Abstract | ||
---|---|---|
We use a simple network which uses negative feedbackof activation and simple Hebbian learning toself-organise in such a way as to produce a featuremap which has the property of identifying the relativeproportions of the components of the input data.Thus it evaluates the angular properties of the inputdata space and ignores the magnitude of the inputdata. When used on unprocessed voice data, the networkis shown to extract both the phase and the frequencyinformation from the raw data. We ... |
Year | DOI | Venue |
---|---|---|
1997 | 10.1007/978-3-7091-6492-1_22 | ICANNGA |
Keywords | DocType | Citations |
hebbian learning | Conference | 2 |
PageRank | References | Authors |
0.55 | 1 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Stephen Mcglinchey | 1 | 9 | 2.94 |
Colin Fyfe | 2 | 508 | 55.62 |