Transcript 419-roass-2005 - Department of Computer Science

74.419 Artificial Intelligence
Course Introduction and ROASS Document
Instructor & Course Info
Course Topics and approximate Schedule
Assignments and Grade Breakdown
The usual Stuff (including ‘How to fail this
course’)
Students introduce themselves
74.419 Artificial Intelligence
- Course Info Time:
Mon, Wed, Fri 2:30-3:20 pm
Prerequisites:
74.319 Introduction to Artificial
Intelligence
Web page:
http://www.cs.umanitoba.ca/~cs419
News Group:
local.cs419
Instructor:
Dr. Christel Kemke
Textbook:
Russell and Norvig: Artificial
Intelligence - A Modern Approach,
Prentice Hall, 1995 / 2003
Instructor Info
Dr. Christel Kemke
524 Machray Hall / 412 Engineering 2
Phone: (204) 474-8674
e-Mail: [email protected]
Web:
www.cs.umanitoba.ca/~ckemke
Office Hours: Tuesday 1:30 - 2:30 pm
Thursday 1:30 - 2:30 pm
by appointment
Course Objectives
Acquire essential and advanced knowledge and
skills in Artificial Intelligence concepts and
methods
Learning
basic and advanced knowledge
Experience
through exercises and projects
Skills
scientific, research and applied
Preparation
for work and advanced studies
Class Structure
Classes will comprise of
• Lectures with Notes
• Online Videos, Movies (PBS, SRI)
• Some Labs (KR, maybe NLP)
• Student Presentations ("project
presentations day", conference style)
• Occasional in-class Exercises
Topics Outline
 The Dream – Flakey and the others
– General Introduction to Agents
– SRI Movie
 Knowledge Representation & Reasoning
–
–
–
–
–
–
–
Warming up with Propositional Calculus
First-Order Predicate Logic including Semantics !
Representation with Frames, Inheritance Hierarchies, ...
Description Logics
Ontology - where elephants sometimes have 3 legs
Allen’s Time Logic (if there is time ...)
Exotic Logics (like Deontic Logic) - Do we have to go to
class? Or can we go to class? Or can we not go?
Topics Outline 2
 Planning
–
–
–
–
–
–
–
–
Introduction with Shakey - The first "Real Robot"
Planning as Search – STRIPS
ABSTRIPS
Partial-Order Planning
Hierarchical Plan Decomposition
Situation Calculus
Standard Problems (everywhere in this section)
Special Topics or maybe more videos?
Topics Outline 3
 Natural Language Processing
– a short introduction to Speech Recognition - is
speech nothing but 'hot air'?
– Overview of Natural Language Processing (NLP) what did you say?
– Syntax, Grammar - "What's up" - is this really a real
sentence?
– Syntactic Sentence Analysis, or Parsing
– a little bit of Semantics - what do you mean?
– something about Discourse and Dialogue - or
"Could you please close the door?"
– Videos and demos
Topics Outline 4 – ‘Free Style’
(optional, depending on time)
Neural Networks
–
–
–
–
General NN Model & Processing
NN Architectures
Learning Paradigms for Neural Networks
some example demos, maybe a video
Evolutionary Algorithms
– General Principles of Evolutionary Computing
– more videos than theory
The Standard Textbook
Stuart Russell and Peter Norvig,
Artificial Intelligence – A Modern Approach,
Prentice Hall, 1995 & 2003
available in The Bookstore, ~ 95 CAD
Textbook: Table of Contents
I. ARTIFICIAL INTELLIGENCE.
1. Introduction.
2. Intelligent Agents.
II. PROBLEM-SOLVING.
3. Solving Problems by Searching.
4. Informed Search and Exploration.
5. Constraint Satisfaction Problems.
6. Adversarial Search.
III. KNOWLEDGE AND REASONING.
7. Logical Agents.
8. First-Order Logic.
9. Inference in First-Order Logic.
10. Knowledge Representation.
In Addition: Description Logics
Non-Standard Logics
Semantics of FOPL
IV. PLANNING.
11. Planning.
12. Planning and Acting in the Real World.
Focus:
Situation Calculus
Partial Order Planning
Hierarchical Planning
V. UNCERTAIN KNOWLEDGE AND REASONING.
13. Uncertainty.
14. Probabilistic Reasoning Systems.
15. Probabilistic Reasoning Over Time.
16. Making Simple Decisions.
17. Making Complex Decisions.
VI. LEARNING.
18. Learning from Observations.
19. Knowledge in Learning.
20. Statistical Learning Methods.
21. Reinforcement Learning.
VII. COMMUNICATING, PERCEIVING, AND
ACTING.
22. Communication.
23. Probabilistic Language Processing.
24. Perception.
25. Robotics.
Focus:
Deterministic Natural Language
Processing (Parsing, a little bit
Semantics)
VIII. CONCLUSIONS.
26. Philosophical Foundations.
27. AI: Present and Future.
Second AI Textbook
Nils J. Nilsson, Artificial Intelligence – A New
Synthesis, Morgan Kaufman, 1998
Another AI Textbook
George Luger and William Stubblefield: Artificial
Intelligence, Addison-Wesley, 1998 and 2001 (CS
319 textbook)
Reference Book - NLP
Daniel Jurafsky / James Martin, Speech and
Language Processing, Prentice Hall, 2000
Reference Book - KR and Logic
Richard A. Frost, Introduction to KnowledgeBase Systems, Collins, 1986
too old to be shown; excellent book.
Assignments
3 Standard Homework Assignments
• Knowledge Representation
• Planning
• Natural Language Processing
1 Individual Research Report
• Essay
• Presentation
• Program
1 Group Project
Group Project
Group Project (with typically 3 Students)
Design and Implementation of an Intelligent Agent
System with Knowledge Base (KB), Planning Module,
and Natural Language Interface (NLI), or an equivalent
project.
examples:
– a Household Robot
– a Mars Rover
– a Scheduling / Planning System with NLI
Students need to write a proposal for the project and a
final report, plus give a presentation with demo.
Grade Breakdown
Homework
Individual Research
Project
Final Exam
30%
10%
10%
50%
---------100%
Occasionally, Bonus Points are issued for exceptional
efforts beyond the requirements.
Course Schedule (approximately)
Introduction & Agents
Knowledge Representation
KR Lab
Natural Language Proc.
Planning
Free Style
Group Project Presentation
Exam preparation
week 1-2
week 3-4
week 5
week 6-7
may swap
week 8-9
week 10-11
week 12
last week
Deadline Policy
Assignments are to be submitted before the due date.
Unless otherwise specified, they have to be dropped
into the 419 slot, left of the entrance to the Cargill
Lab.
If electronic submissions are requested, they have to
be sent to [email protected] .
Extensions to a deadline can be granted only by the
instructor (that's me, Dr. Kemke). In general, no late
assignment will be accepted after the marked
assignments have been returned.
KEEP COPIES OF SUBMITTED ASSIGNMENTS!
Class Communication, Notes, Attendance
 Class Notes will in general be provided via the course
web page.
 Non-web material will be made accessible on-line, per
handout in class, or in a folder on reserve for 74.419 in
the Science Library.
 Class attendance is not checked but students are
responsible for knowing the contents of the classes.
Officially, students are required to attend classes.
 You can use the news group to discuss course related
questions and problems with your 419 class mates.
 Otherwise, questions should be addressed to me
personally or via e-mail.
Illness and other problems
In case of longer times of illness or other
problems like bereavement, which
considerably influence class attendance and
course performance, students are advised to
contact the instructor in order to find
arrangements for continuing successfully with
the course.
In case students encounter other substantial
course-related problems, they are also
advised to contact the instructor or TA.
Misuse of Computer Facilities, Plagiarism, and
Cheating
• These serious offenses will carry sanctions.
Copying of assignments or parts thereof from
anywhere without appropriate references,
cheating on exams, or misusing facilities will
result in punishment ranging from course failure
to prosecution.
• Please see section 7 of the General Academic
Regulations and Requirements in the U of M
General Calendar for more information.
Final Exam
Time and location of the final exam will be
announced by the Student Records Office.
It is your own responsibility to make yourself
aware of the posted exam schedules. You are
obligated to make yourself available for the
writing of the final exam.
How to Fail this Course or
Get a Bad Grade
• A good starting point is not to attend classes on
a regular basis.
• Do not look at the Course Notes either. Just
forget about the whole course web site.
• Never ever ask or talk to fellow students or the
instructor about the course contents. If you
missed a class (or more) or if you can't grasp
something, just hide and play cool.
How to Fail this Course or
Get a Bad Grade
• Don't cooperate with your project partners in the
group project. Tell them you have so many other
things to do, you just don't have time to meet
and work with them.
• Do not come to the presentation of your group
project. Or ask the instructor for a last minute
change of the schedule. (The best time is the
morning of the day when your presentation is
scheduled.)
How to Fail this Course or
Get a Bad Grade
The safest way to fail the course is:
Do not attend the final exam.
Go on holidays during exam time, or hide in a safe
place. Wait until exam time is over.
After Christmas, you get in touch with the Faculty /
Department / Instructor and ask for special
permission to take the exam now. Of course, you
have no evidence of having been terribly ill, or
proof of other serious problems, which would
excuse that you missed the exam.
Course Partner
Every student is asked to have a course partner!
Course Partners have a mutual commitment to help each
other with course related questions and problems.
They have to inform each other about the class contents,
in case one of them missed a class (for good reasons,
not just to dump it on the other one).
In case both course partners have to miss a class, they
are asked to contact other students or the instructor, to
inform themselves about the missed class.
Course partners are also supposed to help each other and
to cooperate - within in reasonable limits - with the goal
to understand the course material and expand their
knowledge.