Conformal Training with Reduced Variance

ICLR 2025 Conference Submission13335 Authors

Keywords: Conformal Training, Conformal Prediction, Optimization, Quantile, Deep Learning, Uncertainty Quantification

TL;DR: We improve performance of conformal training by tackling optimization challenges.

Abstract: Conformal prediction (CP) is a distribution-free framework for achieving probabilistic guarantees on black-box models. {CP} is generally applied to a model post-training. Conformal training is an approach that aims to optimize the CP efficiency during training. In this direction, ConfTr (Stutz et al, 2022) is a technique that seeks to minimize the expected prediction set size of a model by simulating {CP} in-between training updates. Despite its potential, we identify a strong source of sample inefficiency in ConfTr that leads to overly noisy estimated gradients, introducing training instability and limiting practical use. To address this challenge, we propose variance-reduced conformal training (VR-ConfTr), a method that incorporates a variance reduction technique in the gradient estimation of the ConfTr objective function. Through extensive experiments on various benchmark datasets, we demonstrate that VR-ConfTr consistently achieves faster convergence and smaller prediction sets compared to baselines.

Primary Area: unsupervised, self-supervised, semi-supervised, and supervised representation learning

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: 13335