Xie Yuxi’s Home Page

I am a senior undergraduate student at Peking University, pursuing my double major in the specialties of 1) Data Science and Big Data Technology in the School of Electronic Engineering and Computer Science (EECS) and 2) Economics in the National School of Development (NSD). More about my research interest can be found in my Research Statement.

EDUCATION BACKGROUND

  • School of Electronics Engineering and Computer Science, Peking University (Sept. 2016- Jun. 2020)
    B.S., Data Science and Big Data Technology
    Coursework Big Data Management: 99/100; Empirical Methods in Natural Language Processing: 95/100; Digital Image Processing: 95/100; Numerical Methods: 92/100

  • National School of Development, Peking University (Sept. 2017- Jun. 2020)
    B.S., Double Degree in Economics
    Coursework Econometrics: 91/100; Internet Finance and Big Data: 90/100

RESEARCH EXPERIENCE

  • OpenNQG: Integrated Toolkit of Neural Question Generation (NQG) (Mar. 2019- Present)
    Research Intern, Web IR / NLP Group, National University of Singapore (WING-NUS)
    Advisor Prof. Min-Yen Kan
    Reference Yuxi Xie*, Liangming Pan*, Min-Yen Kan. A Package for Neural Question Generation from WING-NUS[Source Code]. https://github.com/YuxiXie/OpenNQG
    • Developed an integrated Python package to provide fair comparison among current SOTA models of NQG
    • Designed specific mechanisms (e.g., masked self-attention) for further development of models
  • Open Domain Semantic Parsing, NLPCC 2019 Shared Task 2 (Apr.- Aug. 2019)
    Research Intern, Institute of Computer Science and Technology, Peking University (ICST-PKU)
    Advisor Prof. Yansong Feng
    Reference Li Zechang, Yuxuan Lai, Yuxi Xie, Yansong Feng and Dongyan Zhao. A Sketch-Based Systemfor Semantic Parsing. 8th CCF International Conference on Natural Language Processing and Chinese Computing (NLPCC2019)
    • Designed algorithms to do semantic parsing step-by-step in a sketch-based way
    • Conducted experiments to test and improve the semantic parsing system
  • Leveraging Semantic Graph for Muti-Hop Reasoning Question Generation (Jul. 2019- Present)
    Research Intern, WING-NUS and ICST-PKU
    Advisor Prof. Min-Yen Kan and Prof. Yansong Feng
    Reference Liangming Pan*, Yuxi Xie*, Min-Yen Kan, Yansong Feng. Combine Graphical and Sequential Information to Generate Multi-Hop Reasoning Questions [Source Code].https://github.com/YuxiXie/GQG
    • Designed and built a model to combine graphical semantic and statistic contextual information to generate questions which contain reasoning among multiple pieces of text
    • Designed and conducted multi-task training, to leverage context selection for better question generation
  • Leveraging Reinforcement Learning to Generate Better Questions (Nov. 2019- Present)
    Research Intern, ICST-PKU
    Advisor Prof. Yansong Feng
    • Aimed at the problem of Exposure Bias (and Evaluation Discrepancy) in Question Generation
    • Proposed to use QA-model as a discriminator to enhance the answerability of generated questions

PROFESSIONAL EXPERIENCE

  • Research Intern, Knowledge Computing Group, Microsoft Research Asia (KC-MSRA) (Oct. 2019- Jun. 2020)
    Mentor Dr. Jinpeng Wang
    • BERT-based Multimodal Learning (Oct. 2019- Present): Developed a BERT-based model to jointly learn representations for multimodal information from slide data
  • Summer Intern, Department of Intelligent Cloud Technology, Baidu, Beijing (Jun.- Oct. 2018)
    • Built and conducted experiments to test and improve the proposed models (including Logistic Regression, Gradient Boosting Decision Tree and XGBoost) of credit card classification for precision marketing
  • Summer Intern, Innovation Team, Business Big Data (BBD), Chengdu (Jun.- Aug. 2017)
    • Proposed a detailed plan to implement the intelligent precision marketing for the government

AWARDS and HONORS

  • People Choice Awards
    • Google Girl Hackathon, Google, Beijing (Mar. 2019)
    • Designed and implemented a Deep Q Network (DQN) based artificial intelligence agent to play board game Splendor

SKILLS

Programming Language excel in Python, excellent in C / C++, familiar with Java, Matlab and Stata
Framework and Library adept in PyTorch, familiar with Keras and Sciki-learn, good at NumPy, NLTK, etc.
Operating System good at Windows and Linux, familiar with Mac
Documentation adept in LaTex, Google Docs and Microsoft Office, familiar with Jupyter Notebook

SERVICE and LEADERSHIP

  • Deputy Minister, Department of Career Development Club of China Center for Economic Research(CCER), Peking University (Sept. 2018 - Jun. 2019)
    • Recruited and led three organizing teams (5 for each on average)
    • Reached out more than 10 alumnus and invited them to give a lecture for on-campus students about their work experience in the field of finance and economics
    • Led the teams to organize more than 5 lectures (3 guests for each on average)
  • Volunteer Teacher, Primary School of Fengyan Village, Guilin (Jan. - Feb. 2017)
    • Taught English for a class of 60 students for 2 weeks