ERIS: Enhancing Privacy and Communication Efficiency in Serverless Federated Learning

2026-02-09

AI Summary

This work addresses the challenge of simultaneously achieving communication efficiency, model accuracy, and privacy preservation in billion-parameter federated learning. To this end, we propose ERIS, a serverless framework that distributes both communication and computation loads across multiple client-side aggregators through model sharding and distributed shifted gradient compression. ERIS achieves convergence rates comparable to FedAvg without relying on cryptographic techniques or noise injection, while effectively mitigating mutual information leakage by increasing the number of aggregators. Experimental results demonstrate that ERIS attains FedAvg-level accuracy on both vision and large language model tasks, substantially reduces communication overhead, and enhances robustness against membership inference and reconstruction attacks.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Computer Vision: Adversarial Attacks & Robustness
• Search and Optimization: Learning to Search

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Security and Privacy: Privacy-enhancing technologies
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Scaling federated learning (FL) to billion-parameter models introduces critical trade-offs between communication efficiency, model accuracy, and privacy guarantees. Existing solutions often tackle these challenges in isolation, sacrificing accuracy or relying on costly cryptographic tools. We propose ERIS, a serverless FL framework that balances privacy and accuracy while eliminating the server bottleneck and distributing the communication load. ERIS combines a model partitioning strategy, distributing aggregation across multiple client-side aggregators, with a distributed shifted gradient compression mechanism. We theoretically prove that ERIS (i) converges at the same rate as FedAvg under standard assumptions, and (ii) bounds mutual information leakage inversely with the number of aggregators, enabling strong privacy guarantees with no accuracy degradation. Experiments across image and text tasks, including large language models, confirm that ERIS achieves FedAvg-level accuracy while substantially reducing communication cost and improving robustness to membership inference and reconstruction attacks, without relying on heavy cryptography or noise injection.

Problem

• federated learning
• privacy
• communication efficiency
• serverless
• large-scale models

Innovation

• serverless federated learning
• model partitioning
• distributed gradient compression
• privacy-accuracy trade-off
• mutual information leakage