CSci 5511: Class Notes

Material will be added to this page during the semester, in particular while a topic is being covered.

General resources on AI

The AI Topics library from AAAI offers a great collection of pages with information about AI and various subjects within AI. This is an excellent place to visit for people who want to get a good start at any of the subfields within AI.
You can find pointers to AI material at http://www.c2i.ntu.edu.sg/AI+CI/Overview/index.html.
Look at WHAT IS ARTIFICIAL INTELLIGENCE?, a paper from John McCarthy.
For another view, look at What is Artificial Intelligence?, an on-line philosophical introduction to AI by Jack Copeland. The article includes some history of AI, the Turing test, and more. Easy to read.

If you are looking for articles, go to CiteSeer at citeseer.ist.psu.edu, a great collection of searchable on-line papers. In addition to providing the papers, it provides links to the references and information on similar papers.

Material from the AIMA textbook

You should go to http://aima.cs.berkeley.edu/lisp/doc/install.html to download the Lisp software from the texbook. Follow the instructions posted there to install the software on your system. We will use the software for the programming assignment and in some homeworks.

Slides from the AIMA textbook will be posted here as we go through the material. They will also be linked in the class schedule in the syllabus.
Chapter 2, Chapter 3, Chapter 4a, Chapter 4a, Chapter 5, Chapter 6, Chapter 7, Chapter 8, Chapter 9,

Applets and demo programs

Tools for Learning Computational Intelligence at UBC has applets, including some for Search, CSP problems, and Planning. Check them out.

Intro to AI

The term Artificial Intelligence was first used in the mid 50s by John McCarthy.
"We propose that a two-month, ten man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire. The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. [proposed by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon. Additional attendees were Trenchard More, Arthur Samuel, Oliver Selfridge, Ray Solomonoff, Alen Newell, and Herbert Simon.]
Artificial Intelligence is an attempt to prove the Physical-Symbol System Hypothesis (PSSH).
From Allen Newell and Herbert Simon, Turing Award winners, "Computer Science as Empirical Inquiry: Symbols and Search," Communications of the ACM, March 1976, pp 113-126.
"A physical symbol system consists of a set of entities, called symbols, which are physical patterns that can occur as components of another type of entity called an expression (or symbol structure). Thus, a symbol structure is composed of a number of instances (or tokens) of symbols related in some physical way (such as one token being next to another). At any instant of time the system will contain a collection of these symbol structures. Besides these structures, the system also contains a collection of processes that operate on expressions to produce other expressions: processes of creation, modification, reproduction and destruction."
In the paper they state the Physical Symbol System Hypothesis: "A physical symbol system has the necessary and sufficient means for general intelligent action.
By necessary we mean that any system that exhibits intelligence will prove upon analysis to be a physical symbol system.
By sufficient we mean that any physical symbol system of sufficient size can be organized further to exhibit general intelligence.
By general intelligent action we wish to indicate the same scope of intelligence as we see in human action."

Search

Notes on search algorithms. The description of the algorithms in these notes is taken from J. Pearl, "Heuristics", Addison-Wesley, 1984.
Examples of how to do depth-first search in Lisp
Example of alpha-beta pruning. Example taken from J. Pearl, "Heuristics", Addison-Wesley, 1984.
for a nice explanation with figures of alpha-beta pruning go to http://www.emunix.emich.edu/~evett/AI/AlphaBeta_movie/sld001.htm
A complex search problem: searching for airfares. ITAsoftware is the leader in producing software that is used to search for airfares, schedule, and availability of flights. To get a glimpse at the complexity, think that there are 10,000 paths from San Franscisco to Boston, and for each path about 10,000 ways of pricing the itinerary.
- For a short description of the problem read Computer Scientists Find Unexpected Depths In Airfare Search Problem, by Sara Robinson of SIAM (Society for Industrial and Applied Mathematics)
- A much more detailed coverage of the problem and its solution is in Computational Complexity of Air Travel Planning, by Carl de Marken.
- Carl de Marcken: Inside Orbitz, a short story of how Lisp and other decisions helped solving the problem and obtain a great commercial success.
How MapQuest Works
an example of how to use constraints to solve Sudoku Puzzles using Python.

Logic, Resolution, and Knowledge Representation

Notes on resolution with propositional calculus and an example of using it.
Notes on unification.
An example of translating from English to predicate calculus with comments on the scope of quantifiers.
Baylor has prepared answers to the practice questions in converting English to predicate calculus questions done in class. Enjoy!
Try answering these practice questions on resolution. Once you are done, check out the answer to one of the questions.
For a better understanding of ontologies, look at the OpenCyc project. In particular, look at the Cyc 101 Tutorial. For another good example look at the Knowledge Sources of the Unified Medical Language System (UMLS).
What is ontology? FAQ from IBM.
The Protege ontology editor from Stanford.
The Semantic Web Wiki

Material on Planning

Practice questions on planning.
Try the applet at http://aispace.org/planning/. It lets you use the STRIPS representation for actions and states, and runs a planning algorithm to produce a plan. It works only in the blocks world, but has a nice graphical interface that let's you follow what happens step-by-step while constructing the plan and let's you execute the plan step by step. Very cute!
Useful papers on planning:
1. The paper by D. Weld, "An introduction to least commitment planning", AI Magazine, Winter 1994, pp 27-61 is a well written tutorial on least commitment planning algorithms, and includes detailed examples.
2. A more recent paper by D. Weld, "Recent Advances in AI Planning", AI Magazine, 1999 (postscript - 50 pages) describes advances in AI planning (it includes Graphplan and SAT planners).
3. The original paper on Graphplan by Avrim Blum and Merrick Furst "Fast planning through planning graph analysis," Artificial Intelligence, vol 90, pp 281-300, 1997 is available from http://www.cs.cmu.edu/~avrim/graphplan.html, the Graphplan home page.
If you want to try a partial order planner, try Sensory Graphplan from the University of Washington. Sensory Graphplan. It is based on Graphplan and supports extensions for conditional effects, quantification, etc. It is written in Lisp, and uses the PDDL langyage.
To learn about and to use a planner based on SAT, look at http://www.cs.rochester.edu/u/kautz/satplan/index.htm This planner combines graphplan with satplan. On the page you'll find links to papers such as Henry Kautz, David McAllester, and Bart Selman, "Encoding Plans in Propositional Logic", Proc. KR-96.

Material on Lisp

The first chapter, second chapter, and the Lisp code of ANSI Common Lisp by Paul Graham are available on the Web.