Definition
Plain language
A search method that builds up a tree of possible moves by simulating many random rollouts.
As stated in the literature
Monte Carlo Tree Search, a heuristic search algorithm that incrementally builds a tree of actions via simulated rollouts and propagates their results.
Why it matters: It's a flexible way to combine learned value estimates with simulation, and it underlies many of the most successful game-playing and planning systems.
For example, an MCTS-based Go program explores promising lines by playing thousands of random games from each candidate move and seeing which leads to the most wins.
Heard on the show
“It runs MCTS — Monte Carlo Tree Search, the same family of algorithms that powered AlphaGo — over thousands of possible workflow variants, scoring each on a small validation set, keeping the best one.”Episode 013 — Why Search Keeps Rediscovering the Same Workflow, and What That Means