Neural Network Pathfinder

A little program that showcases an Artificial Neural Network (ANN) that learns how to pathfind through randomly generated mazes using A* as an error function. Input neurons are fed certain data about the cell, such as its position, distance from the goal, amount of times it's been walked over, and more. The source code is available for download if you wish to play around with different inputs.

A* pathfinding is on the left side, with the ANN being on the right hand side, for easy comparison between the two methods.

Also check out my game, VoxCom!