• Natural language processing
  • Safe AI - Uncertainty in machine learning, Outlier detection
  • Deep generative models

Education

  • The University of Texas at Austin, U.S. - M.S. in CS 2020-2022
  • The University of Tokyo, Japan 2020-2020
    • Department of Computer Science, Graduate School of Information Science and Technology
    • Akiko Aizawa Lab
  • Keio University, Japan - B.E. in Statistics 2016-2020
    • Department of Mathematics
    • Kei Kobayashi Lab
  • Carnegie Mellon University, Pennsylvania - Exchange student 2018-2019
    • Department of Statistics and Data Science

Work Experience

  • SenseTime Japan, Japan - Research Internship 2020
  • Datasection Inc, Japan - Research Internship 2017-2018
    • Research in natural language processing

Publications

Google Scholar

  • Ryo Kamoi, Kei Kobayashi. Why is the Mahalanobis Distance Effective for Anomaly Detection?. arXiv preprint arXiv:2003.00402. 2020. [pdf]
  • Ryo Kamoi, Kei Kobayashi. Out-of-Distribution Detection with Likelihoods Assigned by Deep Generative Models Using Multimodal Prior Distributions. AAAI Workshop on Artificial Intelligence Safety. 2020. [pdf]
  • Ryo Kamoi, Kei Kobayashi. Likelihood Assignment for Out-of-Distribution Inputs in Deep Generative Models is Sensitive to Prior Distribution Choice. arXiv preprint arXiv:1911.06515. 2019. [pdf]

Awards, Scholarships