PtbStolen: Pre-trained Encoder Stealing Through Perturbed Samples

Chuan Zhang, Haotian Liang, Zhuopeng Li, Tong Wu, Licheng Wang*, Liehuang Zhu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Recent years have witnessed the huge success of adopting the self-supervised learning paradigm into pre-train effective encoders [1].

Original languageEnglish
Title of host publicationEmerging Information Security and Applications - 4th International Conference, EISA 2023, Proceedings
EditorsJun Shao, Sokratis K. Katsikas, Weizhi Meng
PublisherSpringer Science and Business Media Deutschland GmbH
Pages1-19
Number of pages19
ISBN (Print)9789819996131
DOIs
Publication statusPublished - 2024
Event4th International Conference on Emerging Information Security and Applications, EISA 2023 - Hangzhou, China
Duration: 6 Dec 20237 Dec 2023

Publication series

NameCommunications in Computer and Information Science
Volume2004
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference4th International Conference on Emerging Information Security and Applications, EISA 2023
Country/TerritoryChina
CityHangzhou
Period6/12/237/12/23

Keywords

  • Encoder stealing attacks
  • Perturbed samples
  • Pre-trained encoders

Fingerprint

Dive into the research topics of 'PtbStolen: Pre-trained Encoder Stealing Through Perturbed Samples'. Together they form a unique fingerprint.

Cite this