Dynamical-Systems Behavior in Recurrent and Non-Recurrent Connectionist Nets An Honors Thesis Presented 1990

Abstract:  A broad approach isdeveloped for training dynamical behaviors in connectionist networks. General recurrent networks are powerful computational devices, necessary for di cult tasks like constraint satisfaction and temporal processing. These tasks are discussed here in some detail. From both theoretical and empirical considerations, it is concluded that such tasks are best addressed by recurrent networks that operate continuously in time|and further, that e ective learning rules for these continuous-time networks must be able to prescribe their dynamical properties. A general class of such learning rules is derived and tested on simple problems. Where existing learning algorithms for recurrent and non-recurrent networks only attempt to train a network's position in activation space, the models presented here can also explicitly and successfully prescribe the nature of its movement through activation space. I am indebted to Jay Rueckl, my advisor, both for his suggestions and for his support. Jay Rueckl and Greg Galperin provided computational facilities that proved indispensable. Iwould also like to thank my family and the many friends whose encouragement saw this project through its nal stages.