Courses at Stanford for Logical AI People
Here are some courses I highly recommend for students interested in
Logical AI. I list the professor I took it under, who is usually the
same teacher year after year. I've also tried to list the courses in
the order that they build on each other.
Classes in mathematics I believe are the most fundamental for anyone
interested in logical AI. To formalize common sense reasoning, you
need to be able to use mathematics to formalize things, and know some
common sense. Most people already have common sense, although I must
admit many academics like myself are a little deficient in that area.
Hence all you need to learn is how to represent concepts in
- Phil 160A: Logic, ?.
- CS 157: Logic and Automated Reasoning , Mike Genesereth.
- CS 323: Algebraic Logic, Vaughan Pratt. More algebra.
- Math290A/B: Model Theory, Solomon Feferman. A must for
logical AI people.
- Math 291A/B: Recursion Theory, Solomon Feferman. Good for
mathematical rigor and concepts, introduction to notions of
computability. Math291B does recursion
theory for reals.
- Math292A/B: Set Theory, Solomon Feferman. Very good for
seeing how to build structure, prove
You should also go to these seminars:
221: Artificial Intelligence, Daphne Koller. A great
introduction to the newest concepts of AI. Includes search, bayes
nets, neural nets, a little logic, MDPs, game trees, CSPs, planning,
satisfiability, decision trees.
- CS 323:
Commonsense Reasoning, McCarthy/Costello. An introduction to
many of John
McCarthy's papers, including circumscription, philosophical
presuppositions of logical AI, elaboration tolerance, and context. We
also explore other fundamental parts of logical AI, including a
brief introduction to logic and model theory, and the frame problem.
- Phil 169: Intensional Logic, Johan van Benthem. Modal
logic, all sorts of good stuff. van Benthem is a
- PHIL 298: Logical Dynamics Johan van Benthem. This course
turns logic into an intriguing toy. van Benthem transforms the
mundane task of proving a formula into an engaging argument, or even
more exciting, a game played against another person who is constantly
trying to prove you wrong! This course is a lot of fun.
- CS 228: Reasoning Under Uncertainty, Daphne Koller. Learn
about Bayes Nets, representational as well as computational problems.
How to represent causality. Even though this isn't logical AI persay, it is
good to know about other approaches to codifying knowledge. Even
though Bayes Nets are built on top of probability distributions, there
is still a lot of formalization that is good to know and see.
Not as necessary but good for other reasons:
- CS 528: Broad Area Colloquium for Artificial Intelligence,
Geometry, Graphics, Robotics and Vision
- Logic Seminar (offered through the math department)
- Logic Lunch (offered through the math department)
Classes I haven't taken. (But would like to!)
- CS 051: Introduction to Quantum Computing, Colin
Williams. Fun and cool! (Especially if you understand and appreciate
- CS 255: Cryptography, Dan Boneh. Beautiful applications
of number theory to get cryptographic solutions. Cryptography is
related to logical AI in that we are both using logic/math to built
objects that satisfy certain properties. (They want functions for
which it is difficult to find a collision, we want formalisms that can
handle the frame problem elegantly.)
- Golf: Golf, Jim Miller. A fun sport to learn. Also a
great way to get a tan.
In general, the rule of thumb is to take any course offered by these
people: (I'm sure that there are others, I just
haven't attended their courses.)
- CS 222: Knowledge Representation, Richard Fikes.
- CS 227: Reasoning Methods in AI, Pandurang Nayak.
- LING 230A/B : Semantics and Pragmatics -- plan to take this quarter
- Phil 160B L Computability and Logic Grisha Mints. Spring 2001.
- Dan Boneh
- Solomon Feferman
- Mike Genesereth
- Daphne Koller
- John McCarthy
- Grigori Mints
- Vaughan Pratt
- Colin Williams
- Johan van Benthem
Last Modified: Wed May 28 10:56:38 2003