AIMA
AIMA

My sparse literature notes of the Artificial Intelligence: A Modern Approach book by Peter Norvig and Stuart Russel.

1. Introduction

This try chapters answer the question of what exactly is AI

2. Intelligent agents

Agents and task environments' characteristics.

Q: Components of an "agent problem"? A: PEAS

  • Performance measures
  • Environment
  • Actuators
  • Sensors

3. Solving problems by searching

The classical searching algorithms that you would see in typical intro algorithms type of books.

4. Search in Complex Environments

What are the Local search methods mentioned in this chapter?

  • Hill climbing heuristic
  • Simulated annealing stochastic heuristic

What search algorithm could be use in nondeterministic environments?:: AND-OR search.

What is the plan that AND-OR search generate?:: contigent plans.

When the environment is partially observable, what state is kept tracked of?:: belief state.

5. Constraint Statisfaction Problem

Getting to know a new type and closer to being useful in real world class of problems.

6. Adversarial search

How would you do search in competitive environments / games.

What are the formulation components that a game is defined by? ?

  • initial state
  • legal action (in each state)
  • result of each action
  • terminal test
  • utility function (to decide who won and what the final score is)

What do you call the characteristic the environment the state is always transparent?:: perfect information.

What is the algorithm to you calculate the best move for you, the worst move for the opponents, turn by turn?:: Minmax algorithm

--SR:!2022-01-06,2,246-->

What is the more optimized version of Minmax algorithm?:: Alpha-beta algorithm.

What optimization does Alpha-beta algorithm do over Minmax algorithm?:: It eliminates subtrees that are provably irrelevant.

In situations where it is still expensive even with Alpha-beta algorithm, what can you do?:: cut the search off at some point, and estimate the utility of the states with an heuristic evaluation function.

"Sometimes you can trade space for time", know what this allures to?:: Precomputed tables for opening and endgame.

7. Logical Agents