Selective Prediction via Training Dynamics

Stephan Rabanser, Anvith Thudi, Kimia Hamidieh, Adam Dziedzic, Israfil Bahceci, Akram Bin Sediq, HAMZA SOKUN, Nicolas Papernot

Primary Area: societal considerations including fairness, safety, privacy

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.

Keywords: selective prediction, training dynamics, example difficulty, forging, reject option, uncertainty quantification, reliability

Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.

TL;DR: We propose a novel approach for selective prediction based on the prediction disagreement evolution of intermediate models with the final model's prediction.

Abstract: Selective prediction is the task of rejecting inputs a model would predict incorrectly on through a trade-off between input space coverage and model utility. Current methods for selective prediction typically impose constraints on either the model architecture or the loss function; this inhibits their usage in practice. In contrast to prior work, we show that state-of-the-art selective prediction performance can be attained solely from studying the (discretized) training dynamics of a model. We propose a general framework that, given a test input, monitors metrics capturing the instability of predictions from intermediate models obtained during training w.r.t. the final model's prediction. In particular, we reject data points exhibiting too much disagreement with the final prediction at late stages in training. The proposed scoring mechanism is domain-agnostic (i.e., it works for both discrete and real-valued prediction) and can be flexibly combined with existing selective prediction approaches as it does not require any train-time modifications. Our experimental evaluation on image classification, regression, and time series forecasting problems shows that our method beats past state-of-the-art accuracy/utility trade-offs on typical selective prediction benchmarks.

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: 4024