mlsanityData to 14 Oct 2026

A Unifying Post-Processing Framework for Multi-Objective Learn-to-Defer Problems

arXiv 2407.12710 ↗First posted 17 Jul 2024cs.LG cs.AIcited by 5

Not on the map. Why?

Abstract (arXiv)

Learn-to-Defer is a paradigm that enables learning algorithms to work not in isolation but as a team with human experts. In this paradigm, we permit the system to defer a subset of its tasks to the expert. Although there are currently systems that follow this paradigm and are designed to optimize the accuracy of the final human-AI team, the general methodology for developing such systems under a set of constraints (e.g., algorithmic fairness, expert intervention budget, defer of anomaly, etc.) remains largely unexplored. In this paper, using a d-dimensional generalization to the fundamental lemma of Neyman and Pearson (d-GNP), we obtain the Bayes optimal solution for learn-to-defer systems under various constraints. Furthermore, we design a generalizable algorithm to estimate that solution and apply this algorithm to the COMPAS and ACSIncome datasets. Our algorithm shows improvements in terms of constraint violation over a set of baselines.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 32 of its 70 references that are in the map. First posted on arXiv 17 Jul 2024.

Builds on

32 of 70 references.

References · 32

  1. 2016Equality of Opportunity in Supervised Learningcited by 2,007
  2. 2017Selective Classification for Deep Neural Networkscited by 852
  3. 2021Retiring Adult: New Datasets for Fair Machine Learningcited by 449
  4. 2017Algorithmic decision making and the cost of fairnesscited by 403
  5. 2020Consistent Estimators for Learning to Defer to an Expertcited by 271
  6. 2017Predict Responsibly: Improving Fairness and Accuracy by Learning to Defercited by 233
  7. 2019The Algorithmic Automation Problem: Prediction, Triage, and Human Effortcited by 101
  8. 2021Differentiable Learning Under Triagecited by 76
  9. 2023Who Should Predict? Exact Algorithms For Learning to Defer to Humanscited by 70
  10. 2019Leveraging Labeled and Unlabeled Data for Consistent Fair Binary Classificationcited by 60
  11. 2020Classification with Rejection Based on Cost-sensitive Classificationcited by 55
  12. 2022Fair and Optimal Classification via Post-Processingcited by 51
20 more
  1. 2022Sample Efficient Learning of Predictors that Complement Humanscited by 46
  2. 2022Human-Algorithm Collaboration: Achieving Complementarity and Avoiding Unfairnesscited by 45
  3. 2018Fairness Under Compositioncited by 43
  4. 2020Classification Under Human Assistancecited by 42
  5. 2019Regression Under Human Assistancecited by 40
  6. 2020Selective Classification via One-Sided Predictioncited by 38
  7. 2020Selective Classification Can Magnify Disparities Across Groupscited by 36
  8. 2023Predictor-Rejector Multi-Class Abstention: Theoretical Analysis and Algorithmscited by 36
  9. 2024Learning to Defer to a Population: A Meta-Learning Approachcited by 27
  10. 2023In Defense of Softmax Parametrization for Calibrated and Consistent Learning to Defercited by 26
  11. 2022Bayes-Optimal Classifiers under Group Fairnesscited by 24
  12. 2018Investigating Human + Machine Complementarity for Recidivism Predictionscited by 19
  13. 2023Unprocessing Seven Years of Algorithmic Fairnesscited by 18
  14. 2023Plugin estimators for selective classification with out-of-distribution detectioncited by 14
  15. 2023Post-hoc Bias Scoring Is Optimal For Fair Classificationcited by 14
  16. 2022Consistent Multiclass Algorithms for Complex Metrics and Constraintscited by 13
  17. 2023Regression with Cost-based Rejectioncited by 11
  18. 2024Bayes-Optimal Fair Classification with Linear Disparity Constraints via Pre-, In-, and Post-processingcited by 10
  19. 2023Fair Classifiers that Abstain without Harmcited by 8
  20. 2021Neyman-Pearson Multi-class Classification via Cost-sensitive Learningcited by 6

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 2; 2026: 2.

5 papers.

  1. 2024Realizable H-Consistent and Bayes-Consistent Loss Functions for Learning to Defercited by 26
  2. 2025Mastering Multiple-Expert Routing: Realizable H-Consistency and Strong Guarantees for Learning to Defercited by 16
  3. 2025Budgeted Multiple-Expert Deferralcited by 13
  4. 2026When More Experts Hurt: Underfitting in Multi-Expert Learning to Defercited by 3
  5. 2026Density-Ratio Losses for Post-Hoc Learning to Defercited by 1