|
|
|
 |
Prof. Lisa Hellerstein
Spring 2004, Wed 3:30-5:45PM
A great challenge in computer research is to develop computer programs that can learn, generalize, and adapt. Research in machine learning has led to the development of computer programs for performing a variety of learning tasks, including recognizing handwritten digits, distinguishing good and bad credit applications, and navigating a robot in an unknown environment.
This course will introduce some of the primary approaches to machine learning, including inductive inference of decision trees, neural network learning, genetic algorithms, statistical learning methods, and reinforcement learning. It will discuss the ideas behind these techniques and experimental methods for evaluating the effectiveness of different techniques. The course will also introduce some fundamental theoretical problems in machine learning, such as determining which concepts can be learned by efficient algorithms.
The course will have homeworks (which will include written exercises, small programming assignments, and running and evaluating computer experiments), midterm and final exams, and a final project.
Prerequisites for the course are CS540 and CS603
or equivalent and an undergraduate course in probability/statistics. The
course is open to exceptional undergraduates who have senior standing
and at least a 3.5 GPA, and who have taken CS201 and Math 341 or equivalent.