TABATA Koji Associate Professor
Imagine that there are 10 slot machines in front of you.
You don’t know which machine has the highest probability of hitting.
If you can play these machines 1000 times in total, how do you play these machines?
This problem, called as a multi-armed bandit problem, models the dilemma of exploitation-exploration dilemma of knowledge.
I am interested in machine learning, especially multi-armed bandit problems and their extensions and applications.
- Koji Tabata, Atsuyoshi Nakamura, Tamiki Komatsuzaki, “Classification Bandits: Classification Using Expected Rewards as Imperfect Discriminators”, Pacific-Asia Conference on Knowledge Discovery and Data Mining: PAKDD 2021 International Workshops: PAKDD 2021 Workshops, WSPA, MLMEIN, SDPRA, DARAI, and AI4EPT, 57-69, 2021.
- Koji Tabata, Atsuyoshi Nakamura, Junya Honda and Tamiki Komatsuzaki, “A bad arm existence checking problem: How to utilize asymmetric problem structure?”, Machine Learning, 109(2), 327-372, 2020.
- Aurelien Pelissier, Atsuyoshi Nakamura, Koji Tabata,“Feature selection as Monte-Carlo Search in Growing Single Rooted Directed Acyclic Graph by Best Leaf”, SIAM International Conference on Data Mining, 450-458 (2019).