Intelligent Information Processing
We conduct research on intelligent information processing using computers, focusing on such fields as "natural language processing," "game AI," and "reinforcement learning. Our research ranges from the development of new algorithms and deep learning methods to the construction of practical information processing systems.
Research field 1
Natural language processing
We conduct research to enable computers to freely handle "natural languages" such as English and Japanese. Our research ranges from basic natural language processing techniques such as part-of-speech inference and syntactic analysis to applications such as machine translation and question answering, as well as how to express the meaning of words using computers.
Research field 2
Game AI
Realizing game players on a computer can be a great help in building models of human intellectual activity. We are investigating new methods, including algorithmic and statistical methods, for a variety of games such as Shogi, Go, and real-time strategy games.
Research field 3
Reinforcement Learning
Reinforcement learning, which allows agents to acquire better behavioral strategies by repeated trial-and-error in the environment using "rewards" as cues, is attracting attention as a central technology for realizing next-generation AI. In our laboratory, we are studying the algorithms and applications of reinforcement learning.