Student research opportunities

Artificial Intelligence for playing video games

Project Code: CECS_851

This project is available at the following levels:
CS single semester, Honours, Summer Scholar, Masters

Keywords:

Video Games, Artificial Intelligence, Computer Vision, Machine Learning, Planning, Knowledge Representation, Reasoning

Supervisors:

Dr Charles GRETTON
Dr Stephen Gould
Assoc Professor Jochen Renz

Outline:

The task of the project is to develop an AI system that can autonomously play video games such as Angry Birds. In the case of Angry Birds, the AI system must solve a number of different AI problems:
1. Computer Vision:
detect the location and category of objects
2. Machine Learning:
a. learn trajectories (where to shoot from in order to hit a given location)
b. learn properties of object categories
3. Knowledge Representation and Reasoning:
a. develop a suitable representation of objects
b. infer the consequences of a shot depending on the physiscs of the game
4. Planning:
determine the best sequence of shots

Other video games consist of similar problems that need to be solved.

Participants in this project can work in teams or alone and can either work on some or all of the different sub-problems involved in successfully playing a video game. Participants can propose an interesting video game themselves.

This project builds on existing work on this topic and participants will work closely with one of the participating academics.