On the loss of learning capability inside an arrangement of neural networks

Abstract

We analyze the loss of information and the loss of learning capability inside an arrangement of neural networks. Our method is new and based on the formulation of non-unitary Bogoliubov transformations in order to connect the information between different points of the arrangement. This can be done after expanding the activation function in a Fourier series and then assuming that its information is stored inside a Quantum scalar field.

arXiv comment6 pages, work presented at Neurips 2019 inside the Workshop Machine Learning and the Physical Sciences

References in corpus (5)

• Black Holes as Brains: Neural Networks with Area Law Entropy
• Area Law Micro-State Entropy from Criticality and Spherical Symmetry
• Finding Critical States of Enhanced Memory Capacity in Attractive Cold Bosons
• Bogoliubov transformations for amplitudes in black-hole evaporation
• Black-Hole evaporation from the perspective of neural networks