SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2305.14910 · 2023

From Shortcuts to Triggers: Backdoor Defense with Denoised PoE

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Language models are often at risk of diverse backdoor attacks, especially data poisoning. Thus, it is important to investigate defense solutions for addressing them. Existing backdoor defense methods mainly focus on backdoor attacks with explicit triggers, leaving a universal defense against various backdoor attacks with diverse triggers largely unexplored. In this paper, we propose an end-to-end ensemble-based backdoor defense framework, DPoE (Denoised Product-of-Experts), which is inspired by the shortcut nature of backdoor attacks, to defend various backdoor attacks. DPoE consists of two models: a shallow model that captures the backdoor shortcuts and a main model that is prevented from learning the backdoor shortcuts. To address the label flip caused by backdoor attackers, DPoE incorporates a denoising design. Experiments on SST-2 dataset show that DPoE significantly improves the defense performance against various types of backdoor triggers including word-level, sentence-level, and syntactic triggers. Furthermore, DPoE is also effective under a more challenging but practical setting that mixes multiple types of trigger.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2305.14910")
get_code_for_paper("2305.14910")
have("2305.14910")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2305.14910.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2305.14910.svg)](https://syntology.ai/paper/2305.14910)

Connect an agent — have() is free.