Kuang-Ting, Lee


I specialize in the fields of deep learning and computer vision. Look for the positions of machine learning engineer and data scientist. 

About Me

  • Master of Computer Science
  • 5+ years experience in programming
  • 2+ years experience in research & development of deep learning and computer vision 


  • National Chung Hsing University, Master Degree, Computer Science and Engineering (2019 - 2021) 

  • National Cheng Kung University, Bachelor Degree, Materials Science and Engineering (2013 - 2018)

Work Experience

  • HMSHost Kennebunk South Travel Plaza (June, 2017 - Sep., 2017)
    • Part-time Employee
    • Work & Travel USA 2017

Publications and Awards

  • 2020 International Computer Symposium (ICS)
    • K. -T. Lee, C. -Y. Chiou and C. -R. Huang, "One-Class Novelty Detection via Sparse Representation with Contrastive Deep Features," 2020 International Computer Symposium (ICS), 2020, pp. 61-66, doi: 10.1109/ICS51289.2020.00022.
  • First Place Award

    • International ICT Innovative Services Awards 2020

    • Few-Shot Self-Supervised Learning for AI-based Automatic Optical Inspection in Industry 4.0 

  • Second Place Award

    • International ICT Innovative Services Awards 2019

    • AI based Traffic Condition Recognition and Visualization in Modern Smart City 

Professional Skills

  • Operating Systems
    • Windows, Linux
  • Programming Languages

    • Python, Java,  C++, C, Matlab, SQL

  • Libraries and Frameworks
    • PyTorch, Tensorflow, Keras, Scikit-learn, OpenCV
  • Development Tools

    • PyCharm, Google Colab, Jupyter, Git 

Research Fields

  • Master Thesis
    • Learning Contrastive Features

      for One-Class Novelty Detection

  • Deep Learning
    • One-class classification
    • Contrastive learning
    • Data augmentation
    • Transformer
  • Machine Learning
    • Dictionary learning


  • Deep Learning and Its Applications    
  • Advanced Data Mining and Big Data Analysis 
    • Hadoop, Spark, K-means, Adaboost
  • Database Systems
    • Relational database, SQL
  • Computer Vision
    • Camera calibration, Descriptor, OpenCV
  • Digital Image Processing
    • Panorama, Wavelet image transformation, Morphology, Image restoration and reconstruction
  • Pattern Recognition
    • Bayesian decision theory, Derivation of the Back-Propagation algorithm


  • Learning Contrastive Features for One-Class Novelty Detection
    • Objective
      • Train a one-class classifier to identify unseen images from corrupted datasets
    • Contributions
      • Propose a novel data augmentation to clearly isolate features of negative samples
      • Outperform the detection performance of state-of-the-art methods reported in CVPR, NIPS on MNIST, CIFAR10, CIFAR100 benchmarks
    • Development tools
      • Use PyTorch to build the network architecture and the pipeline of training and evaluation 
      • Use Scikit-learn to visualize the extracted features


  • English — Advance
    • TOEFL iBT 80
  • Mandarin — Native speaker

Contact Info