CSU520 Artificial Intelligence, Fall 2004

Prof. Patrick Wang, Ph.D. and IAPR Fellow       pwang@ccs.neu.edu
Office 260 WVH            Tel. 617-373-3711
Office Hours: Thursday: 1:30-2:30pm, or by appointment

Class place and time: 108 WVG

Introduction

Weekly schedule

Course administration and rules

Lecture Slides

I. Introduction

This course provide you basic introductions to AI methods for modeling and implementing intelligent behavior in computers.   The focus of the course is on techniques for representing  knowledge about the world in formal, computer-understandable languages and creating programs that seek to achieve goals and/or solve problems within these knowledge frameworks. A major application to pattern recognition/computer vision in virtual environment through interactive learning will be discussed.

The three main approaches we will study are:

• formal logic representations of world/problem knowledge with various forms of logical inference for solving problems,

• state space representations of world/problem knowledge  with various forms of search for solving problems, and

• the application of machine learning techniques to create systems that can learn from examples and improve their performance with experience.

We will consider some applications of these ideas, in expert advisory systems, planning systems, 3D object recognition systems, and natural language systems.

Topics Covered:
I. Search
Problem solving through search
Heuristic search
Application to game playing
II. Logical inference
Propositional and predicate Logic
Techniques for automated inference
Application to query answering and/or logic programming
III. Machine Learning
Computational frameworks for systems that learn
Learning from examples
- Decision tree learning
Neural nets

Students will have hands-on assignments involving programming all or part of a problem-solving system, setting up problem data, and observing the performance of the system in at least two of the three areas.

Other Optional Topics:
Natural language or speech understanding
AI planning
Knowledge representation and semantic networks
Uncertain reasoning models
-- Bayesian inference
-- Default (non-monotonic) reasoning
Perception, Vision and Pattern Recognition using AI, Robotics and Applications

We will also briefly consider some alternative approaches such as semantic nets, frames, conceptual graphs, and case-based reasoning systems.

Other recommended references:

ARTIFICIAL INTELLIGENCE: STRUCTURES AND STRATEGIES FOR COMPLEX PROBLEM SOLVING 4th Edition, by George F. Luger, Addison Wesley publisher (2002)
ARTIFICIAL INTELLIGENCE 3rd Edition, by Patrick Winston, Addison Wesley publisher (1993)
PATTERN CLASSIFICATION, 2nd Edition, by R. Duda, P. Hart, and D, Stork, Wiley, (2001)
Handbook of Pattern Recognition & Computer Vision2nd ed, by C.H.Chen, L.F.Pau and P.S.Wang, World Scientific Pub. Co (WSP), (1999)

II. Weekly Schedule (Approximate)

Mid Term Exam Hint: Please note that as a reminder and as discussed in class, that mid term exam will cover alpha-beta cut-off search, min-max tree (e.g. whoever play first will win as example given in the class), state space search (e.g. 4x4 tiles, also known as 16-puzzle problem as discussed in class and notes), resolutiojn principle problem e.g. in HW3, textbook cha9-10, class slides and handouts), NLP natural language problem (e.g. CFL, grammars), Degree of ambiguity, learnability, recognizability (as discussed in class, handouts, reading papers, /pwang/home/publications, and /pwang/home/teaching/pr/prai/C1.htm selected page notes etc), hope this helps. Any further question you may have, please feel free to contact prof Wang x3711, 260WVH, good luck.

III. Course administration and rules

There will be 3 homework assignments/projects of varying credit depending on difficulty, a 90 minute midterm exam, and a  final exam.  The final will count 30% of  the course grade, the homework 40%, the midterm 25%, and class attendance/participation 5%.  However, you  must pass all three aspects of the course (the homework and the exams and class attendance/participation) to pass the course.   Late homework submitted within two days (48 hours) after the due date will be deducted 15%, within one week (7 days), will be deducted 30%, and no homework will be accepted after one week after due date.

Exams are open book, open notes.   Students MAY NOT share books, notes, calculators or any other items during exams.

Homework will be a combination of written and programming exercises. Homework  assignments are INDIVIDUAL assignments unless you are clearly told otherwise - it is not acceptable to turn in  the same homework paper or program as another person.

 Normally, failure to turn in an assignment or take an exam results in a  grade of 0 for that assignment.  Under extraordinary circumstances, your  grade on the final exam may be used to replace the missed assignment.

 No one will be excused from taking the final exam, so please do not  schedule a vacation or business trip on that day.