Jianmin Guo

Jianmin Guo

Ph.D candidate

Tsinghua University

Biography

My name is Jianmin Guo (郭建敏). I am a Ph.D. candidate at School of Software Engineering, Tsinghua University, Beijing, China. I received the BS degree in software engineering from Beijing University of Posts and Telecommunications, Beijing, China, in 2017. My research interests are software testing, mainly focusing on deep learning testing, adversarial testing of recurrent neural networks and audio adversarial attacks, etc.

Interests

  • Deep learning testing
  • Adversarial testing of recurrent neural networks
  • Audio adversarial attacks

Education

  • PhD in Software Engineering, 2017-Present

    Tsinghua University

  • BSc in Software Engineering, 2013-2017

    Beijing University of Posts and Telecommunications

Preprints

RNN-Test: Towards Adversarial Testing for Recurrent Neural Network Systems

First step towards testing for majority sequential domains of RNN systems, i.e. tasks with sequential outputs without class labels
Jianmin Guo, Yue Zhao, Quan Zhang, Yu Jiang
Last updated on Jan 18, 2021

Publications

Coverage Guided Differential Adversarial Testing of Deep Learning Systems

Coverage guided testing work based on DLFuzz
Jianmin Guo, Yue Zhao, Yu Jiang, Houbing Song
PDF

Leopard: Identifying vulnerable code for vulnerability assessment through program metrics

Static analysis tool for discovering vulnerabilities through program metrics
Xiaoning Du, Bihuan Chen, Yuekang Li, Jianmin Guo, Yaqin Zhou, Yang Liu, Yu Jiang
PDF

Dlfuzz: Differential fuzzing testing of deep learning systems

Differential fuzzing framework for deep learning testing
Jianmin Guo, Yu Jiang, Yue Zhao, Quan Chen, Jiaguang Sun
PDF Code

Experience

 
 
 
 
 

Ph.D. candidate

Tsinghua University

Sep 2017 – Present Beijing

Core works include:

  • RNN-Test: Design the testing framework for Recurrent Neural Networks (RNN), the first step towards sequential tasks like language modelling, speech recognition, etc.
  • DLFuzz: Design the deep learning testing framework, combining knowledge of fuzz testing and producing more adversarial input images with higher neuron coverage w.r.t. DeepXplore (SOSP’17).
 
 
 
 
 

Research assistant

Nanyang Technological University

Jan 2017 – May 2017 Singapore
Leopard: Extract and assess program metrics of tested software programs to identify vulnerability-prone functions and CVEs, assisted with the fuzzing tools.

Accomplish­ments

Reviewer

Mar 2018

Mainly for

  • 软件学报
  • American Journal of Software Engineering and Applications (AJSEA)

Contact

  • guojm17@mails.tsinghua.edu.cn
  • +86 18811082936
  • Tsinghua University, Beijing, 100084