• Ph.D. in Science (Machine Learning) from the University of Tokyo in 2023 (Supervisor: Prof. Issei Sato)

• Special Postdoctoral Researcher in the Imperfect Information Learning Team (PI: Prof. Masashi Sugiyama) at RIKEN Center for Advanced Intelligence Project (RIKEN AIP)

• Researcher at ACT-X Program, JST (Oct. 2021 - Present)

• I’m an alumni of the Issei Sato Laboratory andSugiyama-Yokoya-Ishida Laboratory

• I previously worked as a research fellow at JSPS (DC2; Apr. 2021 - Mar. 2023)

Research Interests

• Robustness, generalization, and stability in approximate inference
• Likelihood-free inference (e.g., approximate Bayesian computation)
• Bayesian inference

CV

My CV is here.

News

• Sep. 22, 2023: Our papar “Time-Independent Information-Theoretic Generalization Bounds for SGLD” has been accepted by NeurIPS2023! We sincerely appreciate the anonymous reviewers and ACs for their insightful feedback.

• Mar. 23, 2023: I have received a Ph.D. degree and Dean’s award for outstanding achievement! I would like to express my gratitude to Prof. Issei Sato, Prof. Masashi Sugiyama, all lab members, and my family.

• Mar. 12, 2023: I gave a presentation on our γ-ABC paper at the satellite event “Bayesian computing without exact likelihoods” held at BayesComp2023 in Levi, Finland. I am grateful to Prof. Robert for inviting me!

• Jul. 14, 2022: I was selected as the outstanding reviewer of ICML2022 (top 10% of reviewers)! Thanks to ICML2022 area chairs for selecting me.

• Mar. 8, 2022: Our presentation at IBIS2021 got the excellent presentation award (3 / 127 presentations)!

• Dec. 1, 2021: Our paper “Multilevel Monte Carlo Variational Inference” has been accepted by Journal of Machine Learning Research (JMLR)!

• Oct. 15, 2021: I was selected to receive a NeurIPS 2021 Outstanding Reviewer Award (top 8% of NeurIPS2021 reviewers)! Thanks to NeurIPS2021 area chairs for selecting me.

• Sep. 22, 2021: My research proposal has been accepted by JST ACT-X Program! The details of the accepted proposals are here (Japanese only).

• Apr. 10, 2021: Our γ-ABC method has been merged into the ABCpy Python library for Likelihood-Free Inference 🎉 Thank you so much, Lorenzo (University of Oxford)! [RIKEN AIP News]