Wonkyung Lee received the BE degree in Electrical & Electronic Engineering from Yonsei University, Seoul, South Korea, in 2019. He is currently working toward the PhD degree in Electrical & Electronic Engineering department from Yonsei University. His recent research focuses on building efficient deep learning models for visual tasks. He is also interested in developing alternative self-attention mechanisms for visual recognition.

Profile

  • Undergraduate of Electrical & Electronics Engineering at Yonsei Univ. (2011.03 ~)
  • Research & Development, ESTsoft, Seoul, South Korea (2015.11 ~ 2018.08)
  • Ph.D candidate at the School of Electrical & Electronic Engineering, Yonsei Univ. (2019.09 ~ )

Publications

  • (TPAMI2020) Learning Semantic Correspondence Exploiting an Object-level Prior. [Paper] [Project]
    • Junghyup Lee*, Dohyung Kim*, Wonkyung Lee, Jean Ponce, and Bumbsub Ham (*equal contribution)
  • (ECCV2020) Learning with Privileged Information for Efficient Image Super-Resolution. [Paper] [Project]
    • Junghyup Lee*, Dohyung Kim*, Wonkyung Lee*, and Bumbsub Ham (*equal contribution)

Interesting Areas

  • Computer vision
  • Efficient deep learning
  • Manifold learning
  • Domain adaptation

Work Experiences

  • Research & Development, ESTsoft, Seoul, South Korea (2015.11 ~ 2018.08)
    • Malware Detection and Prediction
    • Price Optimization
    • Context-based Image Retrieval
  • Development & Service Planning at Startup team Soodal
    • Calendar-SNS service, WhenD [pdf]

Patents

  • [Domestic] 특허 제 10-1863615호 : 신경망 학습 기반의 변종 악성 코드를 탐지하기 위한 장치, 이를 위한 방법 및 이 방법을 수행하기 위한 프로그램이 기록된 컴퓨터 판독 가능한 기록매체 [pdf]
  • [International] PCT/KR2018/005866 : DEVICE FOR DETECTING VARIANT MALICIOUS CODE ON BASIS OF NEURAL NETWORK LEARNING, METHOD THEREFOR, AND COMPUTER-READABLE RECORDING MEDIUM IN WHICH PROGRAM FOR EXECUTING SAME METHOD I S RECORDED [pdf]

Programming Skills

  • Python, PyTorch, Tensorflow
  • JAVA, Android
  • C, C++

Presentations

  • Deep Information Retrieval for Malware Searching System [pdf] @ Blackhat USA 2018