AI learns to play ATARI and DOOM - deep reinforcement learning experiments

preview_player
Показать описание
some of my recent deep reinforcement learning experiments, programmed in my custom CNN framework RYSY; instead of using dogmatic 8x8, 4x4 DQN kernels architecture, I choose modern VGG-net like structure, using only 3x3 kernels, deeper architecture and ELU activation function
Рекомендации по теме
Комментарии
Автор

You didn't just pust a picture of your slavic self at the end of the video

jonathangilliam