1901.10655 , 2019. · arXiv
  38. [OOR + 18] Avital Oliver, Augustus Odena, Colin A Raffel, Ekin Dogus Cubuk, and Ian Goodfellow. Realistic evaluation of deep semi-supervised learning algorithms. In Advances in Neural Information Processing Systems , pages 3235-3246, 2018.
  39. [PBGR19] Joshua C Peterson, Ruairidh M Battleday, Thomas L Griffiths, and Olga Russakovsky. Human uncertainty makes classification more robust. In Proceedings of the IEEE International Conference on Computer Vision , pages 9617-9626, 2019.
  40. [PSM14] Jeffrey Pennington, Richard Socher, and Christopher D Manning. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP) , pages 1532-1543, 2014.
  41. [RBC + 19] Maithra Raghu, Katy Blumer, Greg Corrado, Jon Kleinberg, Ziad Obermeyer, and Sendhil Mullainathan. The algorithmic automation problem: Prediction, triage, and human effort. arXiv preprint arXiv:1903.12220 , 2019. · arXiv
  42. [RBS + 19] Maithra Raghu, Katy Blumer, Rory Sayres, Ziad Obermeyer, Bobby Kleinberg, Sendhil Mullainathan, and Jon Kleinberg. Direct uncertainty prediction for medical second opinions. In International Conference on Machine Learning , pages 5281-5290, 2019.
  43. [RTA + 18] Harish G Ramaswamy, Ambuj Tewari, Shivani Agarwal, et al. Consistent algorithms for multiclass classification with an abstain option. Electronic Journal of Statistics , 12(1):530-554, 2018.
  44. [SMM + 17] Noam Shazeer, Azalia Mirhoseini, Krzysztof Maziarz, Andy Davis, Quoc Le, Geoffrey Hinton, and Jeff Dean. Outrageously large neural networks: The sparsely-gated mixture-of-experts layer. arXiv preprint arXiv:1701.06538 , 2017. · arXiv
  45. [WHK20] Bryan Wilder, Eric Horvitz, and Ece Kamar. Learning to complement humans. arXiv preprint arXiv:2005.00582 , 2020. · arXiv
  46. [ZC15] Chicheng Zhang and Kamalika Chaudhuri. Active learning from weak and strong labelers. In Advances in Neural Information Processing Systems , pages 703-711, 2015.
  47. [ZWL + 19] Liu Ziyin, Zhikang Wang, Paul Pu Liang, Ruslan Salakhutdinov, Louis-Philippe Morency, and Masahito Ueda. Deep gamblers: Learning to abstain with portfolio theory. arXiv preprint arXiv:1907.00208 , 2019. · arXiv

Cited by

Sorted by citations

Authors
Mozannar, H., & Sontag, D.
Affiliations
Massachusetts Institute of Technology (US)
ORCID
Venue
arXiv
Publisher
arXiv
Published
2020
Volume
1
Pages
7076-7087
ISSN
2331-8422 (ISSN-L)
DOI
10.48550/arxiv.2006.01862
arXiv
2006.01862
OpenAlex
W3033575144
Cited by
61
References
6
Citations by year
2025: 8 · 2024: 9 · 2023: 13 · 2022: 10 · 2021: 16 · 2020: 5
Type
posted content
Language
en
Open access
Licence
arxiv.org/licenses/nonexclusive-distrib/1.0/
Locations
  • arXiv (Cornell University) (submittedVersion)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-sa)
  • arXiv (Cornell University)
  • International Conference on Machine Learning
Funders
National Science Foundation (1350965, CAREER)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Algorithms
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
  • Physical Sciences › Computer Science › Artificial Intelligence › Domain Adaptation and Few-Shot Learning
Keywords
learning to defer, consistent estimators, cost-sensitive learning, surrogate loss, cross-entropy loss, decision making under uncertainty

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