# Designing Closed Human-in-the-loop Deferral Pipelines

arXiv: https://arxiv.org/abs/2202.04718 · HTML: https://mlsanity.com/paper/2202.04718/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 49 references that are in the map.
- First posted on arXiv 9 Feb 2022.

## Builds on (most cited first, 16 of 16)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [A Survey on Bias and Fairness in Machine Learning](https://mlsanity.com/paper/1908.09635/) (2019)
- [Automated Hate Speech Detection and the Problem of Offensive Language](https://mlsanity.com/paper/1703.04009/) (2017)
- [A Tutorial on Thompson Sampling](https://mlsanity.com/paper/1707.02038/) (2017)
- [Consistent Estimators for Learning to Defer to an Expert](https://mlsanity.com/paper/2006.01862/) (2020)
- [Predict Responsibly: Improving Fairness and Accuracy by Learning to Defer](https://mlsanity.com/paper/1711.06664/) (2017)
- [Racial Bias in Hate Speech and Abusive Language Detection Datasets](https://mlsanity.com/paper/1905.12516/) (2019)
- [Deep Bayesian Bandits Showdown: An Empirical Comparison of Bayesian Deep Networks for Thompson Sampling](https://mlsanity.com/paper/1802.09127/) (2018)
- [Runaway Feedback Loops in Predictive Policing](https://mlsanity.com/paper/1706.09847/) (2017)
- [Towards Unbiased and Accurate Deferral to Multiple Experts](https://mlsanity.com/paper/2102.13004/) (2021)
- [Deep Gamblers: Learning to Abstain with Portfolio Theory](https://mlsanity.com/paper/1907.00208/) (2019)
- [Direct Uncertainty Prediction for Medical Second Opinions](https://mlsanity.com/paper/1807.01771/) (2018)
- [Mitigating Bias in Adaptive Data Gathering via Differential Privacy](https://mlsanity.com/paper/1806.02329/) (2018)
- [Enhancing the Accuracy and Fairness of Human Decision Making](https://mlsanity.com/paper/1805.10318/) (2018)
- [Crowdsourcing with Fairness, Diversity and Budget Constraints](https://mlsanity.com/paper/1810.13314/) (2018)
- [Auditing for Diversity using Representative Examples](https://mlsanity.com/paper/2107.07393/) (2021)

## Built on by (most cited first, 11 of 11)

- [Human-Algorithm Collaboration: Achieving Complementarity and Avoiding Unfairness](https://mlsanity.com/paper/2202.08821/) (2022)
- [Human-AI Collaboration in Decision-Making: Beyond Learning to Defer](https://mlsanity.com/paper/2206.13202/) (2022)
- [Human Expertise in Algorithmic Prediction](https://mlsanity.com/paper/2402.00793/) (2024)
- [Explaining and Improving Information Complementarities in Multi-Agent Decision-making](https://mlsanity.com/paper/2502.06152/) (2025)
- [Learning To Guide Human Decision Makers With Vision-Language Models](https://mlsanity.com/paper/2403.16501/) (2024)
- [Inclusive Portraits: Race-Aware Human-in-the-Loop Technology](https://mlsanity.com/paper/2311.03567/) (2023)
- [Integrating Expert Judgment and Algorithmic Decision Making: An Indistinguishability Framework](https://mlsanity.com/paper/2410.08783/) (2024)
- [ComplLLM: Fine-tuning LLMs to Discover Complementary Signals for Decision-making](https://mlsanity.com/paper/2602.19458/) (2026)
- [Improving the Efficiency of Human-in-the-Loop Systems: Adding Artificial to Human Experts](https://mlsanity.com/paper/2307.03003/) (2023)
- [MedGellan: LLM-Generated Medical Guidance to Support Physicians](https://mlsanity.com/paper/2507.04431/) (2025)
- [Learning to Guide Human Experts via Personalized Large Language Models](https://mlsanity.com/paper/2308.06039/) (2023)

## Abstract (arXiv)

In hybrid human-machine deferral frameworks, a classifier can defer uncertain cases to human decision-makers (who are often themselves fallible). Prior work on simultaneous training of such classifier and deferral models has typically assumed access to an oracle during training to obtain true class labels for training samples, but in practice there often is no such oracle. In contrast, we consider a "closed" decision-making pipeline in which the same fallible human decision-makers used in deferral also provide training labels. How can imperfect and biased human expert labels be used to train a fair and accurate deferral framework? Our key insight is that by exploiting weak prior information, we can match experts to input examples to ensure fairness and accuracy of the resulting deferral framework, even when imperfect and biased experts are used in place of ground truth labels. The efficacy of our approach is shown both by theoretical analysis and by evaluation on two tasks.