A new dynamical neural network architecture with hierarchical subspaces and forward-backward internetworks allows state variables to relax in warped spaces, forming mappings from periodic inputs and implying an input-output certainty/uncertainty relation.
[2]Amari, S.-I.Theory of adaptive pattern classifiers.IEEE Transactions on Electronic Computers 16(1967), 299–307
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
q-bio.NC 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Relaxing in Warped Spaces: Generalized Hierarchical and Modular Dynamical Neural Network
A new dynamical neural network architecture with hierarchical subspaces and forward-backward internetworks allows state variables to relax in warped spaces, forming mappings from periodic inputs and implying an input-output certainty/uncertainty relation.