Season Greetings

[John Anderson]

profitable year modeling.


John Anderson continues his inexorable progress towards becoming an
empty nester.  Mike Matessa has completed his dissertation (on
communication to support collaborative problem solving) and has gone on
to Nasa Ames.  Raluca Budiu is completing her dissertation on
integration of background knowledge in sentence processing.  On the
other hand ACT-R and cognitive modeling are continuing to grow at CMU. 
Christian Lebiere is now a research scientist in the Human Computer
Interaction Institute and Marsha Lovett is a faculty member in the
Psychology Department.  Cognitive modeling in service of educational
applications is continuing to prosper in Marsha's hands and in the
hands of Al Corbett and Ken Koedinger who are directing the PACT
Center.


The ACT-R project is continuing its LAB (Learning, Application, and
Brain) agenda.  On the learning front the focus is on learning from
instructions, which is pushing issues of knowledge representation and
production compilation.  On the application front Christian Lebiere and
Dan Bothell are working on the Amber air-traffic control project.
Monchu Chen, Scott Douglass, Frank Lee, and Myeong-Ho Sohn are working
on the GT-ASP AAWC project.  On the brain front Jon Fincham, Qulin Qin,
and Myeong-Ho Sohn have been working on relating fMRI studies to ACT-R.
 We can report our first ACT-R fMRI publication:


</fontfamily><fontfamily><param>Geneva</param>Sohn, M.-H., Ursu, S.,
Anderson, J. R., Stenger, V. A., & Carter, C. S.

(2000). The role of prefrontal cortex and posterior parietal cortex in

task-switching. Proceedings of National Academy of Science.
13448-13453.


</fontfamily><fontfamily><param>New_York</param>Hopefully, Jon Fincham
will be shortly following with a similar publication on the Tower of
Hanoi.  We are also involved in projects relating ACT-R to more analog
domains.  This includes Glenn Gunzelmann's work on imagery and spatial
reasoning and Alex Petrov's work on judgements of continuously varying
quantities.


All of this research is feeding into our efforts to define and develop
the ACT-R 5.0 architecture.  With the emerging conception of retrieval
as an action analogous to visual attention, a "buffer model" conception
of ACT-R productions is emerging. The condition of a production tests
against various "buffers" such as the goal buffer, the retrieval
buffer, or the visual buffer and the action causes transformations in
these buffers.  Christian Lebiere and Mike Byrne are working on the
increased integration between ACT-R and R/PM that this conception
requires.  Dan Bothell and Jon Fincham are working on the development
of a new 5.0 environment and a 5.0 course.


Much of this summer's post-graduate school will be devoted to
discussing the issues involved in ACT-R 5.0.  We will also be
discussing the issue of more distributed responsibility for the ACT-R
community.  The distribution of speakers at the post-graduate school
can be taken as a potential model for future summer schools.  We hope
to see you in West Virginia this summer and have you participate in
discussions about the future of ACT-R.


Best Wishes for the New Year.

</fontfamily>