📝 SummarXiv

Abstract

While it is well-established that artificial neural networks are \emph{universal approximators} for continuous functions on compact domains, many modern approaches rely on deep or overparameterized architectures that incur high computational costs. In this paper, a new type of \emph{small shallow} neural network, called the \emph{Barycentric Neural Network} ($\BNN$), is proposed, which leverages a fixed set of \emph{base points} and their \emph{barycentric coordinates} to define both its structure and its parameters. We demonstrate that our $\BNN$ enables the exact representation of \emph{continuous piecewise linear functions} ($\CPLF$s), ensuring strict continuity across segments. Since any continuous function over a compact domain can be approximated arbitrarily well by $\CPLF$s, the $\BNN$ naturally emerges as a flexible and interpretable tool for \emph{function approximation}. Beyond the use of this representation, the main contribution of the paper is the introduction of a new variant of \emph{persistent entropy}, a topological feature that is stable and scale invariant, called the \emph{length-weighted persistent entropy} ($\LWPE$), which is weighted by the lifetime of topological features. Our framework, which combines the $\BNN$ with a loss function based on our $\LWPE$, aims to provide flexible and geometrically interpretable approximations of nonlinear continuous functions in resource-constrained settings, such as those with limited base points for $\BNN$ design and few training epochs. Instead of optimizing internal weights, our approach directly \emph{optimizes the base points that define the $\BNN$}. Experimental results show that our approach achieves \emph{superior and faster approximation performance} compared to classical loss functions such as MSE, RMSE, MAE, and log-cosh.