Regretful Decisions under Label Noise
Keywords: Uncertainty Quantification, Fairness, Model Multiplicity, Clinical Decision Support, Classification, Label Noise
Abstract: Machine learning models are routinely used to support decisions that affect individuals -- be it to screen a patient for a serious illness or to gauge their response to treatment. In these tasks, we are limited to learning models from datasets where the labels are subject to noise. In this work, we study the impact of learning under label noise at the instance level. We introduce a notion of *regret* for this regime, which measures the number of unforeseen mistakes when learning from noisy labels. We show that standard approaches to learn models from noisy labels can return models that perform well at a population level while subjecting individuals to a *lottery of mistakes*. We develop machinery to estimate the likelihood of mistakes at an instance level from a noisy dataset, by training models over plausible realizations of datasets without label noise. We present a comprehensive empirical study of label noise in clinical prediction tasks. Our results reveal how our failure to anticipate mistakes can compromise model reliance and adoption, and demonstrate how we can address these challenges by anticipating and abstaining from regretful decisions.
Primary Area: alignment, fairness, safety, privacy, and societal considerations
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 548
Loading