LGCRDCMar 18, 2024

Improving LoRA in Privacy-preserving Federated Learning

arXiv:2403.12313v1187 citationsh-index: 16ICLR
Originality Incremental advance
AI Analysis

This addresses a problem for federated learning practitioners by offering an incremental improvement to LoRA for enhanced privacy and efficiency in fine-tuning large language models.

The paper tackled the instability of Low-rank adaptation (LoRA) in privacy-preserving federated learning due to data heterogeneity, differential privacy noise, and hyper-parameter sensitivity, proposing Federated Freeze A LoRA (FFA-LoRA) which halved communication costs and improved performance consistency and computational efficiency.

Low-rank adaptation (LoRA) is one of the most popular task-specific parameter-efficient fine-tuning (PEFT) methods on pre-trained language models for its good performance and computational efficiency. LoRA injects a product of two trainable rank decomposition matrices over the top of each frozen pre-trained model module. However, when applied in the setting of privacy-preserving federated learning (FL), LoRA may become unstable due to the following facts: 1) the effects of data heterogeneity and multi-step local updates are non-negligible, 2) additive noise enforced on updating gradients to guarantee differential privacy (DP) can be amplified and 3) the final performance is susceptible to hyper-parameters. A key factor leading to these phenomena is the discordance between jointly optimizing the two low-rank matrices by local clients and separately aggregating them by the central server. Thus, this paper proposes an efficient and effective version of LoRA, Federated Freeze A LoRA (FFA-LoRA), to alleviate these challenges and further halve the communication cost of federated fine-tuning LLMs. The core idea of FFA-LoRA is to fix the randomly initialized non-zero matrices and only fine-tune the zero-initialized matrices. Compared to LoRA, FFA-LoRA is motivated by practical and theoretical benefits in privacy-preserved FL. Our experiments demonstrate that FFA-LoRA provides more consistent performance with better computational efficiency over vanilla LoRA in various FL tasks.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes