Leveraging Soft Prompts for Privacy Attacks in Federated Prompt Tuning

TL;DR

This paper reveals a new privacy vulnerability in federated prompt-tuning, demonstrating that a malicious server can perform membership inference attacks using crafted prompts, challenging existing defenses and highlighting the need for specialized privacy protections.

Contribution

The paper introduces PromptMIA, a novel membership inference attack tailored for federated prompt-tuning, along with theoretical analysis and evaluation on benchmark datasets.

Findings

PromptMIA achieves high attack advantage across datasets.

Standard defenses are ineffective against PromptMIA.

Theoretical lower bounds explain attack success.

Abstract

Membership inference attack (MIA) poses a significant privacy threat in federated learning (FL) as it allows adversaries to determine whether a client's private dataset contains a specific data sample. While defenses against membership inference attacks in standard FL have been well studied, the recent shift toward federated fine-tuning has introduced new, largely unexplored attack surfaces. To highlight this vulnerability in the emerging FL paradigm, we demonstrate that federated prompt-tuning, which adapts pre-trained models with small input prefixes to improve efficiency, also exposes a new vector for privacy attacks. We propose PromptMIA, a membership inference attack tailored to federated prompt-tuning, in which a malicious server can insert adversarially crafted prompts and monitors their updates during collaborative training to accurately determine whether a target data point is in a client's private dataset. We formalize this threat as a security game and empirically show that PromptMIA consistently attains high advantage in this game across diverse benchmark datasets. Our theoretical analysis further establishes a lower bound on the attack's advantage which explains and supports the consistently high advantage observed in our empirical results. We also investigate the effectiveness of standard membership inference defenses originally developed for gradient or output based attacks and analyze their interaction with the distinct threat landscape posed by PromptMIA. The results highlight non-trivial challenges for current defenses and offer insights into their limitations, underscoring the need for defense strategies that are specifically tailored to prompt-tuning in federated settings.