SDBA: A Stealthy and Long-Lasting Durable Backdoor Attack in Federated Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Choe, Minyeong, Park, Cheolhee, Seo, Changho, Kim, Hyunil
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918107458568192
author Choe, Minyeong
Park, Cheolhee
Seo, Changho
Kim, Hyunil
author_facet Choe, Minyeong
Park, Cheolhee
Seo, Changho
Kim, Hyunil
contents Federated learning is a promising approach for training machine learning models while preserving data privacy. However, its distributed nature makes it vulnerable to backdoor attacks, particularly in NLP tasks, where related research remains limited. This paper introduces SDBA, a novel backdoor attack mechanism designed for NLP tasks in federated learning environments. Through a systematic analysis across LSTM and GPT-2 models, we identify the most vulnerable layers for backdoor injection and achieve both stealth and long-lasting durability by applying layer-wise gradient masking and top-k% gradient masking. Also, to evaluate the task generalizability of SDBA, we additionally conduct experiments on the T5 model. Experiments on next-token prediction, sentiment analysis, and question answering tasks show that SDBA outperforms existing backdoors in terms of durability and effectively bypasses representative defense mechanisms, demonstrating notable performance in transformer-based models such as GPT-2. These results highlight the urgent need for robust defense strategies in NLP-based federated learning systems.
format Preprint
id arxiv_https___arxiv_org_abs_2409_14805
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SDBA: A Stealthy and Long-Lasting Durable Backdoor Attack in Federated Learning
Choe, Minyeong
Park, Cheolhee
Seo, Changho
Kim, Hyunil
Machine Learning
Cryptography and Security
Federated learning is a promising approach for training machine learning models while preserving data privacy. However, its distributed nature makes it vulnerable to backdoor attacks, particularly in NLP tasks, where related research remains limited. This paper introduces SDBA, a novel backdoor attack mechanism designed for NLP tasks in federated learning environments. Through a systematic analysis across LSTM and GPT-2 models, we identify the most vulnerable layers for backdoor injection and achieve both stealth and long-lasting durability by applying layer-wise gradient masking and top-k% gradient masking. Also, to evaluate the task generalizability of SDBA, we additionally conduct experiments on the T5 model. Experiments on next-token prediction, sentiment analysis, and question answering tasks show that SDBA outperforms existing backdoors in terms of durability and effectively bypasses representative defense mechanisms, demonstrating notable performance in transformer-based models such as GPT-2. These results highlight the urgent need for robust defense strategies in NLP-based federated learning systems.
title SDBA: A Stealthy and Long-Lasting Durable Backdoor Attack in Federated Learning
topic Machine Learning
Cryptography and Security
url https://arxiv.org/abs/2409.14805