We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CL

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computation and Language

Title: Rethink the Evaluation for Attack Strength of Backdoor Attacks in Natural Language Processing

Abstract: It has been shown that natural language processing (NLP) models are vulnerable to a kind of security threat called the Backdoor Attack, which utilizes a `backdoor trigger' paradigm to mislead the models. The most threatening backdoor attack is the stealthy backdoor, which defines the triggers as text style or syntactic. Although they have achieved an incredible high attack success rate (ASR), we find that the principal factor contributing to their ASR is not the `backdoor trigger' paradigm. Thus the capacity of these stealthy backdoor attacks is overestimated when categorized as backdoor attacks. Therefore, to evaluate the real attack power of backdoor attacks, we propose a new metric called attack successful rate difference (ASRD), which measures the ASR difference between clean state and poison state models. Besides, since the defenses against stealthy backdoor attacks are absent, we propose Trigger Breaker, consisting of two too simple tricks that can defend against stealthy backdoor attacks effectively. Experiments show that our method achieves significantly better performance than state-of-the-art defense methods against stealthy backdoor attacks.
Subjects: Computation and Language (cs.CL); Cryptography and Security (cs.CR); Machine Learning (cs.LG)
Cite as: arXiv:2201.02993 [cs.CL]
  (or arXiv:2201.02993v2 [cs.CL] for this version)

Submission history

From: Lingfeng Shen [view email]
[v1] Sun, 9 Jan 2022 12:34:12 GMT (675kb,D)
[v2] Wed, 16 Feb 2022 07:10:29 GMT (675kb,D)

Link back to: arXiv, form interface, contact.