3 ms·
Randomly sample game states, giving more weight to those that affect balance more strongly and/or occur more frequently.
by eternauta3k 9y ago
Randomly sample game states, giving more weight to those that affect balance more strongly and/or occur more frequently.
- gwern 9y agoRandom sampling wouldn't work because it is the essence of cheese/game-breaking to go into rare or weird edge-cases and parts of state-space. Imagine a broken fighting game where you could win by simply executing 'upper block' the entire game; random sampling will be exponentially unlikely to sample a game with all upper-blocks. But if you used a mechanism like self-play reinforcement learning to explore the game state-space, your agents might autonomously discover game-breaking strategies on their own.