Mixed states on neural network with structural learning

DOI10.1016/S0893-6080(03)00137-0MaRDI QIDQ1887128zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Tomoyuki Kimoto, Masato Okada

Publication date 23 November 2004

Published in Neural Networks (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/cond-mat/0206559

zbMATH Keywords

Associative memoryMixed stateObject recognitionSparse codingStructural learningTransform invariance

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Abstract: We investigated the properties of mixed states in a sparsely encoded associative memory model with a structural learning method. When mixed states are made of s memory patterns, s types of mixed states, which become equilibrium states of the model, can be generated. To investigate the properties of s types of the mixed states, we analyzed them using the statistical mechanical method. We found that the storage capacity of the memory pattern and the storage capacity of only a particular mixed state diverge at the sparse limit. We also found that the threshold value needed to recall the memory pattern is nearly equal to the threshold value needed to recall the particular mixed state. This means that the memory pattern and the particular mixed state can be made to easily coexist at the sparse limit. The properties of the model obtained by the analysis are also useful for constructing a transform-invariant recognition model.

Recommendations

• Bistability of mixed states in a neural network storing hierarchical patterns
• Vector associative memory models
• Coexistence of memory patterns and mixed states in a sparsely encoded associative memory model storing ultrametric patterns
• A comparative study of sparse associative memories
• Storage capacities of sparse associative memories

Cites work

• Bistability of mixed states in a neural network storing hierarchical patterns
• Coexistence of uncorrelated and correlated attractors in a nonmonotonic neural network
• Neural theory of association and concept-formation
• Notions of associative memory and sparse coding
• Self-consistent signal-to-noise analysis and its application to analogue neural networks with asymmetric connections

Cited in

(2)

• Coexistence of memory patterns and mixed states in a sparsely encoded associative memory model storing ultrametric patterns
• State-Dependent Parallel Gaussian Networks With a Common State-Cognitive Helper

This page was built for publication: Mixed states on neural network with structural learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1887128)