Sotetsu KOYAMADA

I am a Ph.D. candidate at Kyoto University. Also, I belong to Recruit Holdings Co., Ltd. as machine learning engieneer and National Institute of Advanced Industrial Science and Technology (AIST) as research assistant.
More details can be found in my cv.

Research interests

My primary research interest is reinforcement learning, and I am particularly interested in both theoretical and practical interface between its algorithms and other fields of machine learning. I am also interested in neural networks, natural language processing, and sensitivity analysis in general as well.

Contact

• Graduate School of Informatics, Kyoto University Yoshidahonmachi 36-1, sakyo-ku, Kyoto-city, Kyoto, Japan. 606-8501
• koyamada-s[at]sys.i.kyoto-u.ac.jp

Reinforcement learning book is now available (in Japanese)

This book is the Japanese translation of “Algorithms for Reinforcement Learning” by C. Szepesvári, published on Sep 21, 2017. I managed the entire translation project and wrote an additional chapter about deep reinforcement learning. The explained algorithms and techniques include DQN, double DQN, dueling architecture, prioritized experience replay, A3C, TRPO, GAE, and AlphaGo. [Kyoritsu Shuppan, Amazon.co.jp]

Education

• Apr 2015 - Present: Ph.D. student of Informatics, Kyoto University, Japan.
  Advisor: Shin Ishii
• Apr 2013 - Mar 2015: Master of Informatics, Kyoto University, Japan.
  Advisor: Shin Ishii
  Master thesis: Principal Sensitivity Analysis and Its Application to Knowledge Discovery in Functional Neuroimaging [pdf]
• Apr 2008 - Mar 2013: Bachelor of Economics, Kyoto University, Japan.
  Advisor: Masaaki Iiyama

Professional experience

• Aug 2016 - Present: Research assistant, National Institute of Advanced Industrial Science and Technology (AIST), Japan.
• Apr 2015 - present: Machine learning engineer, Recruit Holdings Co., Ltd., Japan.
• Oct 2013 - Mar 2015: Research intern, ATR Cognitive Mechanisms Laboratories, Japan.

Research

Books

• S. Koyamada et al.: Japanese translation of “Algorithms for Reinforcement Learning” by C. Szepesvári, Kyoritsu Shuppan.
  ^{[Kyoritsu Shuppan, Amazon.co.jp]}

Publications (refereed)

• S. Koyamada, Y. Kikuchi, A. Kanemura, S. Maeda, and S. Ishii: “Neural sequence model training via α-divergence minimization.” LGNL, ICML Workshop, 2017.
  ^{[arXiv][code]}
• S. Koyamada, M. Koyama, K. Nakae, and S. Ishii: "Principal sensitivity analysis." In Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), 621-632, 2015.
  ^{[publication][arXiv][slide]}
• S. Koyamada, Y. Shikauchi, K. Nakae, and S. Ishii: "Construction of subject-independent brain decoders for human fMRI with deep learning." The International Conference on Data Mining, Internet Computing, and Big Data, 60-68, 2014.

Pre-prints (not refereed)

• S. Koyamada, Y. Shikauchi, K. Nakae, M. Koyama, S. Ishii “Deep learning of fMRI big data: a novel approach to subject-transfer decoding.” arXiv:1502.00093, 2015.
  ^[arXiv]

Conference, workshop and other Events (not refereed)

• S. Koyamada: "Principal Sensitivity Analysis." Machine Learning Summer School 2015 Kyoto, Kyoto, 2015.9.1. (poster presentation)
• S. Koyamada, Y. Shikauchi, K. Nakae, M. Koyama, and S. Ishii: "Knowledge Discovery for Nonliner Classifier in Functional Neuroimaging." 10th AEARU Workshop on Computer Science and Web Technology, Tsukuba, 2015.2.26. (poseter presentation)
• S. Koyamada, Y. Shikauchi, K. Nakae, and S. Ishii: "Learning the subject-independent discriminative features from the largr-scale fMRI database." Neuro2014, Yokohama, 2014.9.13. (poseter presentation)

Skills

• Programming language: Python paper2tmb flashcard / Golang shakyo-golang goscholar gobibtex / C++ / Java / Scala (coursera) / R
• Deep learning framework: PyTorch alpha-dimt-icmlws / Chainer / TensorFlow
• Middlewaree/Infrastructure: Hadoop (coursera1, coursera2) / Spark (certificate) / RDBMS / AWS / GCP (certificate)
• Other tools: Git goscholar / SQL / LaTeX
• Language Japanese (native) / English

Certificates

• Jul 2016: "Functional Programming Principles in Scala." Coursera Verified Certificates [coursera]
• Jul. 4-6, 2016: "Cloudera Developer Training for Apache Spark." Cloudera, Inc. [pdf]
• Jun 30, 2016: "Google Cloud Platform Qualified Data Analyst." [pdf]
• Oct 2015: "Introduction to Big Data." Coursera Verified Certificates [coursera]
• Nov 2015: "Hadoop Platform and Application Framework." Coursera Course Certificates [coursera]

Interview

• Apr 2016: Interview as data scientist of Recruit Technologies Co., Ltd.
  ^{[interview1 (ja)][interview2 (ja)]}

Teaching

• Jul 23, 2014: Teaching assistant. Lecture sessions on deep learning, Kyoto University, Japan.
  ^{[slide (ja)]}
• Oct 2013 - Mar 2014: Teaching assistant. "Introduction to Computer Science," Kyoto University, Japan.
  ^{[course page (ja)]}

Blog

• Sep 2015: Report of KDD2015
  ^{[blog (ja)]}