Día: miércoles 10 de octubre de 2018
Hora: 13:30
Lugar: Sala Grados A - E.T.S. Ingeniería de Telecomunicación

Abstract:

Deep models are nowadays the main actors in the area of machine learning, with convolutional neural networks having gained the highest popularity due to their image processing abilities. This talk aims to share some thoughts on what lies between the theoretical concept and the practical success of the technique. It puts forward some examples of the extensive and sometimes distressing intermediate experimentation: from choosing an appropriate architecture to parameterizing it, all while fighting the overfitting danger generated by the data scarcity of real-world problems.

Biography:

Dra. Ruxandra Stoean is Associate Professor at the Department of Computer Science, University of Craiova, Romania. Her interest in evolutionary computation and support vector machines is revealed by more than 60 published papers, two book chapters and one book published by Springer, while the impact of her research is reflected by an h-index of 8 on Thomson Reuters Web of Science. She was and is involved in several projects related to automated learning in medicine funded by national research agencies. She was awarded the prize Grigore Moisil for Computer Science by the Romanian Academy in 2008 for her research results.