New article: Prediction and identification of physical systems by means of Physically-Guided Neural Networks with meaningful internal layers

We are pleased to present our latest work titled: “Prediction and identification of physical systems by means of Physically-Guided Neural Networks with meaningful internal layers.” It was finally published yesterday, 13 of April 2021.

It´s a great work by Jacobo Ayensa-Jiménez, Mohamed. H. Doweidar, José A. Sanz-Herrera, Manuel Doblaré, where the concept of a physics-guided neural network is presented.
These networks impose the physics of the problem on the inner layers’ neurons to discover state equations between variables. Thus, we provide artificial intelligence with predictive and explanatory capacity and open the black boxes of the NRs. By concentrating the network’s predictive power on the unknown part of the problem, guided networks learn faster, with less data, filter better, and improve predictions in extreme situations.

You can check it out here: https://lnkd.in/djdywP2