Conferencia de Alberto Tonda en la ETSIIT-UGR (20 de enero de 2015)

Dentro de la visita de investigación financiada con el premio al mejor artículo del congreso EvoStar2014 ofrecido por el Programa GENIL (Granada Excellence Network of Innovation Laboratories) asociado al CEI BioTic GRANADA, Alberto Tonda impartirá la conferencia con título «Learning Dynamical Systems with Symbolic Regression» en la ETSIIT-UGR.

Fecha: Martes, 20 de enero de 2015
Hora: 12:30
Lugar: Salón de Grados (ETSIIT)

Title: Learning Dynamical Systems with Symbolic Regression
Author: Alberto Tonda (INRA)

Abstract

Dynamical Systems, composed of Ordinary Differential Equations (ODE), are widely adopted in many fields of science and engineering, to describe phenomena ranging from fluid dynamics to diffusion of heat, from mechanical to biological processes. ODE-based models are usually built by human experts, exploiting both experimental data and theoretical knowledge of the problem at hand.

In this work, we show how to use an established machine learning technique, Symbolic Regression, to automatically derive systems of ODEs starting from data. Preliminary experimental results on the Lotka-Volterra model prove the potential of the approach.

Speaker

Alberto Tonda is a researcher at INRA (the French national institute for agronomy), where he works on semi-supervised modeling of food processes. He completed his PhD in Computer Science at Politecnico di Torino, Italy, in 2010, with a thesis on real-world applications of Evolutionary Computation.