Lecture DEI Series
Speaker: Prof. Robert Sabourin
Affiliation: U. du Québec, Canada
Dynamic ensemble selection systems work by estimating the level of competence of each classifier from a pool of classifiers. Only the most competent ones are selected to classify a given test sample. This is achieved by defining a criterion to measure the level of competence of a base classifier, such as, its accuracy in local regions of the feature space around the query instance. However, using only one criterion about the behavior of a base classifier is not sufficient to accurately estimate its level of competence. In this paper, we present a novel dynamic ensemble selection framework using metalearning. We propose five distinct sets of meta-features, each one corresponding to a different criterion to measure the level of competence of a classifier for the classification of input samples. The metafeatures are extracted from the training data and used to train a meta-classifier to predict whether or not a base classifier is competent enough to classify an input instance. During the generalization phase, the meta-features are extracted from the query instance and passed down as input to the metaclassifier. The meta-classifier estimates, whether a base classifier is competent enough to be added to the ensemble. Experiments are conducted over several small sample size classification problems, i.e., problems with a high degree of uncertainty due to the lack of training data. Experimental results show the proposed meta-learning framework greatly improves classification accuracy when compared against current state-of-the-art dynamic ensemble selection techniques.
Professor Sabourin joined the Université de Montréal Physics department in 1977, and headed the design, experimentation and development of scientific instruments for the Mont Mégantic Observatory. In 1983, he joined École de technologie supérieure (Université du Québec, Montréal), and participated in the creation of the Department of Automated Production Engineering, where he continues to serve as professor, teaching Pattern Recognition, Evolutionary Algorithms, Neural Networks and Fuzzy Systems. In 1992, he also joined the Department of Computer Science of Pontifícia Universidade Católica do Paraná (PUCPR, Curitiba, Brazil), where he participated in the creation of a Master’s program in 1995 and a Doctorate program in 1998. Since 1996, he has been a senior member of the Centre for Pattern Recognition and Machine Intelligence (CENPARMI, Concordia University). Since 2006, he has served as director of the LIVIA (Imaging, Vision and Artificial Intelligence Laboratory) at ÉTS. Professor Sabourin has authored (or co-authored) over 400 scientific publications. According to Google Scholar, his works have been cited over 9600 times (H-index of 49).
His research interests cover feature selection, the learning and selection of ensembles of classifiers, classifiers fusion, change detection and adaptive learning. The methods proposed have been used in the following applications:
This event is fully booked.