Improving LoRA in Privacy-preserving Federated Learning

arXiv:2403.12313v1 Announce Type: cross
Abstract: 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 …

adaptation arxiv computational cs.cr cs.dc cs.lg efficiency federated federated learning fine-tuning good language language models lora low parameter performance popular privacy product task