Evolino for recurrent support vector machines
| dc.creator | Schmidhuber, Juergen | |
| dc.creator | Gagliolo, Matteo | |
| dc.creator | Wierstra, Daan | |
| dc.creator | Gomez, Faustino | |
| dc.date | 2005-12-15 | |
| dc.date.accessioned | 2026-07-07T06:53:51Z | |
| dc.date.available | 2026-07-07T06:53:51Z | |
| dc.description | Traditional Support Vector Machines (SVMs) need pre-wired finite time windows to predict and classify time series. They do not have an internal state necessary to deal with sequences involving arbitrary long-term dependencies. Here we introduce a new class of recurrent, truly sequential SVM-like devices with internal adaptive states, trained by a novel method called EVOlution of systems with KErnel-based outputs (Evoke), an instance of the recent Evolino class of methods. Evoke evolves recurrent neural networks to detect and represent temporal dependencies while using quadratic programming/support vector regression to produce precise outputs. Evoke is the first SVM-based mechanism learning to classify a context-sensitive language. It also outperforms recent state-of-the-art gradient-based recurrent neural networks (RNNs) on various time series prediction tasks. | |
| dc.description | 10 pages, 2 figures | |
| dc.identifier | https://arxiv.org/abs/cs/0512062 | |
| dc.identifier | http://arxiv.org/abs/cs/0512062 | |
| dc.identifier.uri | http://salesiana.dossiersoluciones.com/handle/123456789/105734 | |
| dc.subject | Neural and Evolutionary Computing | |
| dc.subject | F.1.1; I.2.6 | |
| dc.title | Evolino for recurrent support vector machines | |
| dc.type | text |