it-ebooks - Artificial Inteligence - Leonardo Araujo dos Santos

1. 1.1
2. 1.2
3. 1.3
   1. 1.3.1
   2. 1.3.2
      1. 1.3.2.1
      2. 1.3.2.2
      3. 1.3.2.3
      4. 1.3.2.4
      5. 1.3.2.5
      6. 1.3.2.6
      7. 1.3.2.7
      8. 1.3.2.8
   3. 1.3.3
      1. 1.3.3.1
      2. 1.3.3.2
      3. 1.3.3.3
      4. 1.3.3.4
      5. 1.3.3.5
      6. 1.3.3.6
         1. 1.3.3.6.1
      7. 1.3.3.7
      8. 1.3.3.8
      9. 1.3.3.9
      10. 1.3.3.10
      11. 1.3.3.11
      12. 1.3.3.12
      13. 1.3.3.13
      14. 1.3.3.14
      15. 1.3.3.15
   4. 1.3.4
      1. 1.3.4.1
      2. 1.3.4.2
   5. 1.3.5
   6. 1.3.6
4. 1.4
   1. 1.4.1
   2. 1.4.2
   3. 1.4.3
   4. 1.4.4
      1. 1.4.4.1
      2. 1.4.4.2
      3. 1.4.4.3
5. 1.5
   1. 1.5.1
   2. 1.5.2
      1. 1.5.2.1
      2. 1.5.2.2
      3. 1.5.2.3

On this chapter we will learn the basics topics to better understand the book. The following topics will be presented:

• Lua, Torch
• Tensorflow
• Python, numpy
• Matlab

In computer science, an ideal "intelligent" machine is a flexible rational agent that perceives its environment and takes actions that maximize its chance of success at some goal.

Here we will try to explain what means acting rational:

• Maximally achieve pre-defined goals. (Expected utility)
• Goals expressed in terms of utility (non dimensional scalar value that represent "happiness")
• Be rational means maximize your expected utility

Have a rational (for us intelligent) decision, is only related to the quality of the decision(utility), not the process that lead to that decision, for instance:

• You gave the right answer because you brute-force all possible outcomes
• You gave the right answer because you used a fancy state of the art method

Basically you don't care about the method just the decision.

As the AI field evolve what was considered intelligent, is not considered anymore after someone discover how to do it.

A system is rational if he does the right thing with the information available.

Basically we don't understand how our brain work, even with the current advances on neuro-science, and the advances on computation power. We don't know how the brain works and take decisions.
The only cool thing that we know about the brain, is that to do good decisions, we need memory and simulation.

1. Reasoning
2. Knowledge
3. Planning: Make prediction about their actions, and choose the one that
4. Learning
5. Perception
6. Natural Language Processing (NLP)

It's a system (ie: software program) which observes the world through sensors and acts upon an environment using actuators. It directs it's activity towards achieving goals. Intelligent agents may also learn or use knowledge to achieve their goals.

There are a lot of agents types, but here we're going to separate them in 2 classes

• Reflex Agents: Don't care about future effects of it's actions (Just use a if-table)
• Planning Agents: Simulate actions consequences before committing to them, using a model of the world.

Both reflex and planning agents can be rational, again we just care about the result action, if the action maximize it's expected utility, then it is rational.

we need to identify which type of agent is needed to have a rational behavior.

A reflex or planning agent can be sub-optimal, but normally planning is a good idea to follow.

On this book we're going to see a lot of tools used to address planning. For instance searching is a kind of tool for planning.