Research Interests 

My main research interest is developing AI for drug discovery.  In particular, I am focusing on developing new algorithms around Generative AI (= Diffusion models, LLMs) and Reinforcement Learning.  Here are some figures to describe my research.  

My primary focus is on AI-driven molecular design (DNA, RNA, proteins, small molecules, etc.) using diffusion/ language models.

1: Inference-Time Alignment through RL and Search in Diffusion Models.

Liner, Xi and Zhao, Yulai and Wang, Chenyu and Scalia, Gabriele and Eraslan, Gokcen and Nair, Surag and Biancalani, Tommaso and Regev, Aviv and Levine, Sergey and Uehara, Masatoshi (*) Derivative-Free Guidance in Continuous and Discrete Diffusion Models with Soft Value-Based Decoding arXiv preprint arXiv:2408.08252 

Masatoshi Uehara, Xingyu Su, Yulai Zhao, Xiner Li, Aviv Regev, Shuiwang Ji, Sergey Levine, Tommaso Biancalani. Reward-Guided Iterative Refinement in Diffusion Models at Test-Time with Applications to Protein and DNA Design    arXiv preprint arXiv:2502.14944 

Xiner Li, Masatoshi Uehara, Xingyu Su, Gabriele Scalia, Tommaso Biancalani, Aviv Regev, Sergey Levine, Shuiwang Ji.  Dynamic Search for Inference-Time Alignment in Diffusion Models   arXiv preprint arXiv:2503.02039

(morer works) 

These works are summarized at 

Uehara, M., Zhao, Y., Wang, C., Li, X., Regev, A., Levine, S., & Biancalani, T. (2025). Inference-Time Alignment in Diffusion Models with Reward-Guided Generation: Tutorial and Review. arXiv preprint arXiv:2501.09685.

2: Post-Training (RL from Laboratory Feedback) in Diffusion Models.

Chenyu Wang (*) and Masatoshi Uehara (*) and Yichun He and Amy Wang and Tommaso Biancalani and Avantika Lal and Tommi Jaakkola and Sergey Levine and Hanchen Wang, and Aviv Regev. Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design ICLR 2025 

Masatoshi Uehara (*), Yulai Zhao (*), Ehsan Hajiramezanali, Gabriele Scalia, Gökcen Eraslan, Avantika Lal, Sergey Levine, Tommaso Biancalani. Bridging Model-Based Optimization and Generative Modeling via Conservative Fine-Tuning of Diffusion Models  Neurips, 2024  

Masatoshi Uehara (*), Yulai Zhao (*), Kevin Black, Ehsan Hajiramezanali, Gabriele Scalia, Nathaniel Lee Diamant, Alex M Tseng, Sergey Levine (*), Tommaso Biancalani (*) Feedback Efficient Online Fine-Tuning of Diffusion Models   ICML, 2024 

Masatoshi Uehara (*), Yulai Zhao (*), Kevin Black, Ehsan Hajiramezanali, Gabriele Scalia, Nathaniel Lee Diamant, Alex M Tseng, Tommaso Biancalani (*), Sergey Levine (*). Fine-Tuning of Continuous-Time Diffusion Models as Entropy-Regularized Control arXiv preprint arXiv:2402.15194, 2024  

These works are summarized at 

Masatoshi Uehara (*), Yulai Zhao (*), Tommaso Biancalani, Sergey Levine. Understanding Reinforcement Learning-Based Fine-Tuning of Diffusion Models: A Tutorial and Review.  arXiv preprint arXiv:2407.13734

My old works during PhD are summarized as follows: 

• Sample-Efficient Offline Policy Evaluation:  Doubly robust and semiparametrically efficient estimators (1,2), Incorporating deep neural networks with minimax loss (1,2). Slide is here. 

• Robust Offline RL under Insufficient Coverage:  Model-based pessimistic offline RL. Talk is here. 

• Representation Learning in Online RL with Strategic Exploration: Model-based RL in low-rank MDPs.  Talk is here.

• Causal Inference + RL: Accounting for Unmeasured Confounders ([1], [2]), Data combination 

• Instrumental Variable Methods with Deep NNs: Semiparametric IV methods to estimate functionals without identification (+how to perform inference), Nonparametric IV methods without identification  Slide is here 

• Sample-Efficient RL in POMDPs: OPE with general function approximation,  Online RL with general function approximation, Computationally and statistically efficient PAC RL methods.  Slide is here. 

Publication

Red means I am the co-first/corresponding author. Blue means alphabetical order following the convention. The other papers follow the contribution-based ordering.

Conference Proceedings

Chenyu Wang (*) and Masatoshi Uehara (*) and Yichun He and Amy Wang and Tommaso Biancalani and Avantika Lal and Tommi Jaakkola and Sergey Levine and Hanchen Wang, and Aviv Regev. Fine-Tuning Discrete Diffusion Models via Reward Optimization with Applications to DNA and Protein Design ICLR 2025 

Yulai Zhao (*), Masatoshi Uehara (*), Gabriele Scalia, Tommaso Biancalani, Sergey Levine, Ehsan Hajiramezanali. Adding Conditional Control to Diffusion Models with Reinforcement Learning ICLR 2025

Masatoshi Uehara (*), Yulai Zhao (*), Ehsan Hajiramezanali, Gabriele Scalia, Gökcen Eraslan, Avantika Lal, Sergey Levine, Tommaso Biancalani. Bridging Model-Based Optimization and Generative Modeling via Conservative Fine-Tuning of Diffusion Models Neurips, 2024 

Masatoshi Uehara (*), Yulai Zhao (*), Kevin Black, Ehsan Hajiramezanali, Gabriele Scalia, Nathaniel Lee Diamant, Alex M Tseng, Sergey Levine (*), Tommaso Biancalani (*) Feedback Efficient Online Fine-Tuning of Diffusion Models   ICML, 2024 

Kuba, Jakub Grudzien, Masatoshi Uehara, Pieter Abbeel, and Sergey Levine. Functional Graphical Models: Structure Enables Offline Data-Driven Optimization. AISTATS 2024, 2024 

Wenhao Zhan (*),  Masatoshi Uehara (*), Nathan Kallus, Jason D. Lee,  Wen Sun. Provable Offline Reinforcement Learning with Human Feedback. ICLR  (Spotlight), 2024

Wenhao Zhan , Masatoshi Uehara, Wen Sun, and Jason D. Lee.  How to Query Human Feedback Efficiently in RL?.  ICLR (Spotlight), 2024

Masatoshi Uehara (*), Haruka Kiyohara (*), Andrew Bennett, Victor Chernozhukov, Nan Jiang, Nathan Kallus, Chengchun Shi, and Wen Sun. Future-Dependent Value-Based Off-Policy Evaluation in POMDPs. Neurips 2023 (Spotlight). (SLIDE)

Masatoshi Uehara, Nathan Kallus, Jason D. Lee, Wen Sun. Offline Minimax Soft-Q-learning Under Realizability and Partial Coverage  Neurips 2023. 

Kiyohara, Haruka, Masatoshi Uehara, Yusuke Narita, Nobuyuki Shimizu, Yasuo Yamamoto, and Yuta Saito. "Off-Policy Evaluation of Ranking Policies under Diverse User Behavior." In Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pp. 1154-1163. 2023.

Runzhe Wu, Uehara, Masatoshi, and Wen Sun. Distributional offline policy evaluation with predictive error guarantees. ICML, 2023.  (Code)

Masatoshi Uehara, Ayush Sekhari, Jason D. Lee, Nathan Kallus, Wen Sun. Computationally Efficient PAC RL in POMDPs with Latent Determinism and Conditional Embeddings.   arXiv preprint arXiv:2206.12081 ICML 2023.

Andrew Bennett, Nathan Kallus, Xiaojie Mao, Whitney Newey, Vasilis Syrgkanis, and Masatoshi Uehara. Minimax Instrumental Variable Regression and L2 Convergence Guarantees without Identification or Closedness.   COLT 2023.

Andrew Bennett, Nathan Kallus, Xiaojie Mao, Whitney Newey, Vasilis Syrgkanis, and Masatoshi Uehara. Inference on strongly identified functionals of weakly identified functions.   COLT 2023.

Wenhao Zhan, Masatoshi Uehara, Wen Sun, Jason D. Lee.  PAC Reinforcement Learning for Predictive State Representations   ICLR 2023

Masatoshi Uehara, Ayush Sekhari, Jason D. Lee, Nathan Kallus, Wen Sun. Provably Efficient Reinforcement Learning in Partially Observable Dynamical Systems.   Neurips 2022.

Chengchun Shi, Masatoshi Uehara, Jiawei Huang, and Nan Jiang. A minimax learning approach to off-policy evaluation in partially observable markov decision processes. ICML(Long presentation), 2022. (Slide  Code)

Xuezhou Zhang, Yuda Song, Masatoshi Uehara, Mengdi Wang, Wen Sun, and Alekh Agarwal. Efficient reinforcement learning in block mdps: A model-free representation learning approach.  ICML 2022.  Presented at RL THEORY VIRTUAL SEMINAR 2021 by Xuezhou. (Code) 

Masatoshi Uehara, Xuezhou Zhang, and Wen Sun. Representation learning for online and offline rl in low-rank mdps. ICLR (Spotlight), 2022. Oral Paper in Ecological Theory of Reinforcement Learning Workshop at Neurips.    (Talk  Slide)  

Masatoshi Uehara and Wen Sun. Pessimistic model-based offline rl: Pac bounds and posterior sampling under partial coverage. ICLR, 2022. Presented at RL THEORY VIRTUAL SEMINAR 2021. (Talk   Slide )

Jonathan D Chang (*), Masatoshi Uehara (*), Dhruv Sreenivas, Rahul Kidambi, and Wen Sun. Mitigating covariate shift in imitation learning via offline data without great coverage. Neurips, 2021.  (Code)

Nathan Kallus, Yuta Saito, and Masatoshi Uehara. Optimal off-policy evaluation from multiple logging policies. ICML, 2021.  (Code)

Yichun Hu, Nathan Kallus, and Masatoshi Uehara. Fast rates for the regret of offline reinforcement learning. COLT, 2021. Presented at RL THEORY VIRTUAL SEMINAR 2021/11/26 by Yichucn. (“Minor Revision” requested from Mathematics of Operations Research)

Masatoshi Uehara (*), Masahiro Kato (*), and Shota Yasui. Off-policy evaluation and learning for external validity under a covariate shift.  NeurIPS (Spotlight), 2020. (Talk Code)

Nathan Kallus and Masatoshi Uehara. Doubly robust off-policy value and gradient estimation for deterministic policies. NeurIPS, 2020. (Talk ) 

Masatoshi Uehara, Jiawei Huang, and Nan Jiang. Minimax weight and q-function learning for off-policy evaluation. ICML, 2020.  (Code)

Nathan Kallus and Masatoshi Uehara. Statistically efficient off-policy policy gradients.  ICML, 2020. 

Nathan Kallus and Masatoshi Uehara.  Double reinforcement learning for efficient and robust off-policy evaluation ICML, 2020.   (Code )

Masatoshi Uehara, Takeru Matsuda, and Jae Kwang Kim. Imputation estimators for unnormalized models with missing data. AISTATS, 2020.

Masatoshi Uehara, Takafumi Kanamori, Takashi Takenouchi, and Takeru Matsuda. Unified estimation framework for unnormalized models with statistical efficiency. AISTATS, 2020.

Nathan Kallus and Masatoshi Uehara. Intrinsically efficient, stable, and bounded off-policy evaluation for reinforcement learning.  NeurIPS, 2019.

Journal Articles 

Nathan Kallus, Xiaojie Mao, and Masaotshi Uehara. Localized debiased machine learning: Efficient estimation of quantile treatment effects, conditional value at risk, and beyond.  Journal of Machine Learning Research, 2023 

Nathan Kallus and Masatoshi Uehara. Efficient evaluation of natural stochastic policies in offline reinforcement learning. Biometrika, 2023 

Masatoshi Uehara, Danhyang Lee, and Jae Kwang Kim. Semiparametric response model with nonignorable nonresponse.  Scandinavian Journal of Statistics, 2023

Nathan Kallus and Masatoshi Uehara. Efficiently breaking the curse of horizon: Double reinforcement learning in infinite-horizon processes. Operations research,  2021.   (The version at Informs is here. But there are several typos there. )

Takeru Matsuda, Masatoshi Uehara, and Aapo Hyvarinen. Information criteria for non-normalized models. Journal of Machine Learning Research, 2021.

Nathan Kallus and Masatoshi Uehara.  Double reinforcement learning for efficient off-policy evaluation in markov decision processes. Journal of Machine Learning Research, 2020.  (Code)

Unpublished Articles Under Revision  

Masatoshi Uehara , Chengchun Shi, and Nathan Kallus. An overview of off-policy evaluation in reinforcement learning.  arXiv preprint arXiv:2212.06355 (“Minor Revision” requested from Statistical Science )

Masatoshi Uehara, Masaaki Imaizumi, Nan Jiang, Nathan Kallus, Wen Sun, and Tengyang Xie. Finite sample analysis of minimax offline reinforcement learning: Completeness, fast rates and first-order efficiency. arXiv preprint arXiv:2102.02981, 2021. (Rejection with Resubmission from Annals of Statistics)  (SLIDE )

Bennett, Andrew, Nathan Kallus, Xiaojie Mao, Whitney Newey, Vasilis Syrgkanis, and Masatoshi Uehara. Source Condition Double Robust Inference on Functionals of Inverse Problems. arXiv preprint arXiv:2307.13793 (2023)  (Revision from JRSSB) 

Nathan Kallus, Xiaojie Mao, and Masatoshi Uehara. Causal inference under unmeasured confounding with negative controls: A minimax learning approach. arXiv preprint arXiv:2103.14029, 2021.  (Revision from Journal of Machine Learning Research)

Recent Working Drafts

Masatoshi Uehara (*), Yulai Zhao (*), Tommaso Biancalani, Sergey Levine. Understanding Reinforcement Learning-Based Fine-Tuning of Diffusion Models: A Tutorial and Review.  arXiv preprint arXiv:2407.13734

Masatoshi Uehara (*), Yulai Zhao (*), Kevin Black, Ehsan Hajiramezanali, Gabriele Scalia, Nathaniel Lee Diamant, Alex M Tseng, Tommaso Biancalani (*), Sergey Levine (*). Fine-Tuning of Continuous-Time Diffusion Models as Entropy-Regularized Control arXiv preprint arXiv:2402.15194, 2024  

Zihao Li (*), Hui Lan (*), Vasilis Syrgkanis, Mengdi Wang, Masatoshi Uehara. Regularized DeepIV with Model Selection   arXiv preprint arXiv:2403.04236, 2024 

Masatoshi Uehara, Issei Sato, Masahiro Suzuki, Kotaro Nakayama, and Yutaka Matsuo. Generative adversarial nets from a density ratio estimation perspective. arXiv preprint arXiv:1610.02920, 2016. 

----  Tutorials/Talks -----

About Dynamic treatment regime

About 深層CRESTミーティング