Chuyi Kong

Chuyi Kong

CV

Research Interests

  • Controlled Text Generation - Persona Control
  • Fraud Detection

Research Experience

2023/05/08 - now Working as a Research Assistant at The Chinese University of Hong Kong, Shenzhen

  • Reconstructed the Self-Instruct pipeline code for the team
  • Conducted the research on how LLMs can be applied to the user simulator
  • Papers: PlatoLM: Teaching LLMs via a Socratic Questioning User Simulator
    • arxiv: https://arxiv.org/abs/2308.11534v4
    • code: https://github.com/FreedomIntelligence/PlatoLM
    • model: https://huggingface.co/FreedomIntelligence/PlatoLM-7B
      • has achieved SOTA performance from 2023/08 to 2023/10 in Alpaca-Eval and MT-Bench.
    • dataset: https://huggingface.co/datasets/FreedomIntelligence/SocraticChat

04/2019 Construction of Logistics Service Quality Model for Used E-Commerce Platforms (Xianyu App)

  • Preliminarily constructed the evaluation index system in terms of SERVQUAL Model, the overall problems of logistics services of second-hand E-commerce platform, and the factors affecting quality
  • Built the final evaluation index system by Exploratory Factor Analysis (SPSS) through the first questionnaire survey about the logistics services quality of second-hand E-commerce platform
  • Carried out the second questionnaire survey on the Xianyu App to evaluate and analyze the logistics service quality by means of Fuzzy Hierarchical Comprehensive Evaluation (SPSS)
  • Awarded an outstanding thesis at the College level

Education

M.S. in Business and Data Analytics - Information System Stream

  • City University of Hong Kong, 2021
  • GPA: 3.57/4.3
  • Graduated with distinction
  • Awards: Outstanding Student Gold Award with the highest GPA(3.72/4.3) in Sem A at Major Level
  • Courses: Data Mining, Big Data & Social Media Analytics, Advanced Software Construction, Statistical Data Analysis, Database Management Systems, Applied Linear Statistical Modelling
  • Domestic Certification: Statistics

B.S. in Management Science & Engineering - Supply Chain Direction

  • Beijing Normal University, Zhuhai, 2019
  • GPA: 3.42/4
  • Major Rank: 27/225 (top 12%)
  • Awards: Excellent Graduation Thesis at College Level, Third Prize Scholarship at College Level
  • Courses: Calculus, Operations Research, Linear Algebra, Applied Statistics, Forecasting and Decision Making, Logistics Information Systems, Probability Theory and Mathematical Statistics

Skills

  • Programming languages: python, golang, Java, cpp, R, JavaScript, html, css
  • Database: mysql, redis, sql-server, mongodb
  • OS: Linux(Centos/Ubuntu)
  • Libraries/Packages: pytorch, libtorch, numpy, pandas, sklearn, flask, sqlalchemy, django, drf, gin, gorm
  • ML/DL Model: lr, xgb, lgb, catboost, resnet, lstm, gru, transformer, Bert, autoEncoder, GAN
  • Software: SPSS, SASEM, Tableau

Project Experience

01/2023 Online learning blog community construction

  • Responsible for web backend implementation
  • Used Technologies: gin + gorm + machinery + swagger + mysql + redis + vue

09/2022-12/2022 Server-side Chatbot “Big Head” v2.0

  • Responsible for model training, back-end implementation, partial front-end implementation
  • Used Technologies: ubuntu + git + pytorch + numpy + matplotlib + django + vue + mysql + redis + celery
  • Used Model: seq2seq(encoder: bert + transformer + gru, decoder: gru + transformer + gpt2)
  • Link
  • Alternate Link

05/2022-06/2022 Food Classification and Network Compression

  • Responsible for code optimization, model training (Kaggle competition)
  • Used Technologies: numpy + pytorch + PIL + tqdm + matplotlib
  • Used Model: Resnet, MobileNet
  • Rank: TOP 4% (56/1406)

06/2021-07/2021 Multiple classifications of Telstra network service outage severity

  • Responsible for code writing, model training (Kaggle competition)
  • Used Technologies: numpy + pytorch + tqdm + matplotlib + Bayesian optimization + grid search cross-validation
  • Used Model: lightgbm, catboost
  • Rank: TOP 20% (limited by computing capacity)

Project Overview

Development

  • python: Built full-stack chatbots, back-end web Q&A platform
  • golang: Built small instant messaging system, back-end blog community
  • java: Built family income and expenditure bookkeeping system, customer information management system, development team staff scheduling system

Algorithm

  • ML: regression, binary classification (rf, lr, fcn), multiclassification (xgb, lgb, catboost)
  • CV: image recognition (vgg16, resnet18-101), image generation (dcgan, wgan, wgan-gp), anomaly face detection (fcn-ae, cnn-ae, resnet18-ae, vae)
  • NLP: extractive question-and-answer prediction (bert), conversational bots (gru+luong attention), sentiment analysis
  • Speech: phoneme recognition (fcn), speaker recognition (transformer, conformer)
  • Other: transfer learning (DaNN), model compression (knowledge distillation, deep separable networks)

Database

  • mysql: designed database with query optimization

Language Skills

  • JLPT-N2
  • CET-6

Hobbies

  • Animate, Comic, Game, Novel