Temporal association
[For the entire collection see Zbl 0745.00005.] The paper reviews main approaches to storage, retrieval and recognition of pattern sequences in Hopfield-like neural networks. The review, extended by some new results, is not complete (and authors declare that), yet it gives many important ideas on how the time-dimension can be introduced into models of associative memories. Several dynamics, evoking successively various patterns, and stabilizing the stored ones (for a moment necessary to ``read them) in mixed transition phases are discussed. The mechanisms include asymmetric couplings, time delayed transmissions, clusters of neurons exciting each other and endowed by bundles of intercluster synapses preventing recall of undesirable states, their noise-driven analogies etc. Questions of persistence time spent on processing of each pattern in the sequence, use of fast synaptic modifications for stabilization, selection of the next state in bifurcation points (as in ``complex sequence generation: which letter follows ``e?!), memory spans needed for proper selections, etc. are discussed as well. Most of suggested architectures do not enable a true learning; the last paragraph on Hebbian learning with delays does. Various temporal relations between the patterns of a learned cycle then represent the whole cycle as a dynamic object. The list of references includes 67 papers, last ones from 1990.
- Statistical mechanics of temporal association in neural networks with transmission delays
- Networks of delay differential equations and associative memories
- Learning temporal sequences by local synaptic changes
- Competition between pattern recall and sequence processing in a neural network storing correlated patterns
- Perceptron-like learning in time-summating neural networks
- W-neural networks in the problem of associative memory
- Neural networks handling sequential patterns.
- Networks of delay differential equations and associative memories
- Chaotic neural network with time delay term for sequential patterns
- A neural network for storing individual patterns in limit cycles
- scientific article; zbMATH DE number 67858 (Why is no real title available?)
- Perceptron-like learning in time-summating neural networks
- Learning temporal sequences by local synaptic changes
- Complex temporal association in neural networks
- scientific article; zbMATH DE number 978202 (Why is no real title available?)
- scientific article; zbMATH DE number 2033130 (Why is no real title available?)
- scientific article; zbMATH DE number 2033137 (Why is no real title available?)
- scientific article; zbMATH DE number 2044706 (Why is no real title available?)
- Statistical mechanics of temporal association in neural networks with transmission delays
- Sequential memory: binding dynamics
- Competition between pattern recall and sequence processing in a neural network storing correlated patterns
- Temporal association in a network of neuronal oscillators
- Using conceptors to manage neural long-term memories for temporal patterns
- How to compress sequential memory patterns into periodic oscillations: general reduction rules
- Time for retrieval in recurrent associative memories
This page was built for publication: Temporal association
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4027096)
