Learning to Generalize from Single Examples in the Dynamic Link Architecture
- 1 September 1993
- journal article
- Published by MIT Press in Neural Computation
- Vol. 5 (5) , 719-735
- https://doi.org/10.1162/neco.1993.5.5.719
Abstract
A large attraction of neural systems lies in their promise of replacing programming by learning. A problem with many current neural models is that with realistically large input patterns learning time explodes. This is a problem inherent in a notion of learning that is based almost entirely on statistical estimation. We propose here a different learning style where significant relations in the input pattern are recognized and expressed by the unsupervised self-organization of dynamic links. The power of this mechanism is due to the very general a priori principle of conservation of topological structure. We demonstrate that style with a system that learns to classify mirror symmetric pixel patterns from single examples.Keywords
This publication has 15 references indexed in Scilit:
- A Learning Algorithm for Boltzmann Machines*Published by Wiley ,2010
- Distortion invariant object recognition in the dynamic link architectureIEEE Transactions on Computers, 1993
- Oscillatory responses in cat visual cortex exhibit inter-columnar synchronization which reflects global stimulus propertiesNature, 1989
- A learning mechanism for invariant pattern recognition in neural networksNeural Networks, 1989
- Coherent oscillations: A mechanism of feature linking in the visual cortex?Biological Cybernetics, 1988
- A Neural Network for Invariant Pattern RecognitionEurophysics Letters, 1987
- Function of the thalamic reticular complex: the searchlight hypothesis.Proceedings of the National Academy of Sciences, 1984
- Visual symmetry recognition by pigeonsPsychological Research, 1982
- Perception of symmetry in infancy.Developmental Psychology, 1981
- On the perception of symmetrical and repeated patternsPerception & Psychophysics, 1974