Introducción a la Inteligencia Artificial

1. Introducción a la Inteligencia Artificial M.C. Juan Carlos Olivares Rojas jolivares@uvaq.edu.mx juancarlosolivares@hotmail.com @jcolivares http://antares.itmorelia.edu.mx/~jcolivar Enero, 2010

2. Competencia Específica Conoce los fundamentos teóricos que sustentan la Inteligencia Artificial

3. Agenda Nuevas Tendencias en el campo de la Inteligencia Artificial. Jueves 28 de enero. Ensayo de la película 2001. Odisea Especial. Jueves 4 de Febrero. Ontología sobre un tema en específico. Lunes 8 de febrero.

4. Outline 1.1 Basic Concepts 1.2 Applications 1.3 Intelligent Systems and Learning 1.4 Semantic Networks 1.5 Description and Match Method

5. Outline 1.6 Analogy Problems 1.7 Abstraction Recognition 1.8 Knowledge Interpretation

6. Basic Concepts • What’s the diference bewtween Artificial Intelligence (AI) and Human Intelligence? • All the sucessfully AI Systems are based on human knowledge and experience. • Most of the AI Systems can be costructed only when the human intelligence can be expresed in easily form (for instance: if x then y).

7. Basic Concepts • AI Systems extend human experts, but never can’t substituting either “taken” most of human intlligence. • AI Systems don’t have common sense and generallity of human beings. • Human Intelligence are very complex for computing.

8. Basic Concepts If a problem can not be described, then can not be programmed • Human Intelligence have these features: Reasoning. Behavior. Use of Metaphores and Analogies. Concepts Creating and Use.

9. Problem Make a Java Program which calculate if a number give for the user is a Perfect Number or not. What are the steps for solving this problem?

10. Inteligence • Capacity to solution all clasess of problems • Intelligence is very subjective. • “Intelligence Distinguished man of animals” • AI is an interdiciplinary science which involves phylosophy, matemathics, biology, electronics, etc,

11. Turing Test Alan M. Turing defined in 1950 one form to check if a machine is intelligent or not. Turing test consist to set two human and one machine in a dark room. The humans and the machine are not visible between their. One human must act like an Interviewer asking some questions to the other participants.

12. Turing Test Turing Test is passed when the interviewer can not distinguished the answer between the human and the machine. The new AI systems required the perception sense to pass the test.

13. AI Genesys Martin Minsky did cotributions to define brain models in computers. ELIZA of Joseph Weizenbaum and JULIA of Mauldin were the first AI Systems with Intelligent Dialagues. The first AI Systems were development for solving some problems like chess.

14. Génesis de la IA In 1956 John McCarthy and Claude Shanon published “Automata Studies” where defined the Automata Theory. In 1956 John McCarthy defined the AI concept, reason why he is considered the AI Father. The AI history is very old. The greeks were the first to use logic to solve a lot of problems.

15. AI Genesys In 1965 Chomsky defined the Formal Languages Theories. McCulloh and Pits in 1943 define the relations between neurons and simple computational elements. In 1962 Rosenblatt defined the Perceptron and the Neuronal Networks Teories.

16. Maze Problem How a person in a maze can be exit without lost? Are there an optimal solution for the problem?

17. Expert Systems They were the first AI comercial product sucessfully. These Systems let to introduce some information in an specific knowledge area into a computer (knowledge database), they act like a human expert. These Systems simulate human reasoning by applicating especific knowledge and inferences.

18. Natural Language Processing It’s a complex problem. For example (in spanish): “Ideas verdes descoloridas duermen furiosamente”, “Ideas furiosamente verdes descoloridas duermen”.

19. Natural Language Processing “El banco cierra a las 3:00” “Las almejas están listas para comer” “Las almejas están listas para [ser] comidas [por nosotros]”

20. Artificial Vision It’s an application of patter recognition, this area have a lot of application such as: Medical Diagnostic Automatic Signal Processing Automatic Industrial Product Automatic Vigilance Systems OCR (Optical Character Recognition)

21. Robotic This science implies the concepts of perception, motion (spatial reasoning), planning. The main problem autonomous robots are interacting with the human-world, because exists many obstacles unexpected events and dinamic environments.

22. Learning This area studies the way in how computers can obtain new knowledge to solve a problem. In this sense, learning means to make a computer which is able to benefit for the experience obtained.

23. Games AI is applied in games to give more realism and complexity. Also AI gives the “Physics”. The n-queens problem consist in putting n chess queens on an n×n chessboard such that none of them is able to capture any other using the standard chess queen's moves. Activitie: Obtain a Solution in a sheet of paper for a 6x6 chessboard. First 100, Second 80, Third 60 pts.

24. Genetic Algorithms It’s a computational technique inspired in biological models which are used to realize eficient search in spatial solution highly huge and complex. Genetic Algorithms are adaptative methods which can used to implement searches and optimization problems. This has given the creation of emergence areas such as evolutionary computation and swarm computing algorithms that rely on events of nature.

25. Cellular Automata It’s a discrete model studied in computing, mathematics, biology and microstructure modeling. It consists of a regular grid of cells, each in one of a finite number of states. The grid can be in any finite number of dimensions. Time is also discrete, and the state of a cell at time t is a function of the states of a finite number of cells (neighborhood) at time t − 1.

26. Cellular Automata These neighbors are a selection of cells relative to the specified cell, and do not change (though the cell itself may be in its neighborhood, it is not usually considered a neighbor). Every cell has the same rule for updating, based on the values in this neighbourhood. Each time the rules are applied to the whole grid a new generation is created.

27. Cellular Automata

28. The Game of Life The Game of Life, also known simply as Life, is a cellular automaton devised by the British mathematician John Horton Conway in 1970. It is the best-known example of a cellular automaton. The "game" is actually a zero-player game, meaning that its evolution is determined by its initial state, needing no input from human players. One interacts with the Game of Life by creating an initial configuration and observing how it evolves.

29. The Game of Life The universe of the Game of Life is an infinite two-dimensional orthogonal grid of square cells, each of which is in one of two possible states, live or dead. Every cell interacts with its eight neighbours, which are the cells that are directly horizontally, vertically, or diagonally adjacent. At each step in time, the following transitions occur:

30. The Game of Life Any live cell with fewer than two live neighbours dies, as if by needs caused by underpopulation. Any live cell with more than three live neighbours dies, as if by overcrowding. Any live cell with two or three live neighbours lives, unchanged, to the next generation. Any dead cell with exactly three live neighbours becomes a live cell.

31. The Game of Life Play the game at: www.bitstorm.org/gameoflife/

32. Virtual Reality It’s one of the most recent applications of AI. It’s consist in the construction of programs which achive to fool the human senses, make it belive that we are floating, running or flying in an airplane. This application has been used in a fligth simulator for pilots, astronauts and drivers.

33. Intelligent Systems and Learning Most of the actual system say that they are intelligents (“smart”). If an application can take autonomous decisions in a real time in independet form, it’s considered intelligent. The main feature of this systems are the “adaptability” like saving energy.

34. Intelligent Systems and Learning The most important feature of an Intelligent System are the way to representing the knowledge, the way in which the information is retrived and the way in which adquire new knowledge (learning). The representation ways (“explicitation”) of knowledge are diverse and it influences in the retrival informtion and learning ways.

35. Intelligent Systems and Learning Always that a model is developed it has two represetation: logical and physical. This representations need “mapping” to working together. When we have a real life problem, this have to mapping in a computer schema for working in a computational system.

36. Intelligent Syst. and Knowledge Tacking back to the Maze Problem ¿How can be represent this model and the knowledge? It can be represented with a matrix, graph, finite state machine, etc. Also it must rules for play this game. If we don’t have the two representations we can not understand and learn the game.

37. Intelligent Systems and Knowledge In general, knowledge s define by laws and particular languages. Languages define rules. The same knowledge is structured in diferents represtentation such as database, semantic networks, frames, conceptual maps, etc., but after all it must have the same meaning (semantics).

38. Maze Generator We must try to don´t generate a loop

39. Semantic Networks They are other simple form to explicity knowledge, They are conformed by graphs which coding knowledge in a taxonomic form. Nodes represent categories and Edges represents the relations between this categories. There are two types of special relatinoships: Is-A y la Have-A.

40. Semantic Networks We can access throught of each concepts to infer knowledge. The scripts are other way to represent knowledge. They are composed by components called slots, these are a set of elements concept-values. Scripts are more easily to ntroduce than mind maps.

41. Semantic Networks

42. Script Script Example: Printers Subset_of: Office_Machine Superset_of: {Laser_Printer, Inject_Printer} Feed_Source: Door_Socket Author: Juan_Perez Date: 07_January_2008

43. Onthologies Other way to represent knowledge with a lot of use recently is Onthology, It’s consist of relations between distinct concepts like definitions. Onthologies can be represented throught languages such as XML. Knowledge representation has a great importance this is the reason because actually we talk about Knowledge Engineering.

44. Semantic Networks Onthologies act like a dictionary. Some elements like agents used this information to represent and retrieve knowledge. Frames are structure used to represent values, restricctions, process, relation, etc. Frames represent with tuples one propertie of an object. Object-Oriented Programming was originated by Frames.

45. Concept Mind

46. Onthology

47. Activity Use a Protege Onthology Editor for Make a Onthology about Mexican Soccer.

48. The Description&Match Method It’s used for AI problem solving and It’s one the most basic method. The first step consists to identified all features of an object. Later, It realice a seach in a well-define set of objects. It needs two very import methods: the extractor and evaluator of knowledge.

49. The Description and Match Method • When the match process is doing, one posibility is the object don’t be the same pattern in the knowledge Database. This is the reason because We need a Similarity Function. • For Example (In Spanish): AMOR Love a person or thing for over all things Word composeb by 4 characters: ‘A’, ‘M’, ‘O’ and ‘R’ yuxtapuestos

50. The Description&Match Method AMOR = AMOR Exact Macth AMOR = ROMA 0% similarity but contains for characters Amor = AMOR 25% similarity, contains all character but in uppercase Amor = Cariño 0% similarity but the same meaning Amor = Amar 75% it’s a consequence