Federated Conformal Predictors for Distributed Uncertainty Quantification

TL;DR

This paper introduces Federated Conformal Prediction (FCP), a novel method for uncertainty quantification in federated learning that addresses data heterogeneity through a new partial exchangeability concept, with strong theoretical guarantees and empirical success.

Contribution

We develop FCP, extending conformal prediction to federated learning by relaxing exchangeability assumptions, enabling rigorous uncertainty quantification in heterogeneous distributed data environments.

Findings

FCP provides valid uncertainty intervals under partial exchangeability.

FCP achieves high empirical accuracy on vision and medical datasets.

Theoretical guarantees support FCP's reliability in federated settings.

Abstract

Conformal prediction is emerging as a popular paradigm for providing rigorous uncertainty quantification in machine learning since it can be easily applied as a post-processing step to already trained models. In this paper, we extend conformal prediction to the federated learning setting. The main challenge we face is data heterogeneity across the clients - this violates the fundamental tenet of exchangeability required for conformal prediction. We propose a weaker notion of partial exchangeability, better suited to the FL setting, and use it to develop the Federated Conformal Prediction (FCP) framework. We show FCP enjoys rigorous theoretical guarantees and excellent empirical performance on several computer vision and medical imaging datasets. Our results demonstrate a practical approach to incorporating meaningful uncertainty quantification in distributed and heterogeneous environments. We provide code used in our experiments https://github.com/clu5/federated-conformal.