INESC TEC researcher awarded in two conferences

INESC TEC researcher awarded in two conferences

Carlos Soares, a researcher at the Manufacturing Systems Engineering Unit (UESP) at INESC TEC, received a Best Paper Award for two articles in which he describes the application of metalearning techniques to two different tasks: a priori selection of parameters for Support Vector Machines and selection of models for continuous flows of data (data streams).

04th December 2012

The paper "Meta-learning Algorithm for Periodic Selection in Time-changing Data" was presented during the Brazilian Symposium on Neural Networks 2012 (SBRN), which took place between 20 and 25 October in Brazil. The article was written in co-authorship with two other researchers.

The other article, entitled "Combining Multi-Objective Optimization Approach with Meta-Learning for SVM Parameter Selection", was also co-written with three researchers, and it was awarded at the International Conferences on Systems, Man and Cybernetics Society (IEEE SMC) 2012, which took place between 14 and 17 October in Seoul, South Korea.
"The Metalearning consists of using machine learning techniques to model the behavior of machine learning algorithms (hence the 'meta')", explains Carlos Soares.

BIP, November 2012