Design and applications of cost-sensitive classifiers
Etienne Grossmann (Instituto Superior TA©cnico, Portugal)
NICTA SML SEMINARDATE: 2005-11-16
TIME: 09:00:00 - 10:00:00
LOCATION: RSISE Seminar Room, ground floor, building 115, cnr. North and Daley Roads, ANU
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
In this talk, I will give an overview of my research career and then focus on machine learning. My work in that field addresses some of the challenges posed by today's abundance of electronic data, sometimes labeled, but mostly unlabeled, and the desire to extract meaningful information from it. I will present my work on the design of cascades of classifiers and on decision trees that are related to the Adaboost method. Time allowing, I will also present briefly a supervised learning approach to solve a computer vision task.
BIO:
Etienne Grossmann obtained his PhD in ECE in 2002, from the Superior Technical Institute of Lisbon, Portugal, and a Bachelor in Mathematics from the University of Provence, France. Etienne is presently at the Center for Visualization and Virtual Environments of the University of Kentucky. His research has covered two main areas: the first is the part of computer vision related to geometry and estimation, for example 3D reconstruction of man-made scenes and self-calibration of "non-parametric" image sensors; the second area is machine learning, more specifically the development of generic supervised learning methods that have good generalization ability, low computation complexity and can be trained in a practical way. On this last topic he has developped variants of Adaboost that aim at these objectives.
