Conformal Prediction for Federated Uncertainty Quantification Under Label Shift
This work addresses uncertainty quantification for federated learning systems, particularly under label shift, with incremental improvements to existing conformal prediction methods.
The paper tackles uncertainty quantification in federated learning under label shift by developing a federated conformal prediction method using quantile regression and importance weighting, which provides theoretical guarantees for valid coverage and differential privacy and outperforms current competitors in experiments.
Federated Learning (FL) is a machine learning framework where many clients collaboratively train models while keeping the training data decentralized. Despite recent advances in FL, the uncertainty quantification topic (UQ) remains partially addressed. Among UQ methods, conformal prediction (CP) approaches provides distribution-free guarantees under minimal assumptions. We develop a new federated conformal prediction method based on quantile regression and take into account privacy constraints. This method takes advantage of importance weighting to effectively address the label shift between agents and provides theoretical guarantees for both valid coverage of the prediction sets and differential privacy. Extensive experimental studies demonstrate that this method outperforms current competitors.