Week 11

Week 11

1. Inteligencia artificial

   • AI is that in addition to letting us build really interesting and intelligent devices, it tells us something about how humans are intelligent.
     • Baxter
   • Beyond Board Games
     • Reinforcement learning
   • More complex games
     • We need a good evaluation function to specify which boards are better, even if nobody has yet won or lost.
   • Al Search
     • Search tree
       • Search problems
         • This entire class of problems, where we’ll build a search tree and compare all the paths to see how well they fulfill some criterion, we’ll call search problems.
       • ◦This entire class of problems, where we’ll build a search tree and compare all the paths to see how well they fulfill some criterion, we’ll call search problems.
       • Is a tree data structure used for locating specific values
     • Depth first search, where we go all the way down one path and then go back up to the top and start down another path
       • Breadth first search, where we look at all the top-level nodes, then all the nodes one level down, then all the nodes two levels down, and so on
   • Adversarial Search
     • This is how game-playing systems work
     • How to make it....
       • minimax strategy: choosing the move which leaves our opponent with the worst set of possibilities.