3 ms·
I think current LLM architectures are limiting the strategies it learned. MCTS requires recursion but GPT is always executing a fixed number of steps. Allowing
by LeanderK 4y ago
I think current LLM architectures are limiting the strategies it learned. MCTS requires recursion but GPT is always executing a fixed number of steps. Allowing language models more flexibility by a variable number of steps, for example through recursion of the model, would eliminate this hard bound but they are harder to design and train. We have just been able to train GPT-sized models.
- sebzim4500 4y agoI'm sure the MuZero chess policy network would reach much higher than 1400, and that has no notion of recursion either. (And also wasn't taught the rules explicitly)
- LeanderK 4y agoYes, I wouldn't say it's impossible, but that's just a hard limit from the architecture. MCTS would have to terminate after a few iterations. And the number of steps is not that big, just massively parallel.