ICIAP 2015 Best Paper Award goes to Ilja Kuzborskij, Idiap/EPFL Ph.D. student, Francesco Orabona, and Barbara Caputo

International Conference on Image Analysis and Processing (ICIAP) 2015 was held in Genova, Italy, on 7-11 September. ICIAP 2015 is the 18th edition of a series of conferences organized biennially by the Italian Member Society (GIRPR) of the International Association for Pattern Recognition (IAPR). The conference covers both classic and most recent trends in computer vision, pattern recognition and image processing, addressing both theoretical and applicative aspects.

GIRPR, the Italian Group of Researchers affiliated to IAPR, upon recommendation of the Award Committee, conferred the award during the conference, in recognition of the originality of the research and for the significant contribution to the field.

The paper addresses problem of learning from very few examples by leveraging upon the vast amount of source (prior) knowledge in the form of pre-trained hypotheses. In particular, authors propose a simple and efficient way to select and combine source subset that improves performance on the target task. State-of-art empirical findings are reinforced by strong theoretical results.

To download the paper, click on the following link: Transfer Learning through Greedy Subset Selection

More information about the award can be found at: http://www.iciap2015.eu/program/awards.html

Websites of the authors: Ilja Kuzborskij, Francesco Orabona, Barbara Caputo.