Projects

Current

Question-Answering (QA) algorithms and applications:
  • Developing Question-answering (QA) systems as voice assistants (e.g. Siri).
  • Developing a community Question-answering system, as well as vertical QA systems.
  • Developing backend of Android apps and chatting robots on the WeChat platform.

Graduate

Sentiment Analysis for Marketing Support of Corporations
2012/09 -- 2013/07

  • Designed a crawler to crawl microblog messages relevant to some corporations in Weibo (Sina's microblogging service).
    • Designed a separated proxy module for crawling without being blocked
    • Contributed in web-page parsing
  • Sentiment analysis on microblog messages based on lexicons.
    • Constructed lexicons automatically from the Weibo corpus
    • Lexicons include sentiments (positive, negative) and moods (emotions: happiness, anger, sadness, fear, suprise)
  • Built a system to visualize microblog messages and their sentiments (moods).

NExT: Internet Crawling, Archiving and Indexing. Visualization
2011/05 -- 2012/08

  • Built PrEV: A preservation and retrieval system for Web 2.0 user-generated contents.
    • Aggregated more than 50 resources (including web pages, microblogs, images, videos) from over ten research groups in National University of Singapore and Tsinghua University.
    • Live update every three minutes. Loosely coupled design of archiving, indexing and interface layers.
    • Occasionally available at http://sui.ddns.comp.nus.edu.sg/prev/
  • Built RESTful APIs to contribute to the research communities.
    • Key-based authentication, four levels of rate limiting and individual resource access control.
    • Forum-based user community management.
  • Built StrmWrd: HTML5 widgets for comparative keywords and trends visualization for the PrEV system.
  • Built OrgSense: A brand monitoring system on social media.
    • Organizes micro-posts of given organizations into coherent threads of discussions, identifies
      influential users of the organizations, and discovers emerging topics and their sentiments.
    • Contributed in system design, data crawling, sentiment analysis and visualization.
Software Interface Programs Compatible for Electronic Whiteboard from Multiple Manufacturers
2010/07 -- 2010/08
  • A Social Practice experience required for all Tsinghua PhDs. Group leader of 15 members.
  • Worked for Nanhao Company in Hengshui, Hebei Province
    • Developed software programs worked for running multiple utility software programs on Nanhao board, by
    • Studying the protocol of different software and boards, creating virtual serial ports and communications
Active Learning for Fast Drug Discovery
2009/09 -- 2010/03
  • Research project in Carnegie Mellon University, instructed by Jeff Schneider
  • Developed active learning algorithms based on multi-armed bandit machines for screening the best compound to accelerate predicting the performances of molecules
    The algorithms wisely select the most probable molecules to prevent from conducting plenty of biochemical experiments, thus save both time and money
Research on Chinese female skin conditions
2008 -- 2009
  • Lab research project collaborating with P&G Company
    • Mining key factors for Tone, Spots and Hydration of skin from questionnaires
    • Predict those targets with one's life habits
  • Contributed on core algorithm development and program interface design
Development of Prediction Software for Biochemical products based on Machine Learning Technology
2007 -- 2009
  • Including two lab research projects
    • Anti-germ performance prediction for detergents
    • Shelf-life prediction for personal beauty care products
  • Collaboration with P&G Company
  • Contributed on
    • core algorithm development for both projects
    • software program interface design and implementation for shelf-life project
SNS Website Development, ZeroLimits!
2008 -- 2009
  • One of the co-founders and core developers in ZeroLimits!
  • Contributed on SNS website organization, development and maintenance
Teaching assistant
2008 -- 2009
  • 2008/09 -- 2009/01: Discrete Maths(II)
    • Graph Theory and Algebra Structure
    • Required course for Computer Science Dept. undergraduates
    • In charge of both written and programming assignments
  • 2009/03 -- 2009/07: Microcomputer Technology
    • CPU, I/O interface, Pentium instructions, interrupt, DMA, etc.
    • Required elective course of Computer Science Dept. undergraduates
    • In charge of experiments, including ASM, FPGA and DMA

Undergraduate

Research on Image Recognition Technology for Industrial Robots
2007 -- 2008
  • Lab research project collaborated with ABB Company
  • Implemented an image recognition algorithm based on Fourier-Mellin Transform on TI's digital signal processor (DSP)
  • My bachelor’s thesis is based on this work
Automatic Template Extraction and Format-based Clustering on Dynamic Web-pages
2006 -- 2007
  • Tsinghua Student Research Training (SRT) Project
  • Project team: two students (including me) and two instructors.
  • Clustering similar web-pages and extract their template based on HTML tags
    In charge of format-based clustering algorithms
Intelligent Robot Based on Mobile Phone Platform
2006 -- 2007
  • Attend for Tsinghua Challenging Cup Contest (requires student-only members and outstanding innovations)
    See
    MRobot for details.
  • Project team: four students (including me). I'm the second author. The other students are from Department of Automation and Academy of Arts and Design.
    I contributed on software architecture design, multi-sensors combination, code implementation and debugging
  • The project won the Gold Medal and the Best Presentation Prize