English

Instance-Adaptive Online Multicalibration

Machine Learning 2026-05-22 v2

Abstract

We study online multicalibration beyond the worst-case. We give a single, efficient algorithm which dynamically interpolates between benign and worst-case sequences by adaptively refining a dyadic grid of prediction values. Its error is controlled by the number of leaves in the refinement tree. Our analysis recovers the known O~(T2/3)\widetilde O(T^{2/3}) worst-case-optimal rate for online multicalibration, while simultaneously automatically adapting to easier instances: in the marginal stochastic setting it obtains a rate of O~(T)\widetilde O(\sqrt T), and for piecewise-stationary means with JJ segments its rate is O~(JT)\widetilde O(\sqrt{JT}). More generally, the rate depends on a threshold-complexity measure of the predictable mean process relative to the group family. We show that this dependence is tight up to logarithmic factors.

Keywords

Cite

@article{arxiv.2605.09273,
  title  = {Instance-Adaptive Online Multicalibration},
  author = {Zhiming Huang and Jamie Morgenstern and Aaron Roth and Claire Jie Zhang},
  journal= {arXiv preprint arXiv:2605.09273},
  year   = {2026}
}

Comments

We tightened the analysis and added a comparison to the concurrent work of Liu et al. (arXiv:2605.11490)

R2 v1 2026-07-01T13:01:07.472Z