Current Projects

  • Recommendation Systems (Nov. 2017 to present)
    • Practical Recomm. Systems with billions of users at ByteDance
  • Modeling user sequence behavior (Jul. 2017 to present)
    • Understand how to model sequence behavior and enhance multiple applications such spammer detection

Previous Projects

  • Machine reading comrehension (Mar. 2017 to Nov. 2017)
    • Build an AI bot which can pass the medical licensing examination
  • Smart question answering (Oct. 2016 to Nov. 2017)
    • Deep learning based question answering with attentions
    • Answer clustering and generation
    • Serve multiple products in Baidu
  • Mining large scale temporal-spatial data (Apr. 2015 to Oct. 2016)
    • Site optimization for offline stores, e.g., Starbucks
    • Audience extension and customer recommendations
  • Wisdom of crowdsourced stock platform SeekingAlpha (Sep. 2013 to Jan. 2016)
    • Sentiment analysis on Seeking Alpha articles and StockTwits messages
    • Practical investment strategy based on crowdsourced articles
    • Understanding and detecting stock manipulation
  • Measuring anonymous OSN Whisper (Feb. 2014 to May 2014)
    • Topic modeling for abusive contents in Whisper
    • Modeling and prediction of user engagement and stickness
  • Detecting malicious crowdsourced workers (Apr. 2013 to Feb. 2014)
    • Modeling malicious crowdsourced worker behavior
    • Detection with supervised machine learning methods
    • Study the robustness in adversary scenarios
  • Measurement of Sina Weibo (Jul. 2012 to Aug. 2013)
    • Measure the unique function of comments
    • Reveal how function impacts the user behavior
  • Graph Sampling (Mar. 2010 to Jul. 2011)
    • Propose algorithms of unbiased sampling in directed graphs.
    • Evaluation beyond degree distribution.
  • GNU radio platform (Sep. 2009 to Dec. 2009)
    • Write a program for calculating the SNR of the radio.
    • Compare the work of GNU radio in different conditions.