The project implements machine learning to teach our main agent, the cat, to catch certain objects (white balls), in the most efficient time using a dynamic reward system generated in an environment supported through artificial intelligence. Developed on Unity with C Sharp and Python for the initiation.
Our objective is to teach the cat to catch the correct objects using a punishment and reward system. The environment resets every time the agent meets specific conditions to improve and learn in real time.
Our algorithm can be scalable and applied for different sectors and fields such as artificial intelligence in game design or smart systems such as robotics.
The ball objects have been programmed to retreat from the cat while the cat has been tasked to catch the balls, except for the red balls which punishes the cat with a minus(-1) count which allows it to learn, you can think of it as learning from your mistakes.
Please comment, reply, share and ask for any details. Will add github link very soon and modify accordingly.