We will be creating a machine learning agent that will learn to fight. Our learning agent will use the pixels that our agent sees as input. Then we will put it into a computer vision leaning algorithm. The output will be the mouse clicks/moves it makes.
We will be using a computer vision object recognition algorithm to detect the environment. We will also use a reinforcement learning algorithm to teach our agent how fight.
We will make the agent fight past iterations of itself taking time into account and confirm that it is improving. Not sure if we are going to use a nuero dynamic programming based algorithm or not. Our moonshot will be if the agent can beat our team.
Our scheduled appointment is on Wednesday, May 3rd at 3:00pm