The  Weizmann  Institute  of  Science
                  Faculty of Mathematics and Computer Science


                            Computer Science Seminar

                                Michael Kositsky
                         Department of Computer Science
                     University of Massachusetts at Amherst


                                 will speak on


                            A Cluster Memory Model 
                       for Learning Sequential Activities

Abstract:
This study presents a computational model for learning to generate complex
extended actions from examples.  The continuous trajectory is represented in
terms of elementary actions. The sequence of elementary actions move the plant
through intermediate states that form compact clusters in state space.  The
compact clusters allow efficient interpolation and generation of new
trajectories.

An important part of this research is a set of psychophysical tests aimed at
revealing the cluster representation hypothetically used by the brain.  The
experiment includes the entire motor learning process, starting from scratch up
to using highly skilled actions.





                      The lecture will take place in the 
                     Lecture Hall, Room 1, Ziskind Building
                           on Monday, March 20, 2000
                                    at 14:30