CSC104 Issues in Artificial Intelligence
Paper
Due 9 May 4:00PM

For collaborations, I naturally expect more. I would like the paper to be longer and more in depth. I would say 7-9 pages. More depth and detail is more important than increased length.

The topic is up to you. It should be on some issue or issues in Artificial Intelligence, either directly related to one of the ten readings, or relevant to those readings. Original research is not necessary. It is fine with me to rely solely on the ten readings. Original thoughts is more important than original research. Look over the readings, and think of which ones engaged you most deeply, on which topics you have the strongest opinions, the most original thoughts. A thoughtful rebuttal of a single reading would be fine as a topic. So would exploring any topic more deeply, either through the links on the syllabus, or via a Web search. With search engines so good by now, it is relatively easy to quickly locate relevant literature on the Web.

About the only thing I don't want is a summary of one or more readings, unless in a concise form to pave the way for your own argument. Write this as if I will be the primary reader; so you don't need to spend much time explaining the papers to me. I would prefer your paper not to be filled with unsupported opinion and reaction: "I like Searle's argument; the Churchlands don't convince me." It is fine to say that but I would like to see justification, a reasoned argument yea or nea, as opposed to just opinions.

I ask for a commitment to a topic by Tuesday Apr 22nd. I will comment on any rough draft given to me by Thursday the 24th. in time to return by our last class Thursday 1 May. This is 25% of your grade, so it would be wise to take it seriously.

I will maintain a list of possible topics as they occur to me, just to give you some ideas (not to constrain you!). For currency, check the last update date: .

• The Turing Test: Fifty Years Later. Is it still relevant? Is the Loebner competition useful, or a hindrance?
• The Loebner Competition. History, relevance, analysis.
• Hard Games: Bridge, Go, Poker. What can computers currently do? How do the algorithms work? What are future prospects, both for the games and for larger implications in AI?
• CYC. Explore the Web site, read some criticism.
• Searle's Chinese Room. Find the latest criticisms and rebuttals by Searle, and assess the significance of the argument. See The Mind's I ed. Hofstadter and Dennett.
• Machine Translation. The current status and future prospects.
• Ned Block's Humungous Look-up Table. Locate his paper (see Notes3a), read it, and critically assess it.
• Emotions. Are emotions necessay for intelligence? Perhaps look at the work of Antonio Damasio (The Feeling of What Happens), or of Joseph LeDoux (The Emotional Brain).
• Neural Nets: Status and Prospects. What is the current state of the art in neural networks? Are they living up to their promise, or do we have another instance of early success in toy domains (NETTALK) not followed up by comparable advances?
• Learning. What is the current status of Machine Learning? Does any of it deserve the term "learning"?
• Dawkins on Biomorphs. Read enough of The Blind Watchmaker to thoroughly understand Biomorphs, experiment with on-line applets, and draw conclusions for the possibilities of genetic algorithms.
• Genetic Programming. Find some of Koza's GP examples, figure out what's going on in detail, and assess whether the results are truly impressive. [Our library has several books and conference proceedings written/edited by Koza.]
• Genetic Algorithms, et al.. Investigate the annual GECCO conferences. See the syllabus for a link to the 2002 conference. Penetrate it, and perhaps a few earlier ones, and this summer's, enough to get a feel for the variety and direction of the field. Write about that.
• Mutations? Explore the question we left unresolved in class: has research in genetic algorithms actually shown evolutionary biologists something they did not previously appreciate? Or was it more ignorance on the part of the computer scientists? Start with that passage in Levy's chapter ("one in 10M genes experience mutations") and try to find references.
• Data Mining. Find a particular, interesting data mining application, read a paper on it, critically evaluate it, and draw conclusions.
• Scientific Discovery. Study one of the several successful scientific discovery programs (see Notes6b) via on-line papers, and critically assess its success. Are its discoveries really "discoveries"?
• What is Brooks up to Now?. Find out what is the latest research of Rodney Brooks, report, and assess the prospects for his research program.
• RoboSoccer. Write a detailed report on the latest RoboSoccer competition, and assess progress, relevance, and prospects.
• A-Life. Look for ALife Open Problems, find the original paper, and assess the significance of this list.