Matches in SemOpenAlex for { <https://semopenalex.org/work/W2059700573> ?p ?o ?g. }
Showing items 1 to 42 of
42
with 100 items per page.
- W2059700573 abstract "Neural networks: a new approach to pattern recognitionM. Schmutz, M. Rueff, U. MussigmannFraunhofer -Institute for Manufacturing Engineeringand Automation (IPA), Prof. Dr. -Ing. H.J. Warnecke,Nobelstr. 12, 7000 Stuttgart 80, GermanyAbstractThere are reviewed and discussed some of the basic ideas of a newapproach to pattern recognition which is based on the collective dy-namical properties of neural networks.1. IntroductionThe flexibility and ease with which human beings and animals per-form nearly instantly pattern recognition tasks is in striking con-trast to the abilities of present day man -made devices. Recent ad-vances in the modelling of neural networks, which were stimulated toa large extent by the Hopfield -model [1], promise to be a step to-wards closing this gap.The upsurge of interest in neural networks which exhibit featuresof associative memory is strongly influenced by recent developmentsin mathematics (theory of cellular automata) and physics (theory ofnonlinear dynamical systems, spin glasses) which gave new insight inthe self -organized emergence of computational properties of systemshaving a large number of simple equivalent processing elements suchas neurons. In this paper we review the basic ideas of this approachwhich overcomes some of the limitations of the earlier attempts(perceptrons [2,3], linear associative memory [4]).2. Pattern recognition as a nonlinear dynamical processFor simplicity, consider the set of discrete binary patterns con-sisting of N pixels. Each of the 2N patterns may be represented bya vector s= (si...,sN) where the components si may take two values,say si =1: black or si = -l: white.Now let us assume that the vectors s form the state space of adiscrete dynamical system given by means of a Boolean function F(s)" @default.
- W2059700573 created "2016-06-24" @default.
- W2059700573 creator A5023106272 @default.
- W2059700573 creator A5027372427 @default.
- W2059700573 creator A5081512215 @default.
- W2059700573 date "1988-03-22" @default.
- W2059700573 modified "2023-10-05" @default.
- W2059700573 title "Neural Networks: A New Approach To Pattern Recognition" @default.
- W2059700573 doi "https://doi.org/10.1117/12.942834" @default.
- W2059700573 hasPublicationYear "1988" @default.
- W2059700573 type Work @default.
- W2059700573 sameAs 2059700573 @default.
- W2059700573 citedByCount "1" @default.
- W2059700573 crossrefType "proceedings-article" @default.
- W2059700573 hasAuthorship W2059700573A5023106272 @default.
- W2059700573 hasAuthorship W2059700573A5027372427 @default.
- W2059700573 hasAuthorship W2059700573A5081512215 @default.
- W2059700573 hasConcept C153180895 @default.
- W2059700573 hasConcept C154945302 @default.
- W2059700573 hasConcept C41008148 @default.
- W2059700573 hasConcept C50644808 @default.
- W2059700573 hasConceptScore W2059700573C153180895 @default.
- W2059700573 hasConceptScore W2059700573C154945302 @default.
- W2059700573 hasConceptScore W2059700573C41008148 @default.
- W2059700573 hasConceptScore W2059700573C50644808 @default.
- W2059700573 hasLocation W20597005731 @default.
- W2059700573 hasOpenAccess W2059700573 @default.
- W2059700573 hasPrimaryLocation W20597005731 @default.
- W2059700573 hasRelatedWork W2033914206 @default.
- W2059700573 hasRelatedWork W2042327336 @default.
- W2059700573 hasRelatedWork W2046077695 @default.
- W2059700573 hasRelatedWork W2146076056 @default.
- W2059700573 hasRelatedWork W2163831990 @default.
- W2059700573 hasRelatedWork W2378160586 @default.
- W2059700573 hasRelatedWork W2386387936 @default.
- W2059700573 hasRelatedWork W2996038082 @default.
- W2059700573 hasRelatedWork W3003836766 @default.
- W2059700573 hasRelatedWork W3047965787 @default.
- W2059700573 isParatext "false" @default.
- W2059700573 isRetracted "false" @default.
- W2059700573 magId "2059700573" @default.
- W2059700573 workType "article" @default.