A unified framework for connectionist systems

DOI10.1007/BF00317773MaRDI QIDQ1117857zbMATH OpenOpenAlexWikidataFDO

Authors Richard M. Golden

Publication date 1988

Published in Biological Cybernetics (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1007/bf00317773

zbMATH Keywords

\textit{M. A. Cohen} and \textit{S. Grossberg} networkadditive dynamical system summarizing functionback-propagationClassificationconnectionist systemsHopfield modellearning processesMAP estimationmaximum a posteriorimaximum likelihood estimationneural network modelssubjective probability function

Mathematics Subject Classification ID

Applications of statistics (62P99) Learning and adaptive systems in artificial intelligence (68T05) Artificial intelligence (68T99) Physiological, cellular and medical topics (92Cxx) Other natural sciences (mathematical treatment) (92F05)

Review text

The paper proposes to treat connectionist systems (neural network models) within MAP (maximum a posteriori) estimation framework. The fundamental theorem assigns to every additive dynamical system summarizing function (specifying activity levels of visible neurons X for a given set of connection strengths) the subjective probability function representing the network's belief that X occurs in its environment. Classification and learning processes are then considered and evaluated from this viewpoint for the most popular neural networks paradigms [back-propagation; \textit{M. A. Cohen} and \textit{S. Grossberg} network, IEEE Trans. Syst. Man Cybern. 13, 815-826 (1983; Zbl 0553.92009); Hopfield model].

Recommendations

• A neural network unification
• On connectionist models
• scientific article; zbMATH DE number 1164157
• A framework to deal with interference in connectionist systems
• Connectionist networks for constraint satisfaction
• Towards a unifying logical framework for neural networks
• scientific article; zbMATH DE number 4072428

Cites work

• A universal prior for integers and estimation by minimum description length
• Absolute stability of global pattern formation and parallel memory storage by competitive neural networks
• Asymptotic Properties of Non-Linear Least Squares Estimators
• scientific article; zbMATH DE number 3152611 (Why is no real title available?)
• scientific article; zbMATH DE number 3513115 (Why is no real title available?)
• scientific article; zbMATH DE number 3436645 (Why is no real title available?)
• scientific article; zbMATH DE number 3273293 (Why is no real title available?)
• scientific article; zbMATH DE number 3316587 (Why is no real title available?)
• scientific article; zbMATH DE number 3320868 (Why is no real title available?)
• Maximum Likelihood Estimation of Misspecified Models
• Neural networks and physical systems with emergent collective computational abilities
• Neurons with graded response have collective computational properties like those of two-state neurons
• On Information and Sufficiency
• Probability, Frequency and Reasonable Expectation
• Self-organization and associative memory
• Stimulus and response generalization: A stochastic model relating generalization to distance in psychological space
• Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images
• The brain-state-in-a-box neural model is a gradient descent algorithm
• The Covariance Matrix of the Information Matrix Test

Cited in

(6)

• Solving optimization problems with variable-constraint by an extended Cohen-Grossberg model
• A unified constructive network model for problem-solving
• Abstraction based connectionist analogy processor
• scientific article; zbMATH DE number 1164157 (Why is no real title available?)
• scientific article; zbMATH DE number 1843146 (Why is no real title available?)
• Adapting connectionist learning to Bayes networks

This page was built for publication: A unified framework for connectionist systems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1117857)