About Me

A research scientist working at ACES, Inc. I’m passionate about deep learning and its applications.

• 2022 Ph.D. (Engineering) Graduate School of Engineering, University of Tokyo (Yutaka Matsuo’s Lab)
• 2019 M.S. (Engineering) Graduate School of Engineering, University of Tokyo
• 2017 B.A. (Economics) Department of Economics, University of Tokyo

Preprint

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Multimodal Sequential Generative Models for Semi-Supervised Language Instruction Following. arXiv preprint.

Conference Papers

• Kei Akuzawa, Kotaro Onishi, Keisuke Takiguchi, Kohki Mametani, Koichiro Mori. Conditional Deep Hierarchical Variational Autoencoder for Voice Conversion. APSIPA ASC, December, 2021.

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Information-theoretic regularization for learning global features by sequential VAE. In Proc. of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, online, September, 2021.

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Estimating Disentangled Belief about Hidden State and Hidden Task for Meta-Reinforcement Learning. In Proc. of the Third Learning for Dynamics and Control (L4DC) Conference, online, 2021. [arxiv].

• Yusuke Iwasawa, Kei Akuzawa, Yutaka Matsuo. Stabilizing Adversarial Invariance Induction from Divergence Minimization Perspective. In Proc. of the Twenty-Ninth International Joint Conference on Artificial Intelligence (IJCAI), 2020.

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Adversarial Invariant Feature Learning with Accuracy Constraint for Domain Generalization. In Proc. of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, Würzburg, Germany, September, 2019 [arxiv(preprint), code].

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo, Expressive speech synthesis via modeling expressions by variational autoencoder, in Proc. Interspeech2018, Hyderabard，India, September, 2018 [arxiv, Demo].

Journal Papers

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Information-theoretic regularization for learning global features by sequential VAE, Mach Learn (2021). https://doi.org/10.1007/s10994-021-06032-4
• Yuki Saito, Kei Akuzawa, and Kentaro Tachibana, “Joint adversarial training of speech recognition and synthesis models for many-to-one voice conversion using phonetic posteriorgrams,” IEICE Transactions on Information and Systems, Vol. 103-D, No. 9, pp. xxx–xxx, Sep. 2020. (ACCEPTED)

Workshop Papers

• Kei Akuzawa, Yusuke Iwasawa, Yutaka Matsuo. Adversarial Feature Learning under Accuracy Constraint for Domain Generalization. International Conference of Learning Representation (ICLR19) Limited Label Data Workshop, 2019. (spotlight, 22/66)
  • openreview
• Yusuke Iwasawa, Kei Akuzawa, Yutaka Matsuo. Invariant Feature Learning by Attribute Perception Matching. International Conference of Learning Representation (ICLR19) Limited Label Data Workshop, 2019.
  • openreview

Lecture

• 2020 University of Tokyo, Graduate School of Engineering. “Deeplearning”, lecturer in the 2nd class.

Awards and Grants

• Toyota/Dwango Scholarship for Advanced Artificial Intelligence Researcher, Apr. 2020 - Mar. 2021
• Grant-in-Aid for JSPS Research Fellow, the Japan Society of the Promotion of Science (JSPS), Apr. 2020 - Mar. 2022.
• Toyota/Dwango Scholarship for Advanced Artificial Intelligence Researcher, Apr. 2019 - Mar. 2020
• Student Incentive Award, The Japanese Society for Artificial Intelligence, 2018.
• Specially Selected Bachelor Thesis Award, from the department of economics, The University of Tokyo, 2017.

Other Works

• Slides in paper reading club (Japanese Only)