590CE Computational Evolution and Digital Organisms

CSE 590CE is a weekly seminar on Readings and Research in Computational Evolution and Digital Organisms, open to all graduate students in computational, biological, and mathematical sciences. This course requires no background other than the ability to critically read papers and understand computer programs. It would be ideal to have attendees from various backgrounds to help make the discussions that much more expansive and insightful. Undergraduates can attend by permission of the instructor.

To investigate evolution, co-evolution, and the rise in complexity in digital organisms and ecosystems, the field of Computational Evolution studies adaptive, self-replicating, (usually) asexual digital organisms that are subject to random mutation. In some of the research, survival is aided and abetted by various reward schemes, and in some of the research, the simple ability to out reproduce one's neighbors is the only fitness function. There have been many intriguing results lately, but the field is very much in its infancy.

The goal of this reading course is to figure out this field and where it is heading. Are the results novel and valid? Can the results be explained better some other way? What scientific basis is there in work done so far? What are the prospects for relating the simulation results back to biology? How necessary is it to do so and still get some notion of validity? What are the fundamental issues that need to be addressed and solved? If the approach is flawed in some way, what is the right way to fix it? Can one develop a theoretical framework to explain the results? What could one do to help move the field forward in the best direction?

We will be meeting on Mondays from 1:30 to 2:20 in CSE (Allen Center) 503. The instructor is Daniel Weise. This course is offered for 1-3 credits (1 for reading papers and participating, 2 for presenting papers, and 3 for also completing a course project).

Several of the systems we will be studying are available to use and extend. Those taking the course for 3 credits will be expected to either do experiments using these existing systems, or to create a new system. Based on our reading, the course instructor has a ready list of projects that can easily be attempted and completed in a quarter, or you can propose your own.

The course mailing list is cse590ce@cs.washington.edu. You can manage your subscription via the web.

Reading List

Other papers worth reading

Tierra papers

Avida papers

Other papers

Relevant Links

The person who restarted the field of digital evolution was Tom Ray who created the Tierra system.

The major researchers in computational evolution and artifical life each have their own courses they teach. Charles Ofria's digital evolution course lists many more papers than we will be reading, and also covers some of the theory of digital evolution. Chris Adami was Charles Ofria's dissertation advisor and is among the most prolific digital evolution researchers. His course covers a lot of theory (much of which he created) and biology. Mark Bedau has an interesting wide ranging course that covers a lot on non-computational areas.