Advertisement
Published in:
2020 | OriginalPaper | Chapter
5. Algorithmic Deep Learning
Authors : Michael Paluszek, Stephanie Thomas
Published in: Practical MATLAB Deep Learning
Publisher: Apress
Activate our intelligent search to find suitable subject content or patents.
Select sections of text to find matching patents with Artificial Intelligence. powered by
Select sections of text to find additional relevant content using AI-assisted search. powered by
Abstract
In this chapter, we introduce the Algorithmic Deep Learning Neural Network (ADLNN), a deep learning system that incorporates algorithmic descriptions of the processes as part of the deep learning neural network. The dynamical models provide domain knowledge. These are in the form of differential equations. The outputs of the network are both indications of failures and updates to the parameters of the models. Training can be done using simulations, prior to operations, or through operator interaction during operations.