Yuu Jinnai

I am a researcher at CyberAgent AI Lab, focusing on natural language processing, reinforcement learning, and language model alignment. My work spans from theoretical foundations to practical applications in text generation and AI safety.

Contact: ddyuudd [at] gmail [dot] com
Links: Google Scholar | GitHub | CV

Biography

• Jun. 2023 - Present Researcher, CyberAgent AI Lab
• Apr. 2020 - Jan. 2023 Engineer, Lily MedTech Inc.
• Summer 2019 Research Intern, Microsoft Research Cambridge, UK
• Jun. 2017 - Jan. 2020 Ph.D. student (ABD), Computer Science, Brown University
  Advisor: George Konidaris
• Mar. 2017 - May 2017 Technical Staff, RIKEN Center for Advanced Intelligence Project (AIP)
• Mar. 2017 M.A., Graduate School of Arts and Sciences, University of Tokyo
  Advisor: Alex Fukunaga
• Mar. 2015 B.S., University of Tokyo
  Advisor: Alex Fukunaga

Research Interests

Artificial Intelligence, Reinforcement Learning, Language Model Alignment, Text Generation, Classical Planning, Heuristic Search

Current Research Focus

• Minimum Bayes Risk Decoding - Efficient and effective text generation methods
• Language Model Alignment - Aligning LLMs with human values and preferences

Selected Publications

View all publications

Google Scholar

Highlighted Work

	Best Paper Award: Jinnai Y. 2024. Does Cross-Cultural Alignment Change the Commonsense Morality of Language Models? C3NLP Workshop at ACL 2024. PAPER TALK
	Jinnai Y, et al. 2024. Model-based Minimum Bayes Risk Decoding. ICML-24. PAPER CODE TALK
	Y. Jinnai, et al. 2020. Exploration in Reinforcement Learning with Deep Covering Options. ICLR-20. PAPER

---

All Publications

2025

	Yuu Jinnai and Ukyo Honda. 2025. Annotation-Efficient Preference Optimization for Language Model Alignment. In Findings of the Association for Computational Linguistics (EMNLP-25 Findings). PAPER CODE TALK
	Yuki Ichihara, Yuu Jinnai, Kaito Ariu, Tetsuro Morimura, Eiji Uchibe. 2025. Theoretical Guarantees for Minimum Bayes Risk Decoding. Annual Meeting of the Association for Computational Linguistics (ACL-25). PAPER
	Ayuto Tsutsumi, Yuu Jinnai. 2025. Do Large Language Models Know Folktales? A Case Study of Yokai in Japanese Folktales. In Findings of the Association for Computational Linguistics (ACL-25 Findings). PAPER CODE DATASET
	Yuu Jinnai. 2025. Document-Level Text Generation with Minimum Bayes Risk Decoding using Optimal Transport. Annual Meeting of the Association for Computational Linguistics (ACL-25). PAPER CODE TALK
	Ichihara, Y., Jinnai, Y., Morimura, T., Ariu, K., Abe, K., Sakamoto, M., & Uchibe, E. (2025). Evaluation of Best-of-N Sampling Strategies for Language Model Alignment. Transactions on Machine Learning Research (TMLR) PAPER CODE TALK
	Jinnai, Y., Morimura, T., Ariu, K., & Abe, K. (2024). Regularized Best-of-N Sampling with Minimum Bayes Risk Objective for Language Model Alignment. 2025 Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL-25) PAPER CODE TALK

2024

	Morimura, T., Sakamoto, M., Jinnai, Y., Abe, K., & Ariu, K. (2024). Filtered Direct Preference Optimization. The 2024 Conference on Empirical Methods in Natural Language Processing. (EMNLP-24) PAPER CODE
	Jinnai Y. 2024. Does Cross-Cultural Alignment Change the Commonsense Morality of Language Models? Proceedings of the 2nd Workshop on Cross-Cultural Considerations in NLP (C3NLP Workshop at ACL 2024). Best Paper Award. PAPER TALK MODEL DATASET
	Jinnai Y, Morimura T, Honda U, Ariu K, Abe K. Model-based minimum bayes risk decoding. Proc. 41st International Conference on Machine Learning. (ICML-24) PAPER CODE TALK
	Jinnai Y, Ariu K. Hyperparameter-Free Approach for Faster Minimum Bayes Risk Decoding. In Findings of the Association for Computational Linguistics. (ACL-24 Findings) PAPER CODE TALK
	Jinnai Y, Honda U, Morimura T, Zhang P. Generating Diverse and High-Quality Texts by Minimum Bayes Risk Decoding. In Findings of the Association for Computational Linguistics. (ACL-24 Findings) PAPER CODE TALK
	Ohashi A, Honda U, Morimura T, Jinnai Y. 2024. On the True Distribution Approximation of Minimum Bayes-Risk Decoding. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics. (NAACL-24) PAPER CODE TALK

2021

	Lecarpentier E, Abel D, Asadi K, Jinnai Y, Rachelson E, Littman Michael L. 2021. Lipschitz Lifelong Reinforcement Learning. Proc. 35th AAAI conference on Artificial Intelligence (AAAI-21) arXiv Poster CODE

2020

	Y. Jinnai, J. Park, M.C. Machado, and G.D. Konidaris. Exploration in Reinforcement Learning with Deep Covering Options. Accepted, Proceedings of the Eighth International Conference on Learning Representations. (ICLR-20) PAPER
	Wang L*, Zhao Y*, Jinnai Y, Tian Y, Fonseca R. 2020. AlphaX: eXploring Neural Architectures with Deep Neural Networks and Monte Carlo Tree Search. Proc. 34th AAAI conference on Artificial Intelligence (AAAI-20) *These authors contributed equally to this work. PAPER CODE

2019

	Jinnai Y. Park JW, Abel D, Konidaris G. 2019. Discovering Options for Exploration by Minimizing Cover Time. Proc. 36th International Conference on Machine Learning. (ICML-19) PAPER CODE TALK
	Jinnai Y, Abel D, Hershkowitz E, Littman M, Konidaris G. 2019. Finding Options that Minimize Planning Time. Proc. 36th International Conference on Machine Learning. (ICML-19) PAPER CODE TALK
	Jinnai Y, Abel D, Park JW, Hershkowitz E, Littman M, Konidaris G. 2019. Skill Discovery with Well-Defined Objectives. ICLR Worshop on Structure and Priors in Reinforcement Learning. PAPER
	Abel D, Arumugam D, Asadi K, Jinnai Y, Littman M, Wong L. S. 2019. State Abstraction as Compression in Apprenticeship Learning. Proc. 33rd AAAI Conference on Artificial Intelligence (AAAI-19). PAPER CODE

2018

	Abel D*, Jinnai Y*, Guo Y, Konidaris G, Littman M. 2018. Policy and Value Transfer for Lifelong Reinforcement Learning. Proc. 35th International Conference on Machine Learning. (ICML-18) *These authors contributed equally to this work. PAPER POSTER CODE TALK by D. Abel
	Fukunaga A, Botea A, Jinnai Y, Kishimoto A. 2018. Parallel A* for State-Space Search. Handbook of Parallel Constraint Reasoning, Youssef Hamadi, Lakhdar Sais (eds.), Springer. ISBN 978-3-319-63515-6. BOOK

2017

	Jinnai Y, Fukunaga A. 2017. A Graph-Partitioning Based Approach for Parallel Best-First Search. ICAPS 2017 Workshop on Heuristic and Search for Domain-Independent Planning (HSDIP). PAPER SLIDES CODE
	Jinnai Y, Fukunaga A. 2017. Learning to Prune Dominated Action Sequences in Online Black-box Planning. Proc. 31st AAAI Conference on Artificial Intelligence. (AAAI-17) PAPER SLIDES CODE
	Jinnai Y, Fukunaga A. 2017. On Hash-Based Work Distribution Methods for Parallel Best-First Search. Journal of Artificial Intelligence Research. (JAIR) PAPER CODE
	(Preprint) Fukunaga A., Botea A, Jinnai Y., Kishimoto A. 2017. A Survey of Parallel A*. arXiv 1708.05296 PAPER

2016

	Jinnai Y, Fukunaga A. 2016. Automated Creation of Efficient Work Distribution Functions for Parallel Best-First Search. Proc. 19th International Conference on Automated Planning and Scheduling. (ICAPS-16) PAPER SLIDES VIDEO CODE
	Jinnai Y, Fukunaga A. 2016. Abstract Zobrist Hashing: An Efficient Work Distribution Method for Parallel Best-First Search. Proc. 30th AAAI Conference on Artificial Intelligence. (AAAI-16) PAPER POSTER CODE (PDDL) CODE (sliding-tile, path-finding, MSA)

Grants and Scholarships

• 2015-2017 JASSO scholarship with particularly outstanding academic achievements (approx. $21,000)
• 2017 Department of System Sciences: Grants for Doctoral Students Attending International Conferences (AAAI-17)
• 2016 Initiative on Promotion of Supercomputing for Young or Women Researchers, Supercomputing Division, Information Technology Center, The University of Tokyo
• 2016 NEC C&C Foundation: Grants for Researchers Attending International Conferences (ICAPS-16)
• 2016 Department of System Sciences: Grants for Doctoral Students Attending International Conferences (AAAI-16)

Invited Talks

• Jun. 2025. Introduction to Minimum Bayes Risk Decoding. NLP Colloquium. SLIDES
• Jan. 2018. Automated Deep Learning by Neural Architecture Search. National Institute of Information and Communications Technology, Japan.
• Feb. 2017. Graph search algorithms for classical planning. Discrete Structure Manipulation System Project, Hokkaido University, Japan.

Awards and Honors

• Best Paper Award. 2024. Does Cross-Cultural Alignment Change the Commonsense Morality of Language Models? C3NLP Workshop at ACL 2024.
• Graduated summa cum laude. 2017. Ichiko Memorial Award, Graduate School of Arts and Sciences, University of Tokyo.

Patents

• Medical information provision device, ultrasonic ct imaging device, and medical information provision system. Google Patents
• Tumor detection algorithm for ultrasound computed tomography. Google Patents
• Spiculated mass detection algorithm for ultrasound computed tomography. Google Patents
• Motion artifact detection for ultrasound computed tomography. Google Patents
• Faster ultrasound computed tomography by ultrasound signal separation. Google Patents

Services

• Reviewer: ICML, NeurIPS, AAAI, ICLR, ACL Rolling Review
• Journal Reviewer: Journal of Machine Learning Research, Journal of Artificial Intelligence Research, Knowledge-based Systems
• Program Committee: 3rd Workshop on Cross-Cultural Considerations in NLP (C3NLP), 2nd Workshop on Uncertainty-Aware NLP (UncertaiNLP)