ConfBatch: Conformal Prediction for Adaptive Batch Sizing

Anna Chatzipapadopoulou, Stavros Toumpis, John Pavlopoulos
Proceedings of the Fifteenth Symposium on Conformal and Probabilistic Prediction with Applications, PMLR 329:1090-1092, 2026.

Abstract

We study whether conformal uncertainty can guide adaptive batch size selection during neural network training. We propose ConfBatch that periodically computes conformal prediction sets on held-out data, using their size as an uncertainty score to adjust batch size. On CIFAR-10, the opposite-direction ConfBatch variants achieve the highest accuracies, while standard ConfBatch has the lowest wall time among adaptive batch-size methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v329-chatzipapadopoulou26a, title = {ConfBatch: Conformal Prediction for Adaptive Batch Sizing}, author = {Chatzipapadopoulou, Anna and Toumpis, Stavros and Pavlopoulos, John}, booktitle = {Proceedings of the Fifteenth Symposium on Conformal and Probabilistic Prediction with Applications}, pages = {1090--1092}, year = {2026}, editor = {Ahlberg, Ernst and Johansson, Ulf and Boström, Henrik and Carlevaro, Alberto and Hallberg Szabadváry, Johan and Carlsson, Lars}, volume = {329}, series = {Proceedings of Machine Learning Research}, month = {02--04 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v329/main/assets/chatzipapadopoulou26a/chatzipapadopoulou26a.pdf}, url = {https://proceedings.mlr.press/v329/chatzipapadopoulou26a.html}, abstract = {We study whether conformal uncertainty can guide adaptive batch size selection during neural network training. We propose ConfBatch that periodically computes conformal prediction sets on held-out data, using their size as an uncertainty score to adjust batch size. On CIFAR-10, the opposite-direction ConfBatch variants achieve the highest accuracies, while standard ConfBatch has the lowest wall time among adaptive batch-size methods.} }
Endnote
%0 Conference Paper %T ConfBatch: Conformal Prediction for Adaptive Batch Sizing %A Anna Chatzipapadopoulou %A Stavros Toumpis %A John Pavlopoulos %B Proceedings of the Fifteenth Symposium on Conformal and Probabilistic Prediction with Applications %C Proceedings of Machine Learning Research %D 2026 %E Ernst Ahlberg %E Ulf Johansson %E Henrik Boström %E Alberto Carlevaro %E Johan Hallberg Szabadváry %E Lars Carlsson %F pmlr-v329-chatzipapadopoulou26a %I PMLR %P 1090--1092 %U https://proceedings.mlr.press/v329/chatzipapadopoulou26a.html %V 329 %X We study whether conformal uncertainty can guide adaptive batch size selection during neural network training. We propose ConfBatch that periodically computes conformal prediction sets on held-out data, using their size as an uncertainty score to adjust batch size. On CIFAR-10, the opposite-direction ConfBatch variants achieve the highest accuracies, while standard ConfBatch has the lowest wall time among adaptive batch-size methods.
APA
Chatzipapadopoulou, A., Toumpis, S. & Pavlopoulos, J.. (2026). ConfBatch: Conformal Prediction for Adaptive Batch Sizing. Proceedings of the Fifteenth Symposium on Conformal and Probabilistic Prediction with Applications, in Proceedings of Machine Learning Research 329:1090-1092 Available from https://proceedings.mlr.press/v329/chatzipapadopoulou26a.html.

Related Material