Pervasive Artificial Intelligence Research (PAIR) Labs

Computer Games

Pervasive Artificial Intelligence Research (PAIR) Labs

Computer Games

NCTU – Professor I-Chen Wu

Principal Investigator: Prof. I-Chen Wu

Research：

Studies of Applications with Deep Reinforcement Learning (DRL) Technologies

Core Technologies

This project focus on three classes of DRL applications:

1) Lightweight model, e.g., Go program CGI, tutoring applications and exact methods, and general AlphaZero methods.

2) Complex model, e.g., AI bot of video games.

3) Real-word model, e.g., self-driving model racing cars, robotics packing.

Applications

1.Training top-ranking players (e.g. Taiwan professional national Go players)
2.Computer game tutoring applications
3.AI bot for video games
4.Video game development and testing
5.Real-world applications (e.g. self-driving model racing, robotics packing)

Technology Transfer

1.The strength adjustment techinque for Go tutoring
2.The efficient AI bot training technique for video games
3.DRL related real-world applications (e.g. self-driving model racing, robotics packing)

Technologies with Potential for Technology Transfer or Industry-Academia Collaboration

1.The strength adjustment techinque for Go tutoring
2.The Go board evaluation techinque
3.The strength adjustment techinque for Chess
4.The tsumego judgement techinque
5.General AlphaZero training framework
6.The efficient AI bot training framework for video games
7.The human-like AI bot technique
8.AI bot for video game development and testing
9.Self-driving model racing cars training technique
10.DRL related robotics packing technique

Proposal for Industry-Academia Collaboration

DL/RL/DRL applications (e.g. the strength adjustment techinque for Go tutoring, the efficient AI bot training technique, and robotics DRL applications)