Many people have tried to defend pure DRL w Nature Machine Intelligence article that actually is a hybrid model; the below youtube pointer IS to a pure convnet - but read fine print: “network works decently well for any position less than 6 moves away from solved”
#symbolphobiahttps://twitter.com/AlexRoseGames/status/1186571935611850752 …
-
I think you are overstating this. Some form of basic search is probably innate, but we can learn to search better via experience. A*, AO*, MCTS, Thompson sampling are all very sophisticated search methods; I doubt they are innate.
MCTS is only innate in a few places, like AlphaGo and that hybrid Nature Machine Intelligence Rubik’s solver that DRL people keep sending me without reading. Certainly not innate in humans.
it’s hilarious the way arch-anti-nativists keep building MCTS when they work on games 
