caa a22 4500 463244893 CHVBK 20180405153323.0 cr unu---uuuuu 170326e20070701xx s 000 0 eng 10.1007/s00422-007-0162-4 doi (NATIONALLICENCE)springer-10.1007/s00422-007-0162-4 Learning with incomplete information and the mathematical structure behind it [Elektronische Daten] [Reimer Kühn, Ion-Olimpiu Stamatescu] We investigate the problem of learning with incomplete information as exemplified by learning with delayed reinforcement. We study a two phase learning scenario in which a phase of Hebbian associative learning based on momentary internal representations is supplemented by an ‘unlearning' phase depending on a graded reinforcement signal. The reinforcement signal quantifies the success-rate globally for a number of learning steps in phase one, and ‘unlearning' is indiscriminate with respect to associations learnt in that phase. Learning according to this model is studied via simulations and analytically within a student-teacher scenario for both single layer networks and, for a committee machine. Success and speed of learning depend on the ratio λ of the learning rates used for the associative Hebbian learning phase and for the unlearning-correction in response to the reinforcement signal, respectively. Asymptotically perfect generalization is possible only, if this ratio exceeds a critical value λ c , in which case the generalization error exhibits a power law decay with the number of examples seen by the student, with an exponent that depends in a non-universal manner on the parameter λ. We find these features to be robust against a wide spectrum of modifications of microscopic modelling details. Two illustrative applications—one of a robot learning to navigate a field containing obstacles, and the problem of identifying a specific component in a collection of stimuli—are also provided. Springer-Verlag, 2007 Kühn Reimer Department of Mathematics, King's College, London, UK aut Stamatescu Ion-Olimpiu FESt, Heidelberg and Institut für Theoretische Physik, Universität Heidelberg, Heidelberg, Germany aut Biological Cybernetics Springer-Verlag 97/1(2007-07-01), 99-112 0340-1200 97:1<99 2007 97 422 https://doi.org/10.1007/s00422-007-0162-4 text/html Onlinezugriff via DOI 1 review-article jats NATIONALLICENCE 856 40 https://doi.org/10.1007/s00422-007-0162-4 text/html Onlinezugriff via DOI NATIONALLICENCE 700 1- Kühn Reimer Department of Mathematics, King's College, London, UK aut NATIONALLICENCE 700 1- Stamatescu Ion-Olimpiu FESt, Heidelberg and Institut für Theoretische Physik, Universität Heidelberg, Heidelberg, Germany aut NATIONALLICENCE 773 0- Biological Cybernetics Springer-Verlag 97/1(2007-07-01), 99-112 0340-1200 97:1<99 2007 97 422 Metadata rights reserved Springer special CC-BY-NC licence nationallicence BK010053 XK010053 XK010000 NATIONALLICENCE NATIONALLICENCE NL-springer