So, how difficult do you think it would be for me to take the AlphaZero algorithm and apply it to, say choosing bad Tetris pieces https://arxiv.org/pdf/1712.01815.pdf …
Show this thread
Seriously, does the paper have enough information to work from, or will I also need a Ph.D. in machine learning? I can't tell.
-
-
If you go back in time the length of 1 PhD, you're basically looking at the very beginning of this field (not literally but like 99% of the work has been done since then).
-
Is that a yes?
- 1 more reply
New conversation -
-
-
The results are cool but the papers' explanations are lousy (eg confusion about key being ExIt). The independent invention by Anthony et al is clearer, or at least helpful as an explanation from a different perspective: https://arxiv.org/abs/1705.08439 https://davidbarber.github.io/blog/2017/11/07/Learning-From-Scratch-by-Thinking-Fast-and-Slow-with-Deep-Learning-and-Tree-Search/ …
-
Applying it to Tetris from the perspective of the block-picker 'player' sounds sensible to me, FWIW. I'd suggest starting with pure MCTS and only trying to implement ExIt if that is unsatisfying.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.