Dayton AI Class 0

We had our zeroth Dayton AI class on Monday evening. We are taking Stanford's offering of their Introduction to AI class for the fall term. They offer two tracks, a basic track and an advanced track. The main difference is that for the advanced track you have to do the homework assignments and take the exam, and then you get a certificate of completion. Mostly this preparatory class consisted of Joe and I eating Little Caesar's, generating scrap for the furnace and tapping away on our laptops. We collected our meager knowledge of install-guides and tutorials, so this post will give a run-down of each of those areas for Python and Lisp (if you are a Java hacker and are interested, please consider joining us, the exercises are all worked-out in Java too).

Installing Python Language and Development Environment

Joe will be using Python on the Windows side with the vanilla installer off of the Python website. The installer "just works", no special tweaking required.

Installing Lisp Language and Development Environment

I will be using Steel Bank Common Lisp (SBCL) and the SLIME Emacs mode in Fedora. Installing in my case starts with a one-liner: yum -y install sbcl emacs-slime
I also had to edit my .emacs file (the emacs start-up file in my home directory) to include this:
(setq inferior-lisp-program "sbcl")
(require 'slime)

And then, to run the slime mode you tell emacs: M-x slime. This resulted in an error for me about the swank-loader. Google is my friend, the fix was to find the absolute path to swank-loader.lisp and put that into a variable with M-x customize-variable slime-backend.

Tutorials and Getting Started

I've found that "Dive in to Python" is a pretty good intro to Python, this comes as a package on Fedora, and you can download the pdf too if you are on Windows. Joe says that the Stanford course website links to the Python lessons on the Kahn Academy, and those looked pretty interesting.

On learning Lisp, I thought these couple posts by Peter Christensen collected together lots of good info

Peter actually recommends a book by one of the instructors of the Stanford course (Norvig): Paradigms of Artificial Intelligence Programming: Case Studies in Common Lisp


Well, that should get you up and running (or at least staggering). Please ask questions in comments or on the mailing list.

Even if you aren't able to commit to signing up for the Stanford course officially, you are welcome to come hang out on Monday evenings this fall with Joe and I at the space while we try to hack our Python and Lisp software agents into artificial sentience. Possibly coming to a sumobot near you...

No comments:

Post a Comment