Can Artificial Neural Networks be a Source to Derive the Newton's Second Law?

Nieto-Chaupis H. · Proceedings - 2025 8th Asia Conference on Cognitive Engineering and Intelligent Interaction, CEII 2025 (2025) · DOI: 10.1109/ceii67787.2025.00075

Resumen

In classical physics, the 2nd Newton's law is considered the angular stone from which all classical mechanics is derived in either concepts and equations. This paper explores an unusual path: to derive this 2nd Newton's law from the usage of a multilayer perceptron. It implies that law would be a kind of output from a set of inputs. Therefore, the central task is to investigate the more adequate family of inputs that would construct from the mathematical point of view the 2nd Newton's law. The present paper, proposes the physics scenario contemplated inputs as the most general polynomial structures that form the electric potential. Indeed, weight is associated to gradient operator instead of being a number of functions. Finally, the activation function is given by the nullity of sum. The activation more than to be a simple operator it plays the role as necessary condition to yield the output.

Instituciones: Universidad Autónoma del Perú