Winter 2009
CSS 482 will introduce you to a completely different way of programming, in which you specify rules of behavior, rather than algorithms. This is an especially powerful approach for problems that change often or where solutions involve application of human knowledge, rather than intricate calculations. Since their commercial introduction in the early 1980s, expert systems have undergone tremendous growth, representing the most successful application of artificial intelligence technology. Today, they are used in business, science, engineering, manufacturing, etc. Example applications include: computer configuration, fault diagnosis, computer-aided instruction, data interpretation, planning and prediction, and process control.
This course will have an additional focus on building expert systems applications as part of larger systems, including web-based and enterprise systems. Besides rule-based programming, expert systems operation, and knowledge engineering, topics will include aspects of Java that are useful for developing these systems, such as JavaBeans, serialization, applets, servlets, J2EE, JavaServer Pages, Tomcat, web services, and XML.
Recent Blog Entries
February 25th, 2009: Schedule updated
I seem to have a recurring problem with getting the end date for each quarter right. In our case, I was off by a week. So, please see the updated schedule. I've set the Wumpus World Competition for the last day of classes and the project presentations for finals week. Project reports are due at the end of finals week, 3/20. Of course, you should feel free to hand those in early, if you want.
February 11th, 2009: Grades now available online
I've added a link to the UW Catalyst grade book tool onto the menu of our class web site. You should be able to see your grades now for your written assignments and expert system 1. This is the first time I've used this, and it seems pretty simple, but let me know if there are any issues with it.
January 28th, 2009: Changes galore
A bunch of announcements:
1. I've uploaded some example circuits for expert system 1 to the class discussion forum.
2. I've emailed out the URL for registering your teams' possible meeting times with me next week.
3. I've extended the expert system 1 deadline until next Monday, 2/2, at 3:30PM.
4. I believe I will now be notified when someone posts to the discussion forum. Post something new there and see if it's true!
That's about all the news for now. Gotta run!
January 22nd, 2009: Guest lecture on January 28
"From Computational Complexity to Constraints on the Brain's Visual System to a Predictive Model of Human Vision and Attention"
John Tsotsos
The general problem of visual search can be shown to be computationally intractable in a formal complexity-theoretic sense, yet visual search is widely involved in everyday perception and biological systems manage to perform it remarkably well. Complexity level analysis resolves this contradiction. Visual search can be reshaped into tractability through approximations and by optimizing the resources devoted to visual processing. Architectural constraints can be derived to rule out a large class of potential solutions and the evidence speaks strongly against bottom-up approaches to vision. More importantly, the brain is not performing so-called general purpose vision; the vision problem the brain is solving is re-shaped using the constraints arising from the complexity analysis. In particular, the constraints argue for an attentional mechanism that exploits knowledge of the specific problem being solved. This analysis of visual search performance in terms of attentional influences on visual information processing and complexity satisfaction allows a large body of neurophysiological and psychological evidence to be tied together leading to a predictive model of vision and attention.
The Selective Tuning Model is a proposal for the explanation at the computational and behavioral levels of visual attention in humans and primates. Key characteristics of the model include a top-down coarse-to-fine winner-take-all selection process, a unique competition formulation with provable convergence properties, a task-relevant inhibitory bias mechanism, and selective inhibition in both spatial and feature dimensions for elimination of signal interference that leads to a suppressive surround for attended items. An extensive set of predictions arises many of which have now been supported by experiment. This presentation is an example of how a purely theoretical computational analysis can lead to a derivation of a model (without data-fitting or learning) that can have deep impact on brain science.
January 21st, 2009: Schedule updated
I've updated our schedule to reflect the elimination of written exercise 1 and the move of expert system 1 to it's earlier dates.
