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


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


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

Computer Science > Machine Learning

Title: BackdoorBench: A Comprehensive Benchmark of Backdoor Learning

Abstract: Backdoor learning is an emerging and important topic of studying the vulnerability of deep neural networks (DNNs). Many pioneering backdoor attack and defense methods are being proposed successively or concurrently, in the status of a rapid arms race. However, we find that the evaluations of new methods are often unthorough to verify their claims and real performance, mainly due to the rapid development, diverse settings, as well as the difficulties of implementation and reproducibility. Without thorough evaluations and comparisons, it is difficult to track the current progress and design the future development roadmap of the literature. To alleviate this dilemma, we build a comprehensive benchmark of backdoor learning, called BackdoorBench. It consists of an extensible modular based codebase (currently including implementations of 8 state-of-the-art (SOTA) attack and 9 SOTA defense algorithms), as well as a standardized protocol of a complete backdoor learning. We also provide comprehensive evaluations of every pair of 8 attacks against 9 defenses, with 5 poisoning ratios, based on 5 models and 4 datasets, thus 8,000 pairs of evaluations in total. We further present analysis from different perspectives about these 8,000 evaluations, studying the effects of attack against defense algorithms, poisoning ratio, model and dataset in backdoor learning. All codes and evaluations of BackdoorBench are publicly available at \url{this https URL}.
Comments: 30 pages, 14 figures, 10 tables
Subjects: Machine Learning (cs.LG); Cryptography and Security (cs.CR)
Cite as: arXiv:2206.12654 [cs.LG]
  (or arXiv:2206.12654v1 [cs.LG] for this version)

Submission history

From: Baoyuan Wu [view email]
[v1] Sat, 25 Jun 2022 13:48:04 GMT (16322kb,D)

Link back to: arXiv, form interface, contact.