AI News, Lecture 1 CS6800 Artificial Intelligence: artificial intelligence

Computer Science, PhD

A minimum of 48 semester hours of course work beyond the BS/BA degree is required of all students.

The student must maintain a minimum grade-point average (GPA) of 3.500 among the six core courses satisfying the above course requirements and receive a grade of B or better in each of these courses.

To demonstrate research ability, the student is required to submit to the PhD committee a research or a survey paper in an area of specialty under the supervision of a faculty advisor.

A submitted paper from a student is considered to have fulfilled the paper requirement if: Upon completion of the course and the research paper requirements, the student is admitted to candidacy for the PhD degree.

This requirement is fulfilled when the student works as a teaching assistant (TA) or instructor of record (IoR) for one semester and during this semester: PhD students are expected to satisfy the teaching requirement some time after completing their first year and at least one semester prior to scheduling their PhD defense.  The examination is taken after the student has achieved sufficient depth in a field of study in order to prepare a prospectus for the PhD dissertation.

Prior to taking the examination, the student prepares a dissertation proposal, which describes the proposed research, including the relevant background materials from the literature.

Upon approval of the written proposal, the student has to present the proposed work orally in a public forum, followed by a closed-door oral examination from the dissertation committee.

After the establishment of degree candidacy, a maximum of five years will be allowed for the completion of the degree requirements, unless an extension is granted by the college graduate committee.

Introduction to Artificial Intelligence Evolutionary Computing Henry Kautz.

1

2

3

4

Myopic Local Search The local search methods we have discussed so far are myopic – they only look at the immediate neighborhood of a single state at any one time Simple Parallelism: run many searches in parallel with different random seeds –Prob(Success) = 1 – Prob(Run Fails) k –E.g.: Prob(Run Fails) = 90%, k = 10  Prob(Success) = 65% {

'name': 'Myopic Local Search The local search methods we have discussed so far are myopic – they only look at the immediate neighborhood of a single state at any one time Simple Parallelism: run many searches in parallel with different random seeds –Prob(Success) = 1 – Prob(Run Fails) k –E.g.: Prob(Run Fails) = 90%, k = 10  Prob(Success) = 65%',

'description': 'Myopic Local Search The local search methods we have discussed so far are myopic – they only look at the immediate neighborhood of a single state at any one time Simple Parallelism: run many searches in parallel with different random seeds –Prob(Success) = 1 – Prob(Run Fails) k –E.g.: Prob(Run Fails) = 90%, k = 10  Prob(Success) = 65%',

5

6

'name': 'Genetic algorithms A successor state is generated by combining two parent states Start with k randomly generated states (population) A state is represented as a string over a finite alphabet (often a string of 0s and 1s) Evaluation function (fitness function).',

7

'name': 'Example: 8-Queens Fitness function: number of non-attacking pairs of queens (min = 0, max = 8 × 7/2 = 28) 24/( ) = 31% 23/( ) = 29% etc Normalized Fitness',

'description': 'Example: 8-Queens Fitness function: number of non-attacking pairs of queens (min = 0, max = 8 × 7/2 = 28) 24/( ) = 31% 23/( ) = 29% etc Normalized Fitness',

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

“ Normal” crossover operators will often lead to inadmissible solutions Many specialised operators have been devised which focus on combining order or adjacency information from the two parents Crossover operators for permutations 1 2 3 4 5 5 4 3 2 1 1 2 3 2 1 5 4 3 4 5 {

25

26

27

28

29

30

31

32

33

34

35

36

37

38

Computational Bottleneck Consider networks for sorting 16 numbers: every child must be tested on 2 16 inputs Cost of doing all these tests comprises nearly all of the computation time Strategies: –Parallelize evaluation of children [1,000 node cluster] –Use special hardware (FPGA) to evaluate [Koza et al 1997] {

'name': 'Computational Bottleneck Consider networks for sorting 16 numbers: every child must be tested on 2 16 inputs Cost of doing all these tests comprises nearly all of the computation time Strategies: –Parallelize evaluation of children [1,000 node cluster] –Use special hardware (FPGA) to evaluate [Koza et al 1997]',

'description': 'Computational Bottleneck Consider networks for sorting 16 numbers: every child must be tested on 2 16 inputs Cost of doing all these tests comprises nearly all of the computation time Strategies: –Parallelize evaluation of children [1,000 node cluster] –Use special hardware (FPGA) to evaluate [Koza et al 1997]',

39

40

41

42

Swarm Algorithms, Briefly Idea: –Each insect in a swarm is local search process –At each step, each insect: Looks around its neighborhood Decides which direction looks best Communicates what it found out to (some of) the other insects According to a random coin flip, –Moves in the direction that looks best locally –Moves in the best direction it hears about –Moves in some weighted combination of the above {

'name': 'Swarm Algorithms, Briefly Idea: –Each insect in a swarm is local search process –At each step, each insect: Looks around its neighborhood Decides which direction looks best Communicates what it found out to (some of) the other insects According to a random coin flip, –Moves in the direction that looks best locally –Moves in the best direction it hears about –Moves in some weighted combination of the above',

'description': 'Swarm Algorithms, Briefly Idea: –Each insect in a swarm is local search process –At each step, each insect: Looks around its neighborhood Decides which direction looks best Communicates what it found out to (some of) the other insects According to a random coin flip, –Moves in the direction that looks best locally –Moves in the best direction it hears about –Moves in some weighted combination of the above',

43

44