Artificial Intelligence

1. Artificial Intelligence K. SREENIVAS Modified presentation of GholamReza GhassemSani, Sharif University of Technology

2. Main Reference: Elaine Rich Kevin Knight Artificial Intelligence McGraw-Hill, 1991

3. Definition: Intelligence: “ability to learn, understand and think” (Oxford dictionary). Elaine Rich and Kevin Knight :Study of how to make computers do things at which, at the moment, people are better Herbert Simon: We call programs intelligent if they exhibit behaviors that would be regarded intelligent if they were exhibited by human beings.

4. Definition: • Nilsson, Nils J.:The goal of work in artificial intelligence is to build machines that perform tasks normally requiring human intelligence. • Claudson Bornstein: AI is the science of common sense. • Douglas Baker: AI is the attempt to make computers do what people think computers cannot do

5. Definition: • AI has something to do with • problem solving • It involves Knowledge, Reasoning and Learning and much much more. • Examples: Speech recognition, Smell, Face, Object, Intuition, Inferencing, Learning new skills, Decision making, Abstract thinking

6. What About Things that People Do Easily? • Common sense • Newell, Shaw etc. all built a General Problem Solver that was applied to several common sense tasks. • Moving Around : Robotics • Language : • The problem of understanding spoken language. • The problem of understanding a written language is termed as natural language understanding.

7. Domains of AI • Mundane tasks • Perception -Vision - Speech • Natural language understanding • Formal tasks • Games, • Mathematics • ‘Logic Theorist ‘ was an early attempt to mathematical theorems. • Samuel wrote a checker-playing program that not only played games with opponents but also used the experience to improve its performance. • Chess: Deep Blue Vs Garry Kasporov 1996 • Expert Problem Solving • Medical Diagnosis • Engineering

8. The Origins of AI Birth of AI occurred when Marvin Minsky & John McCarthy organized the Dartmouth Conference in 1956 • brought together researchers interested in "intelligent machines" • for next 20 years, virtually all advances in AI were by attendees • Minsky (MIT), McCarthy (MIT/Stanford), Newell & Simon (Carnegie),… Marvin Minsky John McCarthy

9. 4 questions • What are our underlying assumptions about intelligence? • What kinds of techniques will be useful for solving AI problems? • At what level of detail are we trying to model human intelligence? • How will we know when we have succeeded in building an intelligent program?

10. What is an AI Technique? • Intelligence requires knowledge (less desirable properties) • voluminous • hard to characterize accurately • constantly changing • differ from data by being organized in a way that corresponds to the ways it will be used

11. Knowledge Representation • AI technique is a method that exploits knowledge that should be represented in such a way that: • Knowledge captures generalization • It can be understood by people who must provide it • It can be easily modified to correct errors. • It can be used in variety of situations

12. Example: Tic-Tac-Toe program • complexity • use of generalizations • clarity of knowledge • extensibility

13. Program 1 • Board: 9-element vector 0 : blank, 1 : X , 2 : O • Move table: 39 Rows of 9-element vectors • Algorithm: 1. transform board vectorfrom base 3 to 10 2. use (1) as the move tableindex 3. change the board by using the vector from (2)

14. Program 1 • Program 1: • Data Structures: • Board: 9 element vector representing the board, with 1-9 for each square. An element contains the value 0if it is blank, 1 if it is filled by X, or 2 if it is filled with a O • Move table: A large vector of 19,683 elements ( 3^9), each element is 9-element vector.  • Algorithm: • 1.View the vector as a ternary number. Convert it to a decimal number. • 2.Use the computed number as an index into Move-Table and access the vector stored there. • 3.Set the new board to that vector.

15. Comments: • Advantages: efficient in terms of time, optimal game of tic-tac-toe in theory • Disadvantages: space - move tablespace work - move table error prone - move table three dimension - 327, no longer work at all

16. Program2 • Board: program1 2 : blank, 3 : X, 5 : O • Turn: gamemoves1,2,3,..... odd-numbered move : x even-numbered move : o Algorithm : 3sub procedures Make2: Board[5] or Board [2, 4, 6, or 8], Posswin (p): 18 (3*3*2)for p = X 50 (5*5*2)for p = O • Go (n) : Move to Board [n]

17. Program2 • Program 2: • Data Structure: A nine element vector representing the board. But instead of using 0,1 and 2 in each element, we store2 for blank, 3 for X and 5 for O • Functions: 3 functions are used. • Make2: Returns 5 if the center square of the board is blank, that is, if Board[5] = 2. Otherwise, this function returns any blank no corner square ( 2, 4, 6, or 8).

18. Program2 • Posswin( p ): Returns 0 if player p cannot win on his next move; otherwise, it returns the number of the square that constitutes a winning move. • This function will enable the program both to win and to block the opponent's win. • Posswin() operates by checking, one at a time, each of the rows, columns, and diagonals. Because of the way values are numbered, it can test an entire row (column or diagonal) to see if it is a possible win by multiplying the values of its squares together. If the product is 18 (3 x 3 x 2), then X can win. If the product is 50 (5 x 5 x 2), then 0 can win. If we find a winning row, we determine which element is blank, and return the number of that square.

19. Go( n ) : • Makes a move in square n . This procedure sets Board[n] to 3 if Turn is odd, or 5 if Turn is even. It also increments Turn by one.

20. Strategy

21. Comments: • Less efficient than Program 1 (time) • More efficient (space) • More clarity (strategy) • Easier to change (strategy) • Cannot extend to three dimension

22. Program2' • program 2board • magic square 15 • possible win check: S = sum of two paired owned by a player D = 15 – S if 0 < D < 10 and Board [D] is empty then the player can win

23. Program2' Comments: 1. Checking for a possible win is quicker. 2. Human finds the row-scan approach easier, while computer finds the number-counting approach more efficient.

24. Third approach Program 3: Look ahead at the board positions that result from each possible move. 1. If it is a win, give it the highest rating. 2. Otherwise, consider all the moves the opponent could make next. Assume the opponent will make the move that is worst for us. Assign the rating of that move to the current node. 3. The best node is then the one with the highest rating.

25. Comments • much more complex (time and space) • Extendable • AI technique

26. Criteria for success • How will we know if the machine we have constructed is intelligent? • Turing test. (proposed by Alan Turing) • Need two people and the machine to be evaluated. • One plays interrogator in one room • The machine and other person in another room. • Interrogator asks questions to either computer or to other person and receives response in writing but knows them as A and B and aims determine which is machine and which is the person. • If the machine succeeds in fooling the interrogator the machine is intelligent.

27. Can we measure AI achievement in restricted domain such as chess? • A program can acquire chess rating much same as human player i.e beating rated players. • DENDRAL – Analyzes organic compounds to determine their structures. • Hard to compare its achievements against chemists but… • It has produced analyses that have been published as original research results.

28. Concluding remarks • The problems AI is concerned with are varied, interesting and hard. • Solving such problems would give useful programs and a better understanding of human thought. • Do our best to set criteria to judge if we have solved the problem.

29. Concluding remarks • AI programs are easiest to build with languages designed to support symbolic manipulations such as LISP, PROLOG rather than numeric computation.

30. THANK YOU!!