Junpei Komiyama

Junpei Komiyama

Assistant Professor at MBZUAI. Research: Machine Learning, Multi-armed Bandits, Monte Carlo Tree Search, Algorithmic Bias

Publications

You can also find my articles on my Google Scholar profile.

Conference Papers

Fixed Confidence Best Arm Identification in the Bayesian Setting

Published in Advances in Neural Information Processing Systems (NeurIPS), 2024

This paper addresses fixed confidence best arm identification in the Bayesian setting, providing theoretical analysis and practical algorithms for Bayesian bandit problems.

Recommended citation: Jang, K., Komiyama, J., & Yamazaki, K. (2024). "Fixed Confidence Best Arm Identification in the Bayesian Setting." In Advances in Neural Information Processing Systems (NeurIPS 2024).
Download Paper

Learning Fair Division from Bandit Feedback

Published in International Conference on Artificial Intelligence and Statistics (AISTATS), 2024

This paper develops algorithms for learning fair division mechanisms from bandit feedback, combining fairness considerations with sequential decision-making.

Recommended citation: Yamada, H., Komiyama, J., Abe, K., & Iwasaki, A. (2024). "Learning Fair Division from Bandit Feedback." In International Conference on Artificial Intelligence and Statistics (AISTATS 2024).
Download Paper

Posterior Tracking Algorithm for Classification Bandits

Published in International Conference on Artificial Intelligence and Statistics (AISTATS), 2023

This paper develops posterior tracking algorithms for classification bandit problems, providing efficient methods for learning classifiers in sequential decision-making settings.

Recommended citation: Tabata, K., Komiyama, J., Nakamura, A., & Komatsuzaki, T. (2023). "Posterior Tracking Algorithm for Classification Bandits." In International Conference on Artificial Intelligence and Statistics (AISTATS 2023).
Download Paper

Thresholded Linear Bandits

Published in International Conference on Artificial Intelligence and Statistics (AISTATS), 2023

This paper introduces thresholded linear bandits, extending linear bandit theory to scenarios where only thresholded feedback is available.

Recommended citation: Mehta, N., Komiyama, J., Nguyen, A., Potluru, V., & Grant-Hagen, M. (2023). "Thresholded Linear Bandits." In International Conference on Artificial Intelligence and Statistics (AISTATS 2023).
Download Paper

High-dimensional Contextual Bandit Problem without Sparsity

Published in Advances in Neural Information Processing Systems (NeurIPS), 2023

This paper addresses high-dimensional contextual bandit problems without sparsity assumptions, providing theoretical guarantees and practical algorithms.

Recommended citation: Komiyama, J., & Imaizumi, M. (2023). "High-dimensional Contextual Bandit Problem without Sparsity." In Advances in Neural Information Processing Systems (NeurIPS 2023).
Download Paper

Anytime Capacity Expansion in Medical Residency Match by Monte Carlo Tree Search

Published in International Joint Conference on Artificial Intelligence (IJCAI), 2022

This paper applies Monte Carlo tree search for flexible-capacity mechanism design in medical residency matching, addressing NP-Complete optimization problems.

Recommended citation: Abe, K., Komiyama, J., & Iwasaki, A. (2022). "Anytime Capacity Expansion in Medical Residency Match by Monte Carlo Tree Search." In Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence (IJCAI 2022).
Download Paper

Minimax Optimal Algorithms for Fixed-Budget Best Arm Identification

Published in Advances in Neural Information Processing Systems (NeurIPS), 2022

This paper develops minimax optimal algorithms for fixed-budget best arm identification, providing theoretical guarantees and practical implementations.

Recommended citation: Komiyama, J., Tsuchiya, T., & Honda, J. (2022). "Minimax Optimal Algorithms for Fixed-Budget Best Arm Identification." In Advances in Neural Information Processing Systems (NeurIPS 2022).
Download Paper

Scaling Multi-Armed Bandit Algorithms

Published in ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2019

This paper addresses scalability challenges in multi-armed bandit algorithms, developing methods for handling large-scale bandit problems efficiently.

Recommended citation: Fouche, E., Komiyama, J., & Bohm, K. (2019). "Scaling Multi-Armed Bandit Algorithms." In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2019).
Download Paper

Nonconvex Optimization for Regression with Fairness Constraints

Published in International Conference on Machine Learning (ICML), 2018

This paper develops nonconvex optimization methods for regression problems with fairness constraints, addressing algorithmic bias in machine learning.

Recommended citation: Komiyama, J., Takeda, A., Honda, J., & Shimao, H. (2018). "Nonconvex Optimization for Regression with Fairness Constraints." In Proceedings of the 35th International Conference on Machine Learning (ICML 2018).
Download Paper

Statistical Emerging Pattern Mining with Multiple Testing Correction

Published in ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), 2017

This paper develops statistical methods for emerging pattern mining with multiple testing correction, providing rigorous statistical guarantees for pattern discovery.

Recommended citation: Komiyama, J., Ishihata, M., Arimura, H., Nishibayashi, T., & Minato, S. (2017). "Statistical Emerging Pattern Mining with Multiple Testing Correction." In Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD 2017), 897-906.
Download Paper

Position-based Multiple-play Bandit Problem with Unknown Position Bias

Published in Advances in Neural Information Processing Systems (NIPS), 2017

This paper addresses position-based multiple-play bandit problems with unknown position bias, providing theoretical analysis and practical algorithms.

Recommended citation: Komiyama, J., Honda, J., & Takeda, A. (2017). "Position-based Multiple-play Bandit Problem with Unknown Position Bias." In Advances in Neural Information Processing Systems 30 (NIPS 2017), 5005-5015.
Download Paper

Copeland Dueling Bandit Problem: Regret Lower Bound, Optimal Algorithm, and Computationally Efficient Algorithm

Published in International Conference on Machine Learning (ICML), 2016

This paper provides comprehensive analysis of the Copeland dueling bandit problem, including regret lower bounds, optimal algorithms, and computationally efficient implementations.

Recommended citation: Komiyama, J., Honda, J., & Nakagawa, H. (2016). "Copeland Dueling Bandit Problem: Regret Lower Bound, Optimal Algorithm, and Computationally Efficient Algorithm." In Proceedings of the 33rd International Conference on Machine Learning (ICML 2016), 1235-1244.
Download Paper

Regret Lower Bound and Optimal Algorithm in Finite Stochastic Partial Monitoring

Published in Advances in Neural Information Processing Systems (NIPS), 2015

This paper establishes regret lower bounds and develops optimal algorithms for finite stochastic partial monitoring, extending bandit theory to partial feedback settings.

Recommended citation: Komiyama, J., Honda, J., & Nakagawa, H. (2015). "Regret Lower Bound and Optimal Algorithm in Finite Stochastic Partial Monitoring." In Advances in Neural Information Processing Systems 28 (NIPS 2015), 1792-1800.
Download Paper

Optimal Regret Analysis of Thompson Sampling in Stochastic Multi-armed Bandit Problem with Multiple Plays

Published in International Conference on Machine Learning (ICML), 2015

This paper provides optimal regret analysis of Thompson sampling in stochastic multi-armed bandit problems with multiple plays, establishing theoretical guarantees for Bayesian bandit algorithms.

Recommended citation: Komiyama, J., Honda, J., & Nakagawa, H. (2015). "Optimal Regret Analysis of Thompson Sampling in Stochastic Multi-armed Bandit Problem with Multiple Plays." In Proceedings of the 32nd International Conference on Machine Learning (ICML 2015), 1152-1161.
Download Paper

Regret Lower Bound and Optimal Algorithm in Dueling Bandit Problem

Published in Conference on Learning Theory (COLT), 2015

This paper establishes regret lower bounds and develops optimal algorithms for dueling bandit problems, providing fundamental theoretical contributions to preference-based learning.

Recommended citation: Komiyama, J., Honda, J., Kashima, H., & Nakagawa, H. (2015). "Regret Lower Bound and Optimal Algorithm in Dueling Bandit Problem." In Proceedings of the 28th Annual Conference on Learning Theory (COLT 2015), 1141-1154.
Download Paper

Journal Articles

Strategic Choices of Migrants and Smugglers in the Central Mediterranean Sea

Published in PLoS ONE, 2024

This paper analyzes the strategic interactions between migrants and smugglers in the Central Mediterranean Sea using game-theoretic and machine learning approaches.

Recommended citation: Pham, K. H., & Komiyama, J. (2024). "Strategic Choices of Migrants and Smugglers in the Central Mediterranean Sea." PLoS ONE. Vol. 19 (No. 4) e0300553.
Download Paper

Rate-Optimal Bayesian Simple Regret in Best Arm Identification

Published in Mathematics of Operations Research, 2024

This paper establishes rate-optimal bounds for Bayesian simple regret in best arm identification problems, providing theoretical guarantees for Bayesian bandit algorithms.

Recommended citation: Komiyama, J., Ariu, K., Kato, M., & Qin, C. (2024). "Rate-Optimal Bayesian Simple Regret in Best Arm Identification." Mathematics of Operations Research. Vol. 49 (No.3), 1629-1646.
Download Paper

Finite-time Analysis of Globally Nonstationary Multi-Armed Bandits

Published in Journal of Machine Learning Research, 2024

This paper provides finite-time analysis for globally nonstationary multi-armed bandit problems, extending theoretical guarantees to dynamic environments.

Recommended citation: Komiyama, J., Fouche, E., & Honda, J. (2024). "Finite-time Analysis of Globally Nonstationary Multi-Armed Bandits." Journal of Machine Learning Research. Vol. 25 (No. 112), 1-56.
Download Paper

Bridging Offline and Online Experimentation: Constraint Active Search for Deployed Performance Optimization

Published in Transactions of Machine Learning Research, 2022

This paper develops methods for bridging offline and online experimentation through constraint active search, enabling effective performance optimization in deployed systems.

Recommended citation: Komiyama, J., Malkomes, G., Cheng, B., & McCourt, M. (2022). "Bridging Offline and Online Experimentation: Constraint Active Search for Deployed Performance Optimization." Transactions of Machine Learning Research.
Download Paper