Deep Reinforcement Learning [On going]

In recent years, research on game playing AI has benefited from the advancement of hardware and game research platforms. Variety of game environment that pose different challenge has been used extensively for Deep Reinforcement Learning-based techniques. For example, Arcade Learning Environment (ALE) and Retro Learning Environment (RLE) are popular arcade game environments characterised by visual input and control. AI agents must be able to represent and learn from the visual environment to excel. This project aims to achieve higher accuracy in multitask game playing. Additionally, a website is set up to track current progress on different environments of game playing. [More report soon]