
Traditionally, machine learning can be categorized into three types: Supervised learning, unsupervised learning and reinforcement learning. Recently, the hybrid of reinforcement learning and deep learning has made big contribution into the field of automatic decision making.

We're interested in using modern deep learning techniques into the real-world applications. For example, we're involved in studying healthcare, culture technology and autonomous car problems.

Games naturally produce tons of data from game players and the big data help game designers understand their customers. Especially, game companies attempt to predict game players' future behaviors such as churn and purchase.

There are three interesting research topics in game AI field: Game AI Player, Game Contents Generation, and Game Player Modeling.