WinFLoRA: Incentivizing Client-Adaptive Aggregation in Federated LoRA under Privacy Heterogeneity

TL;DR

WinFLoRA introduces a privacy-aware federated LoRA method that incentivizes clients to contribute low-noise updates by adjusting aggregation weights, significantly enhancing global model accuracy and client utility in privacy-heterogeneous settings.

Contribution

It proposes a novel aggregation weighting scheme in federated LoRA that estimates client noise levels to incentivize low-noise contributions, aligning individual privacy preferences with global performance.

Findings

Achieves up to 52.58% higher global accuracy.

Attains up to 2.56x improvement in client utility.

Effectively handles privacy heterogeneity in federated LLM adaptation.

Abstract

Large Language Models (LLMs) increasingly underpin intelligent web applications, from chatbots to search and recommendation, where efficient specialization is essential. Low-Rank Adaptation (LoRA) enables such adaptation with minimal overhead, while federated LoRA allows web service providers to fine-tune shared models without data sharing. However, in privacy-sensitive deployments, clients inject varying levels of differential privacy (DP) noise, creating privacy heterogeneity that misaligns individual incentives and global performance. In this paper, we propose WinFLoRA, a privacy-heterogeneous federated LoRA that utilizes aggregation weights as incentives with noise awareness. Specifically, the noises from clients are estimated based on the uploaded LoRA adapters. A larger weight indicates greater influence on the global model and better downstream task performance, rewarding lower-noise contributions. By up-weighting low-noise updates, WinFLoRA improves global accuracy while accommodating clients' heterogeneous privacy requirements. Consequently, WinFLoRA aligns heterogeneous client utility in terms of privacy and downstream performance with global model objectives without third-party involvement. Extensive evaluations demonstrate that across multiple LLMs and datasets, WinFLoRA achieves up to 52.58% higher global accuracy and up to 2.56x client utility than state-of-the-art benchmarks. Source code is publicly available at https://github.com/koums24/WinFLoRA.git.