A two-step algorithm for learning from unspecific reinforcement
We study a simple learning model based on the Hebb rule to cope with `delayed', unspecific reinforcement. In spite of the unspecific nature of the information-feedback, convergence to asymptotically perfect generalization is observed, with a rate depending, however, in a non-universal way on le...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article (Journal) |
| Language: | English |
| Published: |
1999
|
| In: |
Journal of physics. A, Mathematical and theoretical
Year: 1999, Volume: 32, Issue: 31 |
| ISSN: | 1751-8121 |
| DOI: | 10.1088/0305-4470/32/31/301 |
| Online Access: | Verlag, Volltext: http://dx.doi.org/10.1088/0305-4470/32/31/301 Verlag, Volltext: http://stacks.iop.org/0305-4470/32/i=31/a=301 |
| Author Notes: | Reimer Kühn and Ion-Olimpiu Stamatescu |
| Summary: | We study a simple learning model based on the Hebb rule to cope with `delayed', unspecific reinforcement. In spite of the unspecific nature of the information-feedback, convergence to asymptotically perfect generalization is observed, with a rate depending, however, in a non-universal way on learning parameters. Asymptotic convergence can be as fast as that of Hebbian learning, but may be slower. Morever, for a certain range of parameter settings, it depends on initial conditions whether the system can reach the regime of asymptotically perfect generalization, or rather approaches a stationary state of poor generalization. |
|---|---|
| Item Description: | Gesehen am 06.03.2018 |
| Physical Description: | Online Resource |
| ISSN: | 1751-8121 |
| DOI: | 10.1088/0305-4470/32/31/301 |