Abstract

A novel modular connectionist architecture is presented in which the networks composing the architecture compete to learn the training patterns. As a result of the competition, different networks learn different training patterns and, thus, learn to compute different functions. The architecture performs task decomposition in the sense that it learns to partition a task into two or more functionally independent vii tasks and allocates distinct networks to learn each task. In addition, the architecture tends to allocate to each task the network whose topology is most appropriate to that task, and tends to allocate the same network to similar tasks and distinct networks to dissimilar tasks. Furthermore, it can be easily modified so as to...

Citations

3339	Pattern Classification and Scene Analysis - Duda, Hart - 1973
2307	Learning internal representations by error propagation - Rumelhart, Hinton, et al. - 1986
976	Multilayer feedforward networks are universal approximators - Hornik, Stinchcombe, et al. - 1989
936	Self-Organized Formation of Topologically Correct Feature Maps - Kohonen - 1982
576	Precis of the modularity of mind - Fodor - 1985
527	The society of mind - Minsky - 1986
439	Beyond Regression: New Tools for Prediction and Analysis - Werbos
290	Connectionist Learning Procedures - Hinton - 1990
251	Feature discovery by competitive learning - Rumelhart, Zipser - 1985
232	A New Approach to Manipulator control: the Cerebellar Model Articulation Controller - Albus - 1975
228	Attractor dynamics and parallelism in a connectionist sequential machine - Jordan - 1986
219	Connectionist models and their properties - Feldman, Ballard
170	A Theory of Networks for Approximation and Learning - Poggio, Girosi - 1989
163	Competitive learning: From interactive activation to adaptive resonance - Grossberg - 1987
156	Serial Order: A parallel Distributed Processing Approach - Jordan - 1986
149	An Analogue Approach to the Travelling Salesman Problem Using an Elastic Net Method - Durbin, Willshaw - 1987
143	Learning Automata: An Introduction - Narendra, Thathachar - 1989
136	Object vision and spatial vision: Two cortical pathways, Trends in Neurosci - Mishkin, Ungerleider, et al.
92	A Dimension Reduction Framework for Understanding Cortical Maps. Nature 343: 644–647 - Durbin, Mitchison - 1990
75	Modular construction of time-delay neural networksfor speech recognition - Waibel - 1989
73	and network design strategies - LeCun - 1989
71	Hopfield; Large automatic learning, rule extraction and generalization - Denker, Schwartz, et al. - 1987
63	Learning logic - Parker - 1982
61	Pattern recognizing stochastic learning automata - Barto, Anandan - 1985
55	Dynamic connections in neural networks - Feldman - 1982
50	The meta-pi network: building distributed knowledge representations for robust pattern recognition - Hampshire, Waibel - 1989
46	Building and understanding adaptive systems: a statistical/numerical approach to factory automation and brain research - Werbos - 1987
45	Adaptation and decorrelation in the cortex - Barlow, Földiák - 1989
42	Cortical connections and parallel processing: Structure and function - Ballard - 1986
42	Une procedure d’âpprentissage pour reseau a seuil asymmetrique - LeCun - 1985
38	Supervised learning and systems with excess degrees of freedom - Jordan - 1998
36	Gradient following without back-propagation in layered networks - Barto, Jordan - 1987
32	A parallel computation that assigns canonical object-based frames of reference. In: Joint conference on artificial intelligence - Hinton - 1981
29	Two problems with backpropagation and other steepest-descent learning procedures for networks - Sutton - 1986
27	and imagining in the cerebral hemispheres: A computational approach - Kosslyn - 1987
23	Cascaded Back Propagation on Dynamic Connectionist Networks - Pollack - 1987
22	Learning sets of filters using back-propagation - Plaut, Hinton - 1987
19	Shape representation in parallel systems - Hinton - 1981
18	Why are "what" and "where" processed by separate cortical visual systems? A computational investigation - Rueckl, Cave, et al. - 1989
17	The programmable blackboard model of reading - McClelland - 1986
16	A winner-take-all mechanism based on presynaptic inhibition feedback - Yuille, Grzywacz - 1989
12	Modular learning in neural networks - Ballard - 1987
11	Aspects of cortical organization related to selective attention and selective impairments of visual attention - Cowey - 1985
9	Parameter Nets - Ballard - 1984
9	The psychological status of overgenerated sentences. Cognition 19.101-‐131 - Freedman, Forster - 1985
9	Nonlinear Dynamics of Artificial Neural Systems - Maxwell, Giles, et al. - 1986
6	Why have multiple cortical areas - Barlow - 1986
6	Properties of a competitionbased activation mechanism in neuromimetic network models - Reggia - 1987
6	Skeleton filters in the brain - Sejnowski - 1981
4	Using a context-sensitive learning network for robot arm control - Yeung - 1989