This is just a little experiment to see if i could make an AI in scratch and it worked out pretty well. Arrow keys to change displayed neural network Space to change to a harder course (one or two cars consistently finishing recomended)
3 cars will race to the finish line. They start by making totally random decisions but the probability of them having different reactions based on their surroundings changes. --> If they beat their record distance or time to the finish line the chance of making the same decisions gets strengthened (reward) --> If they don't the decisions get weakened (punishment) The neural network displays the things they see in the first column and the decisions they take based on what they see in the second column.