Teaching AI to Think Smarter, Not Harder
Channel:
Subscribers:
175,000
Published on ● Video Link: https://www.youtube.com/watch?v=39v3A74t-8A
42FULL VIDEO: • How AI Learned to Reason
These random playthroughs known as rollouts offered a way to estimate a move's strength without having to analyze every possible future Known as Monte Carlo Tree Search However the breakthrough would come from combining Monte Carlo tree search with neural networks we saw earlier using both position and move intuition to guide the rollouts towards the most promising directions