Label Privacy in Split Learning for Large Models with Parameter-Efficient Training

TL;DR

This paper introduces P³EFT, a privacy-preserving split learning algorithm for large models with parameter-efficient fine-tuning, ensuring label privacy with minimal performance loss in API-based training.

Contribution

The study presents P³EFT, a novel split learning method that maintains label privacy during fine-tuning large models with PEFT, outperforming existing privacy methods in accuracy.

Findings

P³EFT achieves competitive privacy and accuracy in NLP tasks.

Analysis shows LoRA's privacy vulnerabilities and how P³EFT mitigates them.

Experimental results on DeBERTa, Flan-T5, and LLaMA-2 demonstrate effectiveness.

Abstract

As deep learning models become larger and more expensive, many practitioners turn to fine-tuning APIs. These web services allow fine-tuning a model between two parties: the client that provides the data, and the server that hosts the model. While convenient, these APIs raise a new concern: the data of the client is at risk of privacy breach during the training procedure. This challenge presents an important practical case of vertical federated learning, where the two parties perform parameter-efficient fine-tuning (PEFT) of a large model. In this study, we systematically search for a way to fine-tune models over an API while keeping the labels private. We analyze the privacy of LoRA, a popular approach for parameter-efficient fine-tuning when training over an API. Using this analysis, we propose P$^3$EFT, a multi-party split learning algorithm that takes advantage of existing PEFT properties to maintain privacy at a lower performance overhead. To validate our algorithm, we fine-tune DeBERTa-v2-XXLarge, Flan-T5 Large and LLaMA-2 7B using LoRA adapters on a range of NLP tasks. We find that P$^3$EFT is competitive with existing privacy-preserving methods in multi-party and two-party setups while having higher accuracy.