Online conformal prediction with decaying step sizes

Anastasios Nikolas Angelopoulos, Rina Foygel Barber, Stephen Bates
2/12/2026

No official code found. Best implementation: aangelopoulos/online-conformal-decaying (9 stars, MIT, last commit 2024-05-05). 1 more below.

Abstract

We introduce a method for online conformal prediction with decaying step sizes. Like previous methods, ours possesses a retrospective guarantee of coverage for arbitrary sequences. However, unlike previous methods, we can simultaneously estimate a population quantile when it exists. Our theory and experiments indicate substantially improved practical properties: in particular, when the distribution is stable, the coverage is close to the desired level for every time point, not just on average over the observed sequence.

View on arXivView PDF

Code Implementations(2)

aangelopoulos/online-conformal-decayingCommunity100%
MIT
90Python, Jupyter NotebookJan 25, 20242 years agoMIT
SamyBraik/OnlineConformalCommunity64%

Présentation basée sur l'article Online conformal prediction with decaying step sizes de Anastasios N. Angelopoulos, Rina Foygel Barber, Stephen Bates.

10Feb 19, 20251 years ago

Similar Papers

Distribution-informed Online Conformal Prediction

86% match1 repo62025

Online Conformal Prediction with Efficiency Guarantees

86% match2025

Exploring the Noise Robustness of Online Conformal Prediction

85% match2025

One-Shot Federated Conformal Prediction

83% match1 repo112026

Backward Conformal Prediction

83% match1 repo42025

Conformal Prediction with Missing Values

83% match2026

Cite this paper

@article{angelopoulos2026online,
  title  = {Online conformal prediction with decaying step sizes},
  author = {Anastasios Nikolas Angelopoulos and Rina Foygel Barber and Stephen Bates},
  year   = {2026},
  eprint = {2402.01139},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/2402.01139},
  journal = {ICML 2024 2024}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2402.01139.svg)](https://codeofpaper.com/paper/2402.01139)

Discussion