【学术沙龙】NISL 12月01日活动预告 - TO BE ON AIR

清华大学网络与信息安全实验室学术沙龙，欢迎关注~ 

This is the Paper Reading Seminar of Network and Information Security Lab (NISL) at Tsinghua University. Tune in for more details!

* 本次分享对外公开直播，线上参会者要求实名备注“姓名-单位”

1.【论文分享】Hidden Trigger Backdoor Attack on NLP Models via Linguistic Style Manipulation

• Presenter: 邓楚云

• Conference: USENIX Security 2022

• Authors:

    LiuXudong Pan, Mi Zhang, Beina Sheng, Jiaming Zhu, and Min Yang

• Abstract:

  This paper proposes LISM (i.e., Linguistic Style-Motivated backdoor attack), a hidden trigger backdoor attack on NLP models which simultaneously achieves improved attack effectiveness (in terms of malicious semantic preservation) and improved attack stealthiness (in terms of the naturalness of generated triggers and successful evasion of existing detectors) over previous attacks.

• Link to paper:

    https://www.usenix.org/system/files/sec22-pan-hidden.pdff

2.【论文分享】GraphFuzz: Library API Fuzzing with Lifetime-aware Dataflow Graphs

• Presenter: 李哲铭

• Conference: ICSE 2022

• Authors:

  Harrison Green, Thanassis Avgerinos

• Abstract:

  This paper proposes GraphFuzz, which models sequences of executed functions as a dataflow graph, thus enabling it to perform graph-based mutations both at the data and at the execution trace level.

• Link to paper:

https://hgarrereyn.github.io/GraphFuzz/research/GraphFuzz_ICSE_2022.pdf