Скачать или смотреть AI Takes On ICE Maze With REINFORCEMENT Learning

Watch as AI beats ICE Maze using reinforcement learning in Gymnasium. See how AI navigates the frozen lake and learns to solve the maze with cross entropy.In this video, I showcase the use of Reinforcement Learning Cross Entropy and Evolutionary Algorithm to beat the challenging ice maze in the Frozen Lake environment from Gymnasium. Watch as we design the neural network and tune the parameters, understand what Reinforcement Learning is and how we debug issues that arise

Note that Cross Entropy is actually quite outdated for anything Reinforcement Learning related, for a maze scenario using a value based method like Q learning would be much much better however i went with Cross Entropy due to how easy it was to implement, this video was just me testing this style of editing really, i'm planning to upload other videos with much more advanced network structures and better games overall in the soon future

And i really enjoyed working with blender, it was a refreshing project among all the coding projects here and there :)

🤖 Github: https://github.com/zeyad-shaban* And I really tortured myself by using shotcut...