Academic Experience & Results

 
 
 
 
 
September 2018 – December 2019
Advisor : Dr. Gang Wang and Dr. Bimal Viswanath

Detecting Bots with Limited Data using Neural Data Augmentation

Virginia Tech


  • Our paper is accepted by IEEE S&P(Oakland) 2020!

          Paper PDF

  • Build a stream-based real-time bot detection system to complement with rule-based method to catch ad-vanced bots.
  • Develop a data synthesis method to enable effective model training with limited labeled data.
  • Validate our system using real-world datasets from 3 different online services.
  • Sustain our system over a long period of time with low-cost retraining.
  • Explore adversarial machine learning and transfer learning on bot detection
 
 
 
 
 
September 2017 – May 2018
Advisor : Dr. Zhanyu Ma

Undergraduate Thesis: Noise Robust Speech Feature Extraction

Beijing University of Posts and Telecommunications


  • Build a multi-task adversarial network based noise-robust feature extractor.
  • Extract bottleneck features from the extractor to enhance the robustness of speaker recognition system.

          PDF

 
 
 
 
 
January 2017 – June 2018
Advisor : Dr. Jia Jia

Affective Computing and Machine Learning

Tsinghua University


  • Construct a benchmark social media dataset for online depression detection and analysis.
  • Extract six groups of discriminant depression oriented features to describe users from different aspects.
  • Detect depression via social media using multi-modal dictionary learning model.
  • Extract and process acoustic and textual information from a large-scale internet voice dialogue dataset.
  • Infer emotion from both acoustic and textual features of voice data by combining an Emotion-oriented Bimodal Deep Autoencoder with a Deep Sparse Neural Network.

          PDF

Publications

. Throwing Darts in the Dark? Detecting Bots with Limited Data using Neural Data Augmentation. To appear in IEEE S&P, 2020.

PDF

. Multi-Task Adversarial Network Bottleneck Features for Noise-Robust Speaker Verification. In IC-NIDC, 2018.

PDF DOI

. Depression detection via harvesting social media: A multimodal dictionary learning solution. In IJCAI-17, 2017.

PDF

Contact

  • tianruihu@vt.edu
  • 2202 Kraft Drive, Knowledge Works II, RM 2202, Blacksburg, VA