I have absolutely no clue if this is going to reach the correct people, but for anyone who is experienced with machine learning, particuarly with reinforcement learning, which is the better algorithm for playing yugioh? PPO or DQN? I've been looking into DQN, but mapping the available moves seems to be quite a hard task, and I have heard that PPO seems to be better suited for a game like yugioh.
If it helps, the model will only be trained on a 40 card deck, against another fixed 40 card deck.
If anyone is particuarly knowledgeable about this area, I would be happy to get in contact :)
yifan
I have absolutely no clue if this is going to reach the correct people, but for anyone who is experienced with machine learning, particuarly with reinforcement learning, which is the better algorithm for playing yugioh? PPO or DQN? I've been looking into DQN, but mapping the available moves seems to be quite a hard task, and I have heard that PPO seems to be better suited for a game like yugioh.
If it helps, the model will only be trained on a 40 card deck, against another fixed 40 card deck.
If anyone is particuarly knowledgeable about this area, I would be happy to get in contact :)
2 months ago | [YT] | 41