Induction and physical theory formation as well as universal computation by machine learning

TL;DR

This paper discusses how machine learning can systematically generate formal models for physical phenomena and explore universal computation, including linear, deep networks, and their theoretical implications.

Contribution

It introduces a framework for using machine learning to form physical theories and analyze universal computation through neural network generalizations.

Findings

Linear models can approximate physical systems.

Deep networks extend modeling to nonlinear phenomena.

Machine learning can potentially realize universal computation.

Abstract

Machine learning presents a general, systematic framework for the generation of formal theoretical models for physical description and prediction. Tentatively standard linear modeling techniques are reviewed; followed by a brief discussion of generalizations to deep forward networks for approximating nonlinear phenomena and universal computers.