Our new X account is live! Follow @wizwand_team for updates
WorkDL logo mark

Defending Pre-trained Language Models as Few-shot Learners against Backdoor Attacks

About

Pre-trained language models (PLMs) have demonstrated remarkable performance as few-shot learners. However, their security risks under such settings are largely unexplored. In this work, we conduct a pilot study showing that PLMs as few-shot learners are highly vulnerable to backdoor attacks while existing defenses are inadequate due to the unique challenges of few-shot scenarios. To address such challenges, we advocate MDP, a novel lightweight, pluggable, and effective defense for PLMs as few-shot learners. Specifically, MDP leverages the gap between the masking-sensitivity of poisoned and clean samples: with reference to the limited few-shot data as distributional anchors, it compares the representations of given samples under varying masking and identifies poisoned samples as ones with significant variations. We show analytically that MDP creates an interesting dilemma for the attacker to choose between attack effectiveness and detection evasiveness. The empirical evaluation using benchmark datasets and representative attacks validates the efficacy of MDP.

Zhaohan Xi, Tianyu Du, Changjiang Li, Ren Pang, Shouling Ji, Jinghui Chen, Fenglong Ma, Ting Wang• 2023

Related benchmarks

TaskDatasetResultRank
Sentiment ClassificationSST2 (test)
Accuracy89.02
214
Sentiment ClassificationIMDB (test)--
144
Text GenerationMedical Chatbot
ASR82.56
42
Backdoor DefenseSST-2--
41
Text ClassificationCR
CA91.45
31
Backdoor DefenseMR
AUC0.99
20
Backdoor DefenseCR
AUC1
20
Backdoor DefenseSubj
AUC0.99
20
Backdoor DefenseTREC
AUC0.99
20
Text ClassificationSST-2
CA95.06
20
Showing 10 of 14 rows

Other info

Code

Follow for update