2.19 수요일

 

-Beam search

 

-Stochastic hill-climbing

 

-Simulated annealing

 

-Hyperparameters

 

-Grid search

 

 챗지피티는 브루트포스로 학습시켜서 리소스 낭비가 심할거래

 

 

-Anytime search. -> 다른 오프라인서치(A* 등) 들과 다름 

          -중간에 멈춰도 결과가 있음

 

 

-Adversarial Search

-Minimax search (e.g. tiktak-toe)

 

 

-Alpha-Beta

            Alpha-beta pruning

• As you are expanding the search tree DFS-style, keep two bounds:
• These represent the range of values the node can take on in order to reach the root
• As you search, tighten the bounds

...

 

 

-Planning Algorithm