# AIMA - Adversarial Search and Games

In which we explore environments where other agents are plotting are against us.

Games offer good places to explore search algorithms in practice since they model the uncertainties and difficulties of real life: the branching factor is high, there’s elements of chance, you might have partial observability, and maybe because a lot of the time you’re working against people.

## Flashcards

What is the name of algorithm used to make optimal decision in perfect information games?

Minimax.

What is the name of the procedure used to prune game search trees?

Alpha-beta prunining.

How do game search algorithms deal with not being able to search the entire game tree?

They use an evaluation function to evaluate states at the maximum depth.

In summary, how does Monte Carlo tree search evaluate states?

Simulating gameplay and averaging the outcomes.

What is the name of the algorithm used when minimax is applied to games of chance?

Expectiminimax.

In game tree search, what’s one way that the evaluation function measures states?

It looks at the different features.

What is the algorithm used to balance exploring and exploiting in Monte Carlo tree search?

Upper confidence bounds applied to trees (UCT)