Permutation binary neural networks: analysis of periodic orbits and its applications

DOI10.3934/DCDSB.2022097MaRDI QIDQ2083330zbMATH OpenOpenAlexFDO

Authors Hotaka Udagawa, Taiji Okano, T. Saito

Publication date 10 October 2022

Published in Discrete and Continuous Dynamical Systems. Series B (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2201.09219

zbMATH Keywords

binary neural networkscellular automataperiodic orbitspermutationstability

Mathematics Subject Classification ID

Dynamical aspects of cellular automata (37B15) Neural networks for/in biological studies, artificial life and related topics (92B20)

Abstract: This paper presents a permutation binary neural network characterized by local binary connection, global permutation connection, and the signum activation function. The dynamics is described by a difference equation of binary state variables. Depending on the connection, the network generates various periodic orbits of binary vectors. The binary/permutation connection brings benefits to precise analysis and to FPGA based hardware implementation. In order to consider the periodic orbits, we introduce three tools: a composition return map for visualization of the dynamics, two feature quantities for classification of periodic orbits, and an FPGA based hardware prototype for engineering applications. Using the tools, we have analyzed all the 6-dimensional networks. Typical periodic orbits are confirmed experimentally.

Recommendations

• A variety of globally stable periodic orbits in permutation binary neural networks
• On permutation symmetries of Hopfield model neural network
• Permutation parity machines for neural synchronization
• Period characteristics analysis of pulse coupled neutral network
• scientific article; zbMATH DE number 4094827

Cites work

• ``Neural computation of decisions in optimization problems
• A full computation-relevant topological dynamics classification of elementary cellular automata
• A nonlinear dynamics perspective of Wolfram's new kind of science. Vol. 1 (Part I--III).
• A simple digital spiking neural network: synchronization and spike-train approximation
• Associative dynamics in a chaotic neural network.
• Chaos in Dynamical Systems
• Chaos, fractals, and noise: Stochastic aspects of dynamics.
• Completely reproducible description of digital sound data with cellular automata
• Design of CAECC - cellular automata based error correcting code
• Neural networks and physical systems with emergent collective computational abilities
• On a coupled relaxation oscillator
• Simple mathematical models with very complicated dynamics
• Symbolic computation using cellular automata-based hyperdimensional computing

Cited in

(1)

• A variety of globally stable periodic orbits in permutation binary neural networks

This page was built for publication: Permutation binary neural networks: analysis of periodic orbits and its applications

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2083330)