Robust Synchronization of Time-Fractional Memristive Hopfield Neural Networks

Yuncheng You · 2025

Read Robust Synchronization of Time-Fractional Memristive Hopfield Neural Networks by Yuncheng You (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2510-20949.txt.

Introduction

Yuncheng You, working alone, studies robust synchronization of time fractional Hopfield networks with memristive couplings and Hebbian learning rules. Fractional time gives the network strong memory and long range dependence on its past, closer to some biological learning than standard models.

Scaled group estimates show the dynamics stay globally dissipative under broad assumptions, so activity cannot blow up. A threshold on the coupling strength between neurons guarantees robust synchronization, and the threshold can be computed from the raw parameters. The threshold falls as the fractional order changes in a stated direction, giving designers a clear tuning rule for memory rich networks.

Catalog record

CREATORYuncheng You
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2510-20949

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Robust Synchronization of Time-Fractional Memristive Hopfield Neural Networks Yuncheng You Abstract . In this paper we study robust synchronization of time-fractional Hopfield neural networks with memristive couplings and Hebbian learning rules. This new model of artificial neural networks exhibits strong memory and long-range path-dependence in learning processes. Through scaled group estimates it is proved that under rather general assumptions the solution dynamics is globally dissipative. The main result established a threshold condition for achieving robust synchronization of the neural networks if it is satisfied by the interneuron coupling strength coefficient. The synchronizing threshold is explicitly computable in terms of the original parameters and strictly decreasing for the fractional order α ∈ ( 0 , 1 ) \alpha\in(0,1) . 2020 Mathematics Subject Classification 26A33, 34A08, 45J05, 68T07, 92B20 1. Introduction Dynamical analysis of artificial neural networks plays an essential role in deep learning and all-around AI applications. In recent two decades many biological and artificial neural network models of differential equations have been proposed and studied on topics of stability and pattern formation mainly by approaches of matrix spectrum estimation and computational simulations. As a type of recurrent neural networks, Hopfield neural networks [ 1 , 2 , 3 ] with memristive synapses [ 4 , 5 ] are extensively used for the preeminent performances in many frontier research applications [ 6 , 7 , 8 , 9 ] . For biological neural network models in terms of diffusive Hindmarsh-Rose equations [ 10 ] , diffusive FitzHugh-Nagumo equations [ 11 ] , and general reaction-diffusion equations [ 12 ] with memristive synapses, the exponential synchronization results have been proved by the author’s group. For the memristive Hopfield neural networks, very recently it was shown [ 13 ] that only approximate synchronization can be achieved instead of complete synchronization, due to mismatched weight dynamics and heterogeneous activation functions among the neurons. In recent two decades, along with fractional-order derivatives and related calculus, researches on time-fractional differential equations have been booming in modeling dynamics of physical and interdisciplinary processes as well as Hopfield neural networks [ 14 , 15 ] , ranging from viscoelastic liquids [ 16 ] , electrochemical processes [ 17 ] , dielectric polarization [ 18 ] , image processing and encryption [ 19 ] , mathematical finance [ 20 , 21 ] , to machine learning [ 22 ] . The fractional differential equation models have the advantage of continuum long-term memory properties in comparison with classical integer-order differential equations with discrete time-delay terms. From the biosimulation point of view, fractional-order formulation of artificial neural networks can be analogously justified by some reported researches on biological neural systems, such as the oculomotor integrator [ 23 ] , which is a neural network in the brainstem that quickly converts a temporary eye velocity signal to a persistent eye position command of fractional order less than one. Another evidence is that fractional-order differential model can be used to efficiently emulate stimulus anticipation and oscillatory neuron firing dynamics [ 24 ] , which is a cognitive preparatory measure triggered by the emotional potentials called SPN (stimulus preceding negativity). In the sense fractional model dynamics is an important aspect for the frontier BMI (brain-machine interface) technology. Synchronization dynamics of fractional-order memristive Hopfield neural networks is an important topic, which is more extensive and complicated than the classical stability properties converging limited to equilibrium states [ 25 , 26 ] , because it also includes the chaotic synchronization and the approximate synchronization. In this paper, we propose a new model of Caputo type time-fractional memristive Hopfield neural networks with Hebbian learning rule. It can also be called fractional memristive Hopfield-Hebbian neural networks (briefly fmHHNN). We shall investigate robust synchronization of the solution trajectories for his model. Consider a fractional-order memristive Hopfield neural network composed of m neuron nodes and denoted by \mathcal{NW}=\{\mathcal{N}_{i}:i=1,2,\cdots,m\} . The state variable u_{i}(t) for each neuron \mathcal{N}_{i},1\leq i\leq m , the symmetric weights w_{ij}(t) updated by the Hebbian learning rule [ 27 , 28 ] before synaptic activation functions f_{j}(u_{j}) , and the memristor’s window function \rho(t) are governed by fractional differential equations of the same fractional order \alpha in this model: (1.1) \begin{split}&D_{c}^{\alpha}u_{i}(t)=-a_{i}u_{i}+\sum_{j=1}^{m}w_{ij}\,f_{j}(u_{j})+k\,\psi_{i}(\rho)u_{i}+J_{i}-P\sum_{j=1}^{m}(u_{i}-u_{j}),\;\;1\leq i\leq m,\\ &D_{c}^{\alpha}w_{ij}(t)=-c_{ij}\,w_{ij}+\lambda_{ij}\,f_{i}(u_{i})f_{j}(u_{j}),\quad 1\leq i,j\leq m,\\ &D_{c}^{\alpha}\rho(t)=\sum_{i=1}^{m}\gamma_{i}u_{i}-b\rho,\quad\text{where}\quad t>0,\quad\alpha\in(0,1).\end{split} In this model, the Caputo type fractional time derivative of order \alpha\in(0,1) denoted by D_{c}^{\alpha} of a scalar function y(t) is defined by D_{c}^{\alpha}y(t)=\frac{1}{\Gamma(1-\alpha)}\int_{0}^{t}(t-s)^{-\alpha}\,y^{\prime}(s)\,ds,\quad t>0, and \Gamma(z)=\int_{0}^{\infty}t^{z-1}e^{-t}\,dt,\,z>0, is the Gamma function, for y(\cdot)\in C[0,\infty) and y^{\prime}(\cdot)\in L^{1}_{loc}[0,\infty) . The coefficient P>0 in ( 1.1 ) is called the interneuron coupling strength. Note that the state variable u_{i}(t) simulates the electric potential of a biological neuron. The initial states of the system ( 1.1 ) will be denoted by (1.2) u_{i}^{0}=u_{i}(0),\quad w_{ij}^{0}=w_{ij}(0),\quad\rho^{0}=\rho(0),\quad 1\leq i,j\leq m. Here w^{0}_{ij}=w^{0}_{ji} for any i,j . The memristive feature of this model is reflected by the memductance-potential synapse term k\psi_{i}(\rho)u_{i} for each neuron \mathcal{N}_{i} . The Hebbian learning rule used in ( 1.1 ) for synaptic weights w_{ij}(t) demonstrated good performance in unsupervised machine learning tasks on certain datasets like MNIST datasets and CIFAR datasets in image recognition and classification [ 29 ] . We assume that the scalar functions f_{i}(s),1\leq i\leq m are locally Lipschitz continuous functions and the memristor window’s functions are \psi_{i}(s)=s(\eta_{i}-s),1\leq i\leq m . The parameters a_{i},b,k,\eta_{i},c_{ij} can be any given positive constants, while the input/bias constants J_{i} and coefficients \lambda_{ij},\gamma_{i} can be any given real numbers. More specific assumptions on parameters and element functions will be made in Section 3. In this paper we shall define robust synchronization for artificial neural network models and investigate this new model ( 1.1 ) of fractional-order memristive Hopfield neural networks to rigorously prove that robust synchronization of this fmHHNN will be triggered if the network coupling strength coefficient P satisfies a computable threshold condition. The rest of the paper is organized as follows. In Section 2 we recall preliminary basics of fractional calculus and time-fractional differential equations. In particular we shall prove three useful tool Lemmas, especially a new form of fractional Gronwall inequality. A definition of robust synchronization will then be introduced. In Section 3 we show the dissipative dynamics of global solutions to the initial value problem ( 1.1 )-( 1.2 ) in terms of the existence of an absorbing set. In Section 4 we shall prove the main result on robust synchronization of the fractional memristive Hopfield-Hebbian neural networks by the approach of direct uniform estimates of the interneuron differencing equations. Conclusions will be in the end. 2. Fractional Analysis Lemmas Fractional-order derivatives and fractional calculus broadly expanded the mathematical fields in modeling physical processes and phenomenological dynamics by fractional differential equations (FDE). The framework of FDE provides a flexible incorporation of longterm memory of continuum past states in the trajectories of real time evolution. The basic theory and some applications of fractional calculus and FDE can be found in books [ 30 , 31 , 32 ] . Caputo fractional derivative D_{c}^{\alpha}y(t) defined in Section 1 is strictly called the left-sided Djrbashian-Caputo fractional derivative [ 32 , 33 ] , usually for absolutely continuous functions y\in AC[0,T) which annihilated the singularity at the initial point t=0 . Compare with Riemann-Liouville fractional derivative and Grünwald-Letnikov fractional derivative, Caputo fractional

Read the complete text