Monte Carlo Methods
Computational algorithms that use repeated random sampling to estimate numerical results, widely used in reinforcement learning and probabilistic inference.
In RL
Monte Carlo methods estimate the value of states by averaging returns from multiple complete episodes. Unlike TD learning, they don't bootstrap (estimate from other estimates) and can work with any environment.
Monte Carlo Tree Search (MCTS)
The algorithm behind AlphaGo. MCTS builds a search tree by repeatedly: selecting promising branches, expanding new nodes, simulating random playouts, and backpropagating results. Combined with neural network evaluation, it achieved superhuman Go play.