CSC 8520, Spring2010
Artificial Intelligence
Weds 6:15 - 9:00, Mendel G92
Dr. Paula Matuszek

Syllabus:

This syllabus is a work in progress; expect changes!
Jan 13: Introduction and Intelligent Agents..
Intro presentation.     Assignment 1.
Note: Additional versions of slides for the book are available at the text book web page.
Lab Presentation.    
Jan 20:: Basic Search..     Reading: Chapters 3 and 4
Basic Search presentation.   
Jan 27: Adversarial Search.    Introduction to Prolog Reading: Chapter 5.
Search presentation.     Prolog presentation.
Prolog Assignment 1.    Example Program.
Dave's Concise Introduction to Prolog.   
Accessing Prolog.    Arithmetic Example.
Feb 3: Constraint Satisfaction. Reading: Chapter 6.
Constraint Satisfaction presentation.
Lab     Simple Puzzle Example    
Position Puzzle Example     Next To Puzzle Example
Assignment 2.    Negatives Puzzle Example
Feb 10:
Feb 17: Logical Agents. Reading: Chapters 7-9.
Logic Presentation.   Lab Presentation
Assignment 3:     Sample Prolog Program     Data for lab example
Project Information.
Feb 24: Planning. Reading: Chapter 10-11
Planning presentation.   Planning Lab
A simple Prolog planner     Abstract data type library used by planner    
Assignment 4     Code for lab exercise
Mar 3: Spring Break.
Mar 10: Midterm.
Important: The midterm will be a take-home. We will not
meet at Villanova on March 10.
Midterm notes.   
Mar 17: Knowledge Representation. Reading: Chapter 12
Knowledge Representation presentation.
Mar 24: Robotics   Reading: Ch 22.
Robotics Presentation.
Mar 31: Machine Learning   Reading: Chapter 18
Machine Learning Presentation.   
Presentation summary due April 5: Mail to Paula.Matuszek@villanova.edu
a one-paragraph writeup of the topic of your presentation. These will be
posted on the class webpage.
April 7: Guest Speaker: Eric Eaton
Guest Lecturer: Dr. Eric Eaton. Dr. Eaton received his
PhD from the University of Maryland, Baltimore County,
with a specialization in Machine Learning. He now researches
machine learning at Lockheed Martin. He will present additional
material in machine learning and some information about his
own research in interactive learning.
ML: KNN and Naive Bayes Presentation.   
ML: SVMs presentation.   
April 14: Natural Language Processing   Readings: Ch 22 and 23
Natural Language Presentation.
April 21:
Student Presentations.
Palanisamy Ramamoorthy: Reinforcement learning: In case of an agent
acting on its environment, the agent receives some evaluation
of its action (reinforcement), and learns from the success and
failure, reward and punishment. Including: Passive Reinforcement
learning: The agent simply watches the world going by and tries
to learn the utilities of being in various states. Active
Reinforcement learning: The agent not simply watches, but also acts.
Kory Kirk: The topic is the Genetic Wavelet Algorithm,
which is a new type of genetic algorithm recently developed by
Syncleus, a company from Philadelphia.
http://wiki.syncleus.com/index.php/DANN:Genetic_Wavelets#-
a wiki article that I wrote on it, an excerpt from my independent
study paper. # Some necessary details of#Genetic algorithms as well.
Andrew Burke:Machine Learning and Web Analytics
Sarah Stroman: Webcrow crossword application. I will provide a
general description about the program and walk through a demonstration
of how it works, then the algorithm and heuristics used as well as
some information about the web mining that it performs.# I will
conclude the presentation by discussing how well it performed,
outlining some of its strengths and weaknesses.
April 28: Student Presentations.
Tyson Kennedy: Perception, imaging and reconstructing the 3D world.
Lavanya Ragothaman: Nature has always helped researchers
in finding solutions for many problems. One such inspiration offered
by Nature is in observing the emergent behavior of social animals
which lead to the development of a field of study known as Swarm
Intelligence (SI). SI identifies the collective behavior of decentralized,
self-organized systems. A population of simple agents can discover a
relatively complex and global intelligent behavior through their interaction
among themselves and with their environment. The multi-agent systems are
characterized with the following properties; each agent has incomplete
information or capabilities for solving the problem and, thus, has a limited
viewpoint; there is no system global control; data are decentralized;
and computation is asynchronous. In addition to these characteristics
the SI systems are self-organizing that are accompanied with feedbacks
both positive and negative, reliance on multiple interactions and along
with certain degree of randomness. The benefits of the SI systems include
collective completion of the task, no need for very complex algorithms and
they are adaptable to the changes in the environment.
Andrew Hampson:Use of Lego Mindstorms in classroom education of
Artifical Intelligence. The purpose of this presentation is the give an
overview of the use of Lego Mindstorms in the instruction of Artificial
Intelligence at the collegiate level.# Examples will be given of projects
from undergraduate and graduate curricula both in the United States and
around the world.# The ability of the system as a toy and teaching aid
will be explored through the hardware, the software, and the extension
languages.# The presentation will end with a hands-on demonstration of
a robot showcasing the information discussed.
Daniel Priece: Brain Simulation. Brain Simulation in the
field of Artificial Intelligence deals with the understanding of
the capabilities of the human, or other animal, brain. The approach
of Brain Simulation in Artificial Intelligence aims to create a
hardware and software based machine to mimic the cognitive capabilities
of a brain. Brain Simulation is seen as an attempt at strong AI, where
a machine is indistinguishable from a brain. Many projects exist in
this field including the "Blue Brain" project, which is a reverse
engineering of the human brain down to a molecular level, and IBM's
brain simulation projects, in which an IBM team was able to match
the processing power of a cat's brain. The race is on to create a
simulation of the human brain, which some experts are saying could
happen within ten years.
May 5: Final Exam Due.
Resolution and Bayes' Examples.