Yuan Tan

Yuan Tan

Lanzhou University

📝 Publications

Linux Security Summit Europe 2024
sym

Safer Seccomp: Dead Syscalls Elimination

Yuan Tan, Zhangjin Wu, Xiao Liu

  • Contributed kernel patches to Linux. Discovered excitement and passion around kernels.
  • Utilized compiler directives to establish reference relationships for all orphaned sections.
  • The excessive use of ‘KEEP()’ in the Linux kernel can be eliminated using the above method.
  • Made the elimination of unused system calls possible through the construction of references between segments.
  • Reduced the size of the kernel image by 7%.
  • Reduced the attack surface that prevents attackers from exploiting trimmed syscalls.
  • The early version has received positive feedback from the maintainer.
Black Hat USA 2022

Human or Not: Can You Really Detect the Fake Voices?

Xin Liu, Yuan Tan, Rui Chong, Xiaokang Zhou

Co-first Author

  • Introduced SiF-DeepVC, an innovative voice clone attack framework.
  • Enhanced AI-synthesized speech to appear more “human” by merging real voice features with synthetic ones.
  • Rendered all existing methods for synthetic speech detection ineffective.
  • The first to underscore the formidable challenge of detecting AI-synthesized speech.
  • Was responsible for implementing the attack framework, reproducing existing detectors, and evaluating them.
ACL 2023

Not The End of Story: An Evaluation of ChatGPT-Driven Vulnerability Description Mappings

Xin Liu, Yuan Tan, Zhenghang Xiao, Jianwei Zhuge, Rui Zhou

  • Initiated a series of experiments that assessed ChatGPT’s performance across diverse cybersecurity domains.
  • Provided evidence that ChatGPT could not accurately classify CVEs.
ISSRE

Hidden-in-Wave: A Novel Idea to Camouflage AI-Synthesized Voices Based on Speaker-Irrelative Features

Xin Liu, Yuan Tan, Xuan Hai, Qingchen Yu, Qingguo Zhou

Co-first Author

  • Invalidated all existing methods for synthetic voice detection.