caa a22 4500 477034373 CHVBK 20180405111301.0 cr unu---uuuuu 170330e19960401xx s 000 0 eng 10.1007/BF00417784 doi (NATIONALLICENCE)springer-10.1007/BF00417784 Improving reinforcement learning in stochastic RAM-based neural networks [Elektronische Daten] [Alistair Ferguson, Hamid Bolouri] RAM-based neural networks are designed to be efficiently implemented in hardware. The desire to retain this property influences the training algorithms used, and has led to the use of reinforcement (reward-penalty) learning. An analysis of the reinforcement algorithm applied to RAM-based nodes has shown the ease with which unlearning can occur. An amended algorithm is proposed which demonstrates improved learning performance compared to previously published reinforcement regimes. Kluwer Academic Publishers, 1996 hardware realisation nationallicence RAM-based nodes nationallicence reinforcement learning nationallicence reward-penalty nationallicence Ferguson Alistair ERDC, University of Hertfordshire, AL10 9AB, Hatfield, Herts, UK aut Bolouri Hamid ERDC, University of Hertfordshire, AL10 9AB, Hatfield, Herts, UK aut Neural Processing Letters Kluwer Academic Publishers; http://www.wkap.nl 3/1(1996-04-01), 11-15 1370-4621 3:1<11 1996 3 11063 https://doi.org/10.1007/BF00417784 text/html Onlinezugriff via DOI 1 research-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/BF00417784 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Ferguson Alistair ERDC, University of Hertfordshire, AL10 9AB, Hatfield, Herts, UK aut NATIONALLICENCE 700 1- Bolouri Hamid ERDC, University of Hertfordshire, AL10 9AB, Hatfield, Herts, UK aut NATIONALLICENCE 773 0- Neural Processing Letters Kluwer Academic Publishers; http://www.wkap.nl 3/1(1996-04-01), 11-15 1370-4621 3:1<11 1996 3 11063 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer