FedP$^2$EFT: Federated Learning to Personalize PEFT for Multilingual LLMs

TL;DR

FedP$^2$EFT introduces a federated learning approach that automatically personalizes PEFT structures for multilingual LLMs, enhancing client-specific performance in low-resource and diverse data settings.

Contribution

It proposes a novel federated learning-to-personalize method using Bayesian sparse rank selection to optimize PEFT structures for each client.

Findings

Outperforms existing personalized fine-tuning methods.

Effective in both simulated and real-world multilingual FL benchmarks.

Enhances client-specific performance in low-resource language scenarios.

Abstract

Federated learning (FL) has enabled the training of multilingual large language models (LLMs) on diverse and decentralized multilingual data, especially on low-resource languages. To improve client-specific performance, personalization via the use of parameter-efficient fine-tuning (PEFT) modules such as LoRA is common. This involves a personalization strategy (PS), such as the design of the PEFT adapter structures (e.g., in which layers to add LoRAs and what ranks) and choice of hyperparameters (e.g., learning rates) for fine-tuning. Instead of manual PS configuration, we propose FedP$^2$EFT, a federated learning-to-personalize method for multilingual LLMs in cross-device FL settings. Unlike most existing PEFT structure selection methods, which are prone to overfitting low-data regimes, FedP$^2$EFT collaboratively learns the optimal personalized PEFT structure for each client via Bayesian sparse rank selection. Evaluations on both simulated and real-world multilingual FL benchmarks demonstrate that FedP$^2$EFT largely outperforms existing personalized fine-tuning methods, while complementing other existing FL methods.