AI General Thread

Started by Legend, Dec 05, 2022, 04:35 AM

0 Members and 5 Guests are viewing this topic.

Legend


I vibe coded an abstract strategy game. Spent two days with my extra pc training an AI to play it, same architecture as AlphaZero. But it sucked and was barely intelligent.

I also tried a more traditional chess like AI but this game just has way too many states. I could only search a single turn forward in a reasonable time, since unlike chess you can't prune all the obviously wrong moves.



Not sure what else I could try. This is a toy problem just to get some practice with these types of AI systems.