CS 448 Machine Learning -- Fall 2011
- Instructor : Jim Levenick
- Office: 206 Ford (6486)
- Office Hours (Fall 2011):
- W9-11
- Or by appt. (send email).
- Email anytime! levenick(at)willamette.edu
- Texts:
- Braitenberg, Vehicles
- Hofstadter & Dennett, The Mind's I
- Rojas Neural Networks
- Merker Consciousness without a Cerebral Cortex
Overview:
This course will address machine learning techniques and philosophical
issues concerning machine learning. Learning techniques will include:
Darwinian optimization techniques (sometimes called artificial evolution
or artificial life), learning in simulated neural networks, plus
state space tree search and modification. Philosophical issues include:
"Can a machine be intelligent?","Can a digital computer without a body be intelligent in the way people are?",
"Is consciousness a precondition of intelligence?",
"Is intelligence without emotion an oxymoron?", and, "How can machine intelligence
be assessed?".
Approach:
Class time will be used mostly for lectures and group discussions
of readings and related issues. Out of class time will be divided
between reading and programming in Java.
The texts address philosophical and technical aspects of machine learning
and intelligence. You will write programs to implement some learning
techniques from scratch; for others, working programs will be provided
to experiment with and/or modify. Students (selected at random at
the beginning of the discussion period) will present the 2 or 3 main ideas in the day's reading.
Evaluation:
Your grade will be determined by exams, labs, and discussions. Labs will
count 40%, exams 60%. There will a mid-term, and a final which will count double the mid-term. Exams will cover philosophical
and technical topics. Labs must be demonstrated on time; late labs will not be accepted (as
solutions will typically be distributed immediately after the due date; you can use the
sample solution to do the next part of the lab if you didn't finish yours).
Plagiarism:
You are welcome to use any and all code you find in books, notes, on disk
or on the Internet -- so long as you have permission and credit your source.
All work you hand in as your own must be your own. The penalty for
misrepresenting another's work as your own (commonly called plagiarism)
is failure in the course.
Finally:
If you feel the class is going too fast, or too slow, or in the wrong direction
-- let me know!