FedP$^2$EFT: Federated Learning to Personalize Parameter Efficient Fine-Tuning for Multilingual LLMs

2025-02-05

AI Summary

To address the challenges of personalizing large language models (LLMs) for low-resource languages in cross-device federated learning (FL)—namely, difficulty in client-specific adaptation and poor generalizability of manually configured parameter-efficient fine-tuning (PEFT) structures—this paper proposes the first collaborative personalized PEFT structure learning framework based on Bayesian sparse rank selection. Our method automatically jointly optimizes both the insertion locations and rank parameters of LoRA modules across clients, eliminating manual design of personalization strategies and mitigating overfitting under data-scarce conditions. The framework unifies federated learning, LoRA-based parameter-efficient fine-tuning, Bayesian sparse modeling, and cooperative hyperparameter optimization across heterogeneous clients. Extensive experiments on synthetic and real-world multilingual FL benchmarks demonstrate significant improvements over existing personalized fine-tuning approaches, while maintaining full compatibility with mainstream federated optimization algorithms.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Search and Optimization: Learning to Search

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
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

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 a range of existing FL methods.

Problem

• Optimize PEFT for multilingual LLMs
• Automate personalized structure selection
• Enhance FL performance in low-data

Innovation

• Federated learning for multilingual LLMs
• Bayesian sparse rank selection
• Personalized PEFT structure optimization