Splines, Rational Functions and Neural Networks
نویسندگان
چکیده
Connections between spline approximation, approximation with rational functions, and feedforward neural networks are studied. The potential improvement in the degree of approximation in going from single to two hidden layer networks is examined. Some results of Birman and Solomjak regarding the degree of approximation achievable when knot positions are chosen on the basis of the probability distribution of examples rather than the function values are extended.
منابع مشابه
Efficient Approximation with Neural Networks: A Comparison of Gate Functions
We compare different gate functions in terms of the approximation power of their circuits. Evaluation criteria are circuit size s, circuit depth d and the approximation error e(s, d). Informally, gate functions γ1 and γ2 are called equivalent if {γ1}-circuits of size s and depth d can be approximated by {γ2}-circuits (and vice versa) of size poly(s), depth O(d) with approximation error e(s, d) ...
متن کاملNon Uniform Rational B Spline (NURBS) Based Non-Linear Analysis of Straight Beams with Mixed Formulations
Displacement finite element models of various beam theories have been developed traditionally using conventional finite element basis functions (i.e., cubic Hermite, equi-spaced Lagrange interpolation functions, or spectral/hp Legendre functions). Various finite element models of beams differ from each other in the choice of the interpolation functions used for the transverse deflection w, tota...
متن کاملUniform Approximation by Neural Networks Activated by First and Second Order Ridge Splines
We establish sup-norm error bounds for functions that are approximated by linear combinations of first and second order ridge splines and show that these bounds are near-optimal.
متن کاملIsogeometric-Based Modeling and Analysis of Laminated Composite Plates Under Transverse Loading
Analysis of the laminated composite plates under transverse loading is considered using the new method of Isogeometric Analysis (IGA). Non-Uniform Rational B-Splines(NURBS)are used as shape functions for modeling the geometry of the structure and also are used as shape functions in the analysis process. To show robustness of the new technique, some examples are represented and are compared with...
متن کاملRational Parametrizations of Neural Networks
A connection is drawn between rational functions, the realization theory of dynamical systems, and feedforward neural networks. This allows us to parametrize single hidden layer scalar neural networks with (almost) arbitrary analytic activation functions in terms of strictly proper rational functions. Hence, we can solve the uniqueness of parametrization problem for such networks.
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1991