Recreating DeepMind's AlphaZero - AI Plays Connect 4 - Part 3: Monte Carlo Tree Search

In this series we re-create DeepMind's AlphaZero to create an Artificial Intelligence Reinforcement Learning algorithm that plays Connect 4 using Python.
View the rest of the series: www.youtube.co....
Full Code: github.com/The...
AI, Deep Learning, Reinforcement Learning, DeepMind, Convolutional Neural Networks, Monte Carlo Tree Search, AlphaGo, AlphaZero, Connect4, ConnectX
bits of code modified from source:
github.com/Jos...

[@daniloilguisonis3462]

I am very happy that u are posting again! I am Data Science student interested in its applications to finance, your videos were very helpful. Hope u keep beign active!

[@daniloilguisonis3462]

I have a question, did u by any chance did Reinforcement Learning on crypto price? If u did, what its your take on that?

[@theprincipalcomponent]

Glad you’re enjoying! I’ve poked around with it, but haven’t spent a substantial amount of time. If I had to guess I think it could work well

[@crayc3]

Hi, great Video! Quick Question: I saw in other videos on mcts that each node always holds the value with respect to the player who is to play in this particular state. So from the perspective of the player whose turn it currently is in that state. Which would mean backing up the value alternating between +1 and -1. And also they would inverse the board after each step so that it's always "player 1" to play. I guess this is to make the network able to play either as player 1 or 2? You see I am confused and i wanted to ask if you have an explanation why one would inverse the state or not? Thanks ;D