cv

General Information

Full Name Pengyang Xie
Date of Birth 2nd September 2000
Languages Mandarin, English

Education

  • 2022 - 2023
    Master of Research
    Imperial College London, London, United Kingdom
    • Artificial Intelligence and Machine Learning.
    • Is self-supervised pre-training at odds with unsupervised learning?
  • 2018 - 2022
    Bachelor of Engineering
    Beihang University, Beijing, China
    • Computer Science and Technology (Major).
    • Applied Mathematics (Minor).

Experience

  • 2021
    Summer Research Intern
    Institute of Neuroscience, Chinese Academy of Sciences, China
    • Supervised by Prof. Min Xu.
    • Analyzed widefield calcium imaging data of murine models using seqCNMF, PCA, ICA, and other algorithms and extracted spatiotemporal characteristics.
    • Analyzed functional connectivity of cerebral ROIs, facial movement, and pupil characteristics during different sleep stages and air-puff stimuli.
  • 2021
    Research Intern
    Institute of Computing Technology, Chinese Academy of Sciences, China
    • Supervised by Prof. Yiqiang Chen.
    • Participated in the related work of multimodal collaborative representation. Based on five multimodal data of vision, hearing, haptic feedback, EMG, and pose, designed an online algorithm to prove that multimodal data has higher reliability than single-modal data in judging human behavior.
  • 2019-2020
    Natural Language Processing Intern
    Innovation Center for Big Data and Brain Computing, Beihang University, China
    • Supervised by Prof. Richong Zhang.
    • Participated in the journal extension for Syntax Encoding with Application in Authorship Attribution, proposed a novel strategy to encode the syntax parse tree of a sentence into a learnable distributed representation
    • Independently reproduced the experimental results of the paper on the personal computer without source code in PyTorch.

Projects

  • 2021
    A4 Paper Method Based English Words Memory Website
    • Developed an English vocabulary learning and memory website, JQKey, based on the A4 Paper method, which connects the vocabs with a specific scene on an A4 size paper or screen called Word Map. Designed algorithms that recommended new words intelligently and built an online community for sharing their Word Maps with nearly 1000 users.
    • Used Vue, echarts, and fabrics to realize the Word Map function and the user's page, and used JavaScript to realize the animation.
  • 2020
    A compiler for the C0 programming language
    • Implemented a compiler supporting error handling for generating MIPS assembly code according to a given C0 grammar. It supports registers allocation and some optimization mechanisms, including copy propagation, simplification of multiplication and division, and modification of for-statement and while-statement.
  • 2020
    Cloud Computing Based Personal Deep Learning Platform
    • Supervised by Prof. Hailong Yang.
    • Developed an online high-throughput computing platform for deep learning based on HTCondor, a job scheduler. It utilizes computing resources rented from the Alibaba cloud. The platform encapsulates the process of creating and terminating computing instances, job operation, and feedback and refines the computing cost to a single task through a bidding strategy.

Other Interests

  • Hobbies: Badminton (even broke my ACL for it), hiking (there're lots of great parks in London, isn't it?), ACG.