Introduction to Artificial Intelligence

Fall, 1998

Instructor

Chris Welty

Contents

  • Announcements/Messages
  • Textbooks
  • Background
  • Course Schedule and Projects
  • Grading Policy

    Messages:
    • 12/10: You can view the catalog of web pages here.
    • 11/16: The project 3 description is ready.
    • 10/28: The Project 2 description is ready.
    • 10/7: There is now a dropbox on the Untangle server called "heuristics". If you put some code in there (follow the rules: name your file the same as your function, name your functions with your name before each one), it will immediately be included on the othello server.
    • 9/28: The othello.lisp file has been updated to include the proper play in the case one player can't move.
    • 9/20: Project 1 is ready. Note that since the assignment was late coming on line, I have extended the deadline to Oct. 5, however it is due before class on that Monday.
    • 9/9: The Index link for the CLRM has been fixed.
    • 9/8: The Common LISP Reference Manual is available on line for LISP reference.
    • 9/2: The lab will be Mondays, 12:30-1:30. Every Week!.

    Schedule:

    Week Topic Chapters Assignment Lab
    8/31 Introduction: AI, LISP AI 1 Homework 1
    9/7 Part I:
    Intelligent Agents    		 Deliberative Systems
    Intelligence as Search    		 LISP Loops, State Spaces AI 7,8 Homework 2
    		 LISP
    9/14 Heuristics, Playing Games AI 9 Project 1 LISP
    9/21 Minimax AI 12 Othello
    9/28 Reactive Systems
    Emergent Intelligence    		 Stimulus-Response AI 2 Homework 3 CL-HTTP
    10/5 S-R Robots SR1 IC
    10/12 Threshold Logic Units AI 3 Project 2
    10/19 Fall Break
    10/26 Part II:
    Knowledge Representation    		 Logic-Based Systems
    Deterministic Behavior    		 Propositional Logic AI 13,14 Robots
    11/2 Predicate Logic AI 15,16 Robots
    11/9 Frames and Description Logics AI 18; SR2 Project 3 Logic
    11/16 Ontological Analysis SR3 Classic
    11/23 Soft Computing Systems
    Probabalistic Behavior    		 Probablilistic Reasoning AI 19.1-19.2 Homework 5 Classic
    11/30 Bayesian Networks AI 19.3-19.8 Project 4 Probablity
    12/7 Conclusion: Wrapup and tearful goodbyes Bayes Nets

    Textbooks:

    Nilsson, Nils. Artificial Intelligence: A New Synthesis. Morgan Kaufman, 1998. [AI]
    Shapiro, Stuart. Common LISP: An Interactive Approach. Computer Science Press, 1992. [CL]

    Supplemental Readings:

  • Brooks, Rodney. Artificial Intelligence through Building Robots. MIT AI Lab Memo 899, May 1986. Available in PS. [SR1]
  • Brachman, R., McGuinness, D., Patel-Schneider, P., Borgida, A. and Resnick, L. Living with CLASSIC: When and How to Use a KL-ONE-Like Language. Principles of Semantic Networks. Morgan Kaufman. Pp. 401-456. May, 1991. [SR2]
  • Welty, Chris. The Ontological Nature of Subject Taxonomies. In N. Guarino, ed., Formal Ontology in Information Systems. IOS Press Frontiers in AI Applictions Series. Trento, Italy. June, 1998. Available in HTML and PS.[SR3]
  • Steele, Guy. Common LISP: The language. Digital Press.

    Background:

    AI is a field which overlaps many other disciplines, delving into the very roots of Computer Science, Philosophy, Psychology, Neurobiology, Cognitive Science, etc. Due to the enormous breadth of the field, it is not possible even to glimpse a significant portion of it in one semester.

    Modern AI is divided across numerous lines. As a discipline of computer science, however, it is mostly concerned with the development and study of intelligent systems. While this is still a vague distinction, we will attempt to illuminate it this semester by cutting it up in two different ways.

    First, we will study systems that interact with the world, which in the modern AI vernacular are called "agent" systems. It is widely believed that true intelligence can only be embodied in systems that are capable of interacting with some world. We will experiment with and contrast the two opposing approaches to building intelligent agents: deliberative and reactive systems. For the former, we will build agents that play the game Othello; for the latter, we will program small robots to exhibit simple insect-like behaviors.

    In the second half of the semester, we will turn our attention to knowledge, and the role it plays in intelligence and intelligent systems. Much of the AI enterprise has been consumed with studying ways in which knowledge can be abstracted out of a system as a separate module, and how that knowledge should be represented in such a way that a system could behave as if it understood what it supposedly knows. We will experiment with and contrast the two opposing approaches to dealing with knowledge: the traditional logic-based approach and the nouveau soft computing approach. For the former, we will model a small knowledge domain in CLASSIC, a knowledge representation system; for the latter, we will experiment with Bayesian Networks to model a domain in which uncertainty is prevalent.

    It should be noted at the outset that this is by no means a comprehensive or even adequately representative list, however it will serve to give a flavor of what AI, as a field of research and practice in computer science, is all about.

    This will be a "hands on" course, and we will be experimenting with four different systems that will require some training, and that you will be expected to use to complete projects. Therefore there will be an extra laboratory session each week, which we will schedule during the first week of class. The purpose of this laboratory session will be to ensure that everyone keeps up and gains an acceptable level of proficiency with the software and hardware that will be used this semester, without detracting from the pedagogical content of the course.

    Grading:

    60%: Programming Projects
    20%: Homeworks
    10%: Participation
    10%: Quizzes

    The homeworks are small problem sets designed to be completed during the lab sessions. Collaboration with other students and the lab leader is expected. The projects will be major programming tasks, and during project periods the lab sessions will be devoted to issues relating to the projects. Collaboration on the Projects is permitted, however what you turn in must be your own work. Blatant copying will result in a zero grade. The late penalty for Projects is 20% per day, with the weekend counting for one day.

    Class participation is not a guaranteed 15%. You must attend classes and labs and demonstrate you have done the reading by participation in class. In addition, my travel schedule this semester is quite busy. For days that I will be absent, I will assign a student to lead the class or lab. This student will be responsible for doing the reading the week before, and meeting with me to discuss the material and the strategy for presentation. Half the participation grade will be based on these presentations. Following such classes, a quiz will be administered to all students. Overall performance on the quizzes will count toward the evaluation of the student leader.