Se ha denunciado esta presentación.
Se está descargando tu SlideShare. ×

Artificial Intelligence AI Topics History and Overview

Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Anuncio
Próximo SlideShare
Artificial intelligence
Artificial intelligence
Cargando en…3
×

Eche un vistazo a continuación

1 de 26 Anuncio

Más Contenido Relacionado

Presentaciones para usted (20)

A los espectadores también les gustó (20)

Anuncio

Más de butest (20)

Anuncio

Artificial Intelligence AI Topics History and Overview

  1. 1. Artificial Intelligence
  2. 2. AI Topics <ul><li>History and Overview </li></ul><ul><li>Machine Learning </li></ul><ul><li>Games and AI </li></ul><ul><li>The Turing test </li></ul><ul><li>Computer Vision </li></ul>
  3. 3. AI Pioneers <ul><li>Alan M. Turing </li></ul><ul><ul><li>“ Computing Machinery and Intelligence” </li></ul></ul><ul><li>Marvin Minksy </li></ul><ul><ul><li>Constructed the first neural net machine </li></ul></ul><ul><li>Herbert Simon, Allen Newell, J.C. Shaw </li></ul><ul><ul><li>Developed the first AI computer program </li></ul></ul>
  4. 4. AI Terms <ul><li>Artificial Intelligence: The capability of a machine to imitate intelligent human behavior </li></ul><ul><li>Artificial Neural Network: A network of neurons with connections of varying strength </li></ul><ul><li>Fuzzy Logic: A superset of Boolean logic which includes truth values between true and false </li></ul><ul><li>Knowledge Base: A collection of knowledge expressed using some formal knowledge representation language </li></ul><ul><li>AI-complete: Describes a problem which presupposes a solution to the “strong AI problem” </li></ul>
  5. 5. Famous AI Programs <ul><li>ELIZA (Joseph Weizenbaum) </li></ul><ul><ul><li>Psychologist </li></ul></ul><ul><li>Deep Blue (IBM) </li></ul><ul><ul><li>Chess program </li></ul></ul><ul><li>Cyc (MCC and Cycorp) </li></ul><ul><ul><li>Multi-contextual knowledge base and inference engine </li></ul></ul><ul><li>HAL (Arthur C. Clarke) </li></ul><ul><ul><li>Space explorer </li></ul></ul>
  6. 6. Machine Learning
  7. 7. What Is Machine Learning? <ul><li>Enabling machines to process data in such a way that it can be to make future decisions </li></ul><ul><li>ML been studied for many years </li></ul><ul><li>ML has many applications in a variety of fields </li></ul>
  8. 8. Methods of Learning <ul><li>Genetic algorithm </li></ul><ul><li>Inductive logic </li></ul><ul><li>Computational learning </li></ul>
  9. 9. Dimensions of Study <ul><li>Representation of experience </li></ul><ul><ul><li>Most learning is based on experience </li></ul></ul><ul><ul><li>Storage values </li></ul></ul><ul><ul><ul><li>Attribute values (length) </li></ul></ul></ul><ul><ul><ul><li>Binary values (yes/no) </li></ul></ul></ul><ul><ul><li>Relations (Difficult) </li></ul></ul><ul><li>Representation of acquired knowledge </li></ul><ul><ul><li>Generalizations </li></ul></ul><ul><ul><li>Logical/discrete vs. numeric/continuous </li></ul></ul>
  10. 10. Dimensions of Study <ul><li>Supervised and unsupervised learning </li></ul><ul><ul><li>Supervised </li></ul></ul><ul><ul><ul><li>Feedback given immediately after an action is taken </li></ul></ul></ul><ul><ul><ul><li>Easy to give examples of correct vs. incorrect behavior </li></ul></ul></ul><ul><ul><li>Unsupervised </li></ul></ul><ul><ul><ul><li>Machine learns on its own with no conditioning </li></ul></ul></ul><ul><li>Inductive learning vs. analytic learning </li></ul><ul><ul><li>Inductive – take all data, make generalizations </li></ul></ul><ul><ul><li>Analytic – offer explanations for new data based on previous data, then simplify </li></ul></ul>
  11. 11. Dimensions of Study <ul><li>Incremental vs. Non-Incremental Learning </li></ul><ul><ul><li>Incremental </li></ul></ul><ul><ul><ul><li>Examine results one-by one </li></ul></ul></ul><ul><ul><ul><li>Less information retained, but faster </li></ul></ul></ul><ul><ul><li>Non-Incremental </li></ul></ul><ul><ul><ul><li>Examine all results at once </li></ul></ul></ul><ul><ul><ul><li>More information retained, but slower </li></ul></ul></ul>
  12. 12. Tasks For Machines <ul><li>Pattern recognition </li></ul><ul><li>Grouping/classification </li></ul><ul><ul><li>Create general descriptions for classes of instances </li></ul></ul><ul><li>Strategizing </li></ul><ul><li>Generating heuristics </li></ul><ul><li>Problem solving </li></ul>
  13. 13. Problem Solving <ul><li>Take a similar problem with a known solution and try to find the answer (analogies) </li></ul><ul><li>Simplify the problem and find a solution that can be used to solve the main problem </li></ul><ul><li>Thresholds </li></ul><ul><li>Decision trees </li></ul><ul><li>Macro-operators (AND, OR) </li></ul>
  14. 14. Issues in Machine Learning <ul><li>Computational complexity </li></ul><ul><li>Ethics </li></ul><ul><li>Correctness </li></ul><ul><ul><li>Would the exact desired learning be constructed? </li></ul></ul><ul><ul><li>What if there is an error in learning? </li></ul></ul>
  15. 15. Games AI <ul><li>Min-Max Trees </li></ul><ul><ul><li>Builds a level of maximizing moves followed by a level of minimizing moves </li></ul></ul><ul><ul><li>Uses evaluate functions to analyze situation </li></ul></ul><ul><li>Alpha Beta Trees </li></ul><ul><ul><li>Like Min-Max Trees </li></ul></ul><ul><ul><li>Discards paths it knows to be useless </li></ul></ul>
  16. 16. Chess Algorithms <ul><li>Most use Alpha-Beta trees to make moves </li></ul><ul><li>Trees helped by additional knowledge </li></ul><ul><ul><li>Transposition Tables </li></ul></ul><ul><ul><li>Endgame Database </li></ul></ul><ul><ul><li>Human Literature </li></ul></ul><ul><li>Deep Blue </li></ul><ul><ul><li>First championship caliber chess player </li></ul></ul>
  17. 17. Other Games <ul><li>Othello – Logistello </li></ul><ul><ul><li>Deep search algorithm </li></ul></ul><ul><ul><li>Can solve most endgames </li></ul></ul><ul><ul><li>Large opening book </li></ul></ul><ul><li>Checkers – Chinook </li></ul><ul><ul><li>Extremely deep search depth </li></ul></ul><ul><ul><li>8 piece endgame database </li></ul></ul>
  18. 18. The Turing Test <ul><li>Motivated to identify intelligence in a computer program. </li></ul><ul><li>Proposed in 1950 by Alan Turing. </li></ul><ul><li>Original Proposal: </li></ul><ul><ul><li>Given a person X, a computer Y, and an interrogator C, C isolated from X and Y. </li></ul></ul><ul><ul><li>C must determine who is the person </li></ul></ul><ul><ul><li>X is intelligent if it can fool C. </li></ul></ul>
  19. 19. Problems with the Turing Test <ul><li>Intelligence may be considered as a continuum. The Turing test only identifies one (very strong) type of intelligence, and thus offers no means to measure . </li></ul><ul><li>Does fooling C really imply intelligence? </li></ul>
  20. 20. Our Proposal <ul><li>Motivated to allow: </li></ul><ul><ul><li>a measure of intelligence. </li></ul></ul><ul><ul><li>more rigid definitions. </li></ul></ul><ul><ul><li>more flexible admission of programs. </li></ul></ul>
  21. 21. Our Proposal <ul><li>Define D as the set of all problems. </li></ul><ul><ul><li>This may be restricted for practical considerations. </li></ul></ul><ul><li>P(D) is therefore the partially ordered set (under inclusion) of all subsets of problems. </li></ul>
  22. 22. Our Proposal <ul><li>Let R be the set of all responses </li></ul><ul><li>P(R) is therefore the partially ordered set of subsets of R. </li></ul><ul><li>Define the Turing Test T as a function between P(D) and P(R). </li></ul><ul><li>Those programs which mimic T on some subset X (pre-image) of P(D) are said to pass T restricted to X . </li></ul>
  23. 23. Our Proposal <ul><li>As P(D) is partially ordered, and by the way D was defined, there are several maximal elements M i in P(D). </li></ul><ul><li>A program that is said to pass T restricted to an M i is said to be an expert in M i . </li></ul><ul><li>In specific applications, one may identify an expert program as intelligent. </li></ul>
  24. 24. Examples <ul><li>Consider the set of Arithmetic Problems </li></ul><ul><li>If a program can solve these problems, it is said to pass T restricted to Arithmetic Problems . </li></ul><ul><li>In practice, one would need to restrict this set. </li></ul>
  25. 25. Examples <ul><li>The set of all Math Problems is a maximal element. </li></ul><ul><li>If a program can solve these problems, it is said to be an Expert in Math Problems . </li></ul>
  26. 26. Sources <ul><li>Encyclopedia of Artificial Intelligence 2 nd ed. Ed. Stuart C. Shapiro. John Wiley & Sons, Inc. New York City, NY, 1992. </li></ul><ul><li>R. Miikkulainen and D. Moriarity. Discovering Complex Othello Strategies Through Evolutionary Neural Networks. University of Texas, USA, 1995. </li></ul><ul><li>B. Moreland. Basic Search Techniques. http://www.seanet.com/~brucemo/topics/topics.htm. USA, 2001. </li></ul><ul><li>J. Schaffer. The Games Computers (And People) Play. University of Alberta, Canada, 2000 </li></ul>

×