31st December 2021*
Applicants should contact the primary supervisor, and submit their Expression of Interest (EOI) and Application as soon as possible.
*unless filled earlier
In this project, HDR students implement and develop of state-of-the-art machine learning and deep learning models, especially in deep reinforcement learning algorithms to easily train intelligent agents for various games. The research goal is to speed up the learning process of multiple agents and allow each agent receives higher rewards in a game scenario. These trained agents can be presented in the demo workshop and can be used for multiple purposes, including testing of game builds and controlling behaviour.
In this project, we used the OpenAI Gym and Unity platform, which have been developed for creating and interacting with simulation environments. Specifically, the Unity ML Agents Toolkit is an open-source Unity plugin that enables games and simulations to serve as environments for training intelligent agents. This project will use this toolkit to develop dynamic multi-agent interaction, and agents can be trained using reinforcement learning, imitation learning, neuro-evolution, or other machine learning methods through a simple-to-use Python API.
Additionally, this project is mutually beneficial for both students and AI researchers as it provides a central platform where advances in AI can be evaluated on rich environments and then made accessible to the industry and research developer communities.
Applicants who require more information or are interested in this specific project should first contact the listed Supervisor. Information and guidance on the application process can be found on the Apply Now website.
Information about scholarships is available on the Scholarships webpage.
Please contact, Zehong Cao for further information.