Abstract | ||
---|---|---|
Although the detection of invariant structure in a given set of input patternsis vital to many recognition tasks, connectionist learning rules tend to focus ondirections of high variance (principal components). The prediction paradigm isoften used to reconcile this dichotomy; here we suggest a more direct approach toinvariant learning based on an anti-Hebbian learning rule. An unsupervised twolayernetwork implementing this method in a competitive setting learns to extractcoherent depth... |
Year | Venue | Keywords |
---|---|---|
1991 | ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 4 | hebbian learning,principal component |
Field | DocType | Volume |
Competitive learning,Semi-supervised learning,Multi-task learning,Computer science,Self-organizing map,Unsupervised learning,Learning rule,Artificial intelligence,Anti-Hebbian learning,Leabra,Machine learning | Conference | 4 |
Citations | PageRank | References |
9 | 26.33 | 9 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Nicol N. Schraudolph | 1 | 1185 | 164.26 |
Terrence J. Sejnowski | 2 | 8278 | 2135.10 |