Kento Nozawa is a 5th year Ph.D. student who does research on machine learning. His current research interest is self-supervised representation learning via theoretical tools such as PAC-Bayes theory. He is supervised by Dr. Issei Sato at Issei Sato Lab at The University of Tokyo. He is also a part-time research assistant at RIKEN AIP. During summer 2019, he visited UCL AI Centre and Inria Lille Nord Europe Modal team to work on PAC-Bayes and contrastive representation learning. In order to live, he has been working for Optuna as a part-time machine learning engineer at Preferred Networks, Inc. since April 2021.
Hopefully, he’ll be graduating in Sept. 2022, so he is on the job market.
Journal / Conference papers
- Han Bao, Yoshihiro Nagano and Kento Nozawa. On the Surrogate Gap between Contrastive and Supervised Losses. In ICML, pages 1585–1606, 2022.
arXiv.Alphabetical ordering and equal contribution.
- Kento Nozawa and Issei Sato. Evaluation Methods for Representation Learning: A Survey. In IJCAI-ECAI Survey Track, pages 5556–5563, 2022.
slides.Extended version: ``Empirical Evaluation and Theoretical Analysis for Representation Learning: A Survey’’. 2022.
- Kento Nozawa and Issei Sato. Understanding Negative Samples in Instance Discriminative Self-supervised Representation Learning. In NeurIPS, pages 5784–5797, 2021.
- Kento Nozawa, Pascal Germain and Benjamin Guedj. PAC-Bayesian Contrastive Unsupervised Representation Learning. In UAI, pages 21–30, 2020.
- Atsunori Kanemura, Yuhsen Cheng, Takumi Kaneko, Kento Nozawa and Shuichi Fukunaga. Imputing Missing Values in EEG with Multivariate Autoregressive Models. In EMBC, pages 2639–2642, 2018.
The full publication list is available on Google Scholar.