Computer science is no more about computers than astronomy is about telescopes. -- E. W. Dijkstra [folklore :)]

About me


Me.

I am a Ph.D. student in the Department of Computer Science at Virginia Tech, under the supervision of Prof. Naren Ramakrishnan. My current project focuses on deep generative model, with an emphasis on time-series tabular data generation. I also widely investigate on related open problems in generative model, neural model attacks, and natural language, etc. Additionally, I've completed programming language focused projects with an emphasis on program syntax and semantics analysis.

Before starting my Ph.D, I received my bachelor degree from Beijing University of Post and Telecommunication in 2016. I spent summers as intern at Microcoft Research Asia, Google, Facebook, and Tsinghua National Laboratory.

Besides, I am also interested in graph theory and competitive algorithm contests. I'm an equipment enthusiast of MMA and cycling.

Office: 900 N Glebe Rd, Virginia Tech Research Center - Arlington,
Arlington VA 22203
Email: shengzx [AT] vt.edu

On- going Projects


stan.

AI Security at Commonwealth Cyber Initiative (CCI)
[1] Generating synthetic network data with generative neural models that is realistic enough to replace real data in machine learning tasks, without the privacy considerations.
The project has been deployed to the CCI AI testbed and everyone is very welcome to play and discuss on our project website. [website]
[2] Scenario, Behavior, and Event. We're rethinking a High-volume Time-series Generation Framework through Netflow Attacks. The project poster is going to participate the VASEM: Securing the Future of Cyberspace Summit at Washington DC on Oct 22, 2022.

Publications


STAN: Synthetic network traffic generation with generative neural models [pdf] [website] [software]
    Shengzhe Xu, Manish Marwah, Martin Arlitt, Naren Ramakrishnan
    MLHat: The 2nd International Workshop on Deployable Machine Learning for Security Defense
    co-located with 26TH ACM SIGKDD Conference on Knowledge Discovery and Data Mining (MLhat at KDD)
    Virtual Space, August 2021

Meditor: inference and application of API migration edits [pdf] [software]
    Shengzhe Xu, Ziqi Dong, and Na Meng
    Proceedings of the 27th International Conference on Program Comprehension (ICPC)
    Montreal, QC, Canada, May 2019

Internship Experience


Machine Learning Intern @Facebook
Ads Core ML
Time-series video-clip summarization
05/2020-8/2020, Menlo Park, US

Research Intern @Microsoft Research Asia
Deep Reinforcement Learning
05/2018-8/2018, Beijing, China

Software Engineer Intern @Google
Input Method Engine
07/2015-11/2015, Beijing, China

Research Assistant @Tsinghua National Laboratory
Tsinghua-Waterloo Joint Research Center for Internet Information Acquisition
Text Mining
01/2015-06/2015, Beijing, China

Presentation


Learning to coordinate with coordination graphs in repeated single-stage multi-agent ... , Summer18 MSRA [pdf]
Don't decay the learning rate, increase the batch size, Spring18 VT ICLR paper [pdf]
Poisoning attacks against support vector machines, Fall17 VT ICML paper [pdf]
Attacking speaker recognition with adversarial speech perturbations, Fall17 VT course project [pdf]
Automatically finding patches using genetic programming, Spring17 VT ICSE paper [pdf]
SVM-KMeans, a semi-supervised learning for outlier detection, Spring16 BUPT UnderGrad Thesis Defense [pdf]

Teaching Assistant


VT CS5560: Fundamentals of Information Security, Spring19
VT CS5704: Software Engineering, Spring18
VT CS1044: Introduction to Programming in C, Fall17
VT CS3114: Algorithm and Data Structure, Fall16

Honor


Award of excellence at Machine Learning group, MSRA 2018
Beijing Outstanding Graduate Student, BUPT 2016
Silver Medal, ACM-ICPC Asia Regionals Mudanjiang site, 2014