Parameter Estimation of Sigmoid Superpositions: Dynamical System Approach

DOI10.1162/089976603322362428MaRDI QIDQ4461330zbMATH OpenOpenAlexWikidataFDO

Authors C. van Leeuwen, Danil Prokhorov, I. Tyukin

Publication date 30 March 2004

Published in Neural Computation (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0207075

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Abstract: Superposition of sigmoid function over a finite time interval is shown to be equivalent to the linear combination of the solutions of a linearly parameterized system of logistic differential equations. Due to the linearity with respect to the parameters of the system, it is possible to design an effective procedure for parameter adjustment. Stability properties of this procedure are analyzed. Strategies shown in earlier studies to facilitate learning such as randomization of a learning sequence and adding specially designed disturbances during the learning phase are requirements for guaranteeing convergence in the learning scheme proposed.

Recommendations

• A method of optimal estimation of parameters of a logistic curve.
• scientific article; zbMATH DE number 4007694
• On sigmoidal functions
• scientific article; zbMATH DE number 4050810
• Maximum likelihood parameter estimation of superimposed signals by dynamic programming

Cites work

• A Global Optimum Approach for One-Layer Neural Networks
• A new adaptive learning rule
• Absolute stability of global pattern formation and parallel memory storage by competitive neural networks
• Adaptive observers for single output nonlinear systems
• Adaptive observers with arbitrary exponential rate of convergence for nonlinear systems
• Adaptive observers with exponential rate of convergence
• Approximation by superpositions of a sigmoidal function
• Degree of Approximation Results for Feedforward Networks Approximating Unknown Mappings and Their Derivatives
• Global adaptive output-feedback control of nonlinear systems. I. Linear parameterization
• On continuous-time parameter identification by using observers
• Optimization by simulated annealing
• Parameter identification for uncertain linear systems with partial state measurements under an H/sup ∞/ criterion
• Parameter identification for uncertain plants using \(H^ \infty\) methods
• Stable adaptive observers for nonlinear time-varying systems
• The unreasonable effectiveness of neural network approximation
• Universal approximation bounds for superpositions of a sigmoidal function

Cited in

(3)

• Knowledge-informed neuro-integrators for aggregation kinetics
• Physics-informed neural networks for aggregation kinetics
• Free energy, value, and attractors

This page was built for publication: Parameter Estimation of Sigmoid Superpositions: Dynamical System Approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4461330)