UPS researcher received the Best Paper Award

Cuenca, miércoles 6 septiembre 2017
Diego Cabrera presenting his paper in an international seminar
Diego Cabrera presenting his paper in an international seminar

 

The paper "Combining reservoir computing and variational inference for efficient one-class learning on dynamical systems" written by researchers Diego Cabrera (UPS), Fernando Sancho (U. de Sevilla) and Felipe Tobar (U.de Chile) was awarded the Best Paper Award during the International Conference on Sensing, Diagnostics, Prognostics and Control 2017 (SDPC 2017) held in Shanghai, China. 

The paper proposes a hybrid model of reservoir computing and deep learning for the learning of complex probability distributions of time series for the modeling of specific states of a dynamic system. "The proposal can be implemented in any national or foreign industry that has rotating machines (at the moment) as an essential part of its production process," explained Cabrera.

At the moment, all the important companies worldwide have this technology, and a continuous monitoring of their condition is required without having to stop production. "Our research proposes to detect possible machine failures through online measurements and avoid significant costs to the company in corrective maintenance, which could affect the cost of the final product. Continuous monitoring of machines can also prevent accidents that can even cause human losses, "said Cabrera.

The research is the result of the joint work of the Research and Development Group on Industrial Technologies (Grupo de Investigación y Desarrollo en Tecnologías Industriales (GIDTEC) of UPS's branch campus in Cuenca, the Department of Computer Science and Artificial Intelligence (CCIA) of the University of Seville and the Mathematical Modelling Center from Universidad de Chile. 

Over 300 scientists and professors from 16 countries from 5 continents attended the SDPC 2017, which is indexed to the IEEEXPLORE database. 258 papers were presented, 148 were approved by the evaluating committee and finally one was chosen as the best paper.

The winning paper will be included in a group of 40 recommended papers for the Chinese Journal of Aeronautics - Elsevier and the Journal of Intelligent and Fuzzy systems.  

 

 

 

La investigación es fruto del trabajo conjunto del Grupo de Investigación y Desarrollo en Tecnologías Industriales (GIDTEC) de la Universidad Politécnica Salesiana sede Cuenca, el Departamento de Ciencias de la Computación e Inteligencia Artificial (CCIA) de la Universidad de Sevilla y el Centro para el Modelamiento Matemático (CMM) de la Universidad de Chile.