Artificial Intelligence

CMPS 240: Artificial Intelligence Spring Quarter  2013

       Summary:  "HOW TO USE KNOWLEDGE TO MAKE GOOD DECISIONS...."

      Lectures: TuTh  10-1145pm  B214 Earth and Marine Sciences

      Instructor: Robert Levinson Office Hours: M2-4pm  after class or
      by appointment.  Phone number: 459-2087 (x92087 on campus)

      E-mail:
      levinson@soe.ucsc.edu (use it!) 

      PREREQs:  Grad. Standing, or instructor's permission.  Previous
      AI course is helpful but not required.

      (CMPS140 this year = approximately chapters 1-5,18,21)


      REQUIRED  BOOKS and TOOLS:

      AI:  "Artificial Intelligence: A Modern Approach" Stuart Russell
      and Peter Norvig Prentice Hall, Third Edition.  Webpage:
      http://aima.cs.berkeley.edu/ (Buy at Baytree Bookstore)


      CH:    the UCSC Reader by Chris Harrison:
          Determining Satisfiability Through Subtraction And a Binary
          Constrained XOR Form (BXCF).
             Includes SAT and PH homework problems in the back.
              (Buy at Baytree Bookstore)


      PH:  ( 8  Probability Problems in back of CH.)

 

        SAT Paper to critique:  (download from

        http://www.louis-coder.com/P_NP_Problem_solved/P_NP_Problem_solved.htm

      

WEBSITE for CHESS TACTICS TRAINING:

     www.chesstempo.com (register for free!)

      OPTIONAL:

       ET:    Jaynes, E.T. and G. Larry Brethorstt
      Probability Theory:  Logic of Science.  Cambridge Press 2003

      Or download from the web:
               http://omega.albany.edu:8008/JaynesBook.html

           Chapters 1-5
                http://bayes.wustl.edu/etj/prob/book.pdf]
      P:
      http://www.dartmouth.edu/~chance/teaching_aids/book_articles/probability_book/amsbook.mac.pdf

 

        

      =========================================================================

      PROJECTS/HOMEWORK


   1.  Chess: 
      Tactics: Get the highest blitz tactics  chess rating you can at
      chesstempo.com. Must practice at least 4 hours per week.  You
      will be tested in realtime at end of quarter so cheating won't
      help!

      Everybody will also get an algorithm score based on the class
      consensus algorithm made up of 12 rules.

   A.   Chess Journal:  (-4 to +4 points possible)
              0. How much time put in? How is my rating doing?
               1. What have I learned about myself and/or chess?

               2.
              Consider  Heuristics/Algorithm for solving chess tactics
               problems.

                Select 1 "best" new heuristic/pattern for class
                consumption.

                  If your heuristic is selected for the "class
                  consensus algorithm" you get 5 bonus points.  The
                  more mathematical/precise the better.  For example:
                      "First consider all checks, they may be
                      checkmate!"

              Approx. 2 pages per checkpoint. Bonus in {-1,0,+1} per
              checkpoint.

   B.   CHESS TACTICS will be  evaluated  and ranked in 2 factors:
           RATING =  min (Final Rating, Final Rating - Improvement )

                           = min (Final Rating, Target Rating) ....
           IMPROVEMENT = Final Rating - Target Rating.

NOTE:
  You must play tactics at a rating of 850 or above by end of quarter.
  You will select the target rating at the beginning of the quarter.


   2.   Applied Probability Theory:  8 Probability Problems in the back
   of CH

   3    Determining Satisfiability Through    Subtraction:
            See Chris Harrison Reader
                 Satisfiability Problems  in back of CH (don't handin -
                 fair game for final).

        You need to write a paper (500-5000 words  ) arguing for the
        correctness or incorrectness of the following paper and
        improving it. This will be due and also presented at  the final
        exam.

            http://www.louiscoder.com/P_NP_Problem_solved/P_NP_Problem_solved.htm
        (your paper will be ranked against the other student papers).



   4.     Probability and Decision Theory
                   Russell-Norvig AI textbook.  Problems are at the end
                   of the chapters and are scheduled in the syllabus.
                   
EVALUATION

Each of the 4 areas will be combined for final grade out of 100.
   1. Chess Tactics: 25 percent    [RATING(7.5) + IMPROVEMENT(7.5) +
   CLASS ALGORITHM(10)  + JOURNAL{-4,4}  BONUSES[0-20]

 2. Probability
   Handout Problems (8 total) 20 percent

  3. AI Problems. 30 percent

  4. 25percent
   3SAT: 10 percent on exam on CH satisfiabilty and methods taught in
   class.  (H0 will be ten percent of your final exam score)  15
   percent on paper review or improvement on P=NP paper.

[Chess rating, chess improvement and the 3SAT paper will be graded
competitively 70-100].

FinalGrade Levels:  

                   90% A- or better

                   80% B- or better  (Passing for Grads.)

                   70% C or better  (Passing for Ugrads.)

      You may work in consistent groups of 1-3 people.

      It is fine and encouraged to discuss homework problems  with
      other  students - BUT CHEATING or ACADEMIC DISHONESTY on any
      course item (such as direct verbatim copying  from a  member
      outside  your  group  or  during an exam) will result in not
      passing the course and other extremely undesirable consequences.
      Written problem solutions should reflect YOUR OWN  UNDERSTANDING
      and be in YOUR OWN WORDS. All references used should be cited.

      NO INCOMPLETES WILL BE GIVEN.
===============================================================