Efficient Federated Conformal Prediction with Group-Conditional Guarantee

2026-03-14

AI Summary

This work addresses the challenge of achieving group-conditional coverage guarantees in federated learning, where calibration data are distributed across multiple clients and exhibit inherent group structure. The authors propose Group-Conditional Federated Conformal Prediction (GC-FCP), the first method to provide rigorous group-conditional coverage guarantees in a federated setting. At its core, GC-FCP constructs mergeable group-stratified coreset summaries from local calibration scores, enabling each client to transmit compact, weighted representations that support efficient communication and server-side aggregation. Empirical evaluations on both synthetic and real-world datasets demonstrate that GC-FCP consistently maintains its theoretical coverage guarantees while achieving predictive performance comparable to centralized conformal calibration baselines.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Natural Language Processing: Safety and Robustness

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Large-scale security measurements

Abstract

Deploying trustworthy AI systems requires principled uncertainty quantification. Conformal prediction (CP) is a widely used framework for constructing prediction sets with distribution-free coverage guarantees. In many practical settings, including healthcare, finance, and mobile sensing, the calibration data required for CP are distributed across multiple clients, each with its own local data distribution. In this federated setting, data can often be partitioned into, potentially overlapping, groups, which may reflect client-specific strata or cross-cutting attributes such as demographic or semantic categories. We propose group-conditional federated conformal prediction (GC-FCP), a novel protocol that provides group-conditional coverage guarantees. GC-FCP constructs mergeable, group-stratified coresets from local calibration scores, enabling clients to communicate compact weighted summaries that support efficient aggregation and calibration at the server. Experiments on synthetic and real-world datasets validate the performance of GC-FCP compared to centralized calibration baselines.

Problem

• federated learning
• conformal prediction
• group-conditional guarantee
• uncertainty quantification
• distributed calibration

Innovation

• Federated Conformal Prediction
• Group-Conditional Coverage
• Mergeable Coresets
• Uncertainty Quantification
• Distributed Calibration