Online Conformal Prediction via Universal Portfolio Algorithms

Tuo Liu, Edgar Dobriban, Francesco Orabona

Keywords: Conformal prediction, online learning, universal portfolio

Abstract: Online conformal prediction (OCP) seeks prediction intervals that achieve long-run $1-\alpha$ coverage for arbitrary (possibly adversarial) data streams, while remaining as informative as possible. Here, we propose UP-OCP, a new approach to OCP, leveraging universal portfolio algorithms. We show strong finite-time bounds on the miscoverage of UP-OCP. In experiments, UP-OCP delivers better size/coverage trade-offs than existing methods.

Submission Number: 15