[AllUsers-ISR] Palestra Prof. Christoph Lampert (31 de Março 2016, 16h, Anf. ISR)

[Jorge Batista]

Caros Colegas, Investigadores e Alunos,

É com prazer que os convido a assistir à palestra "Classifier Adaptation at Prediction Time" que o Prof. Christoph Lampert irá realizar pelas 16h00 do dia 31 de Março de 2016 no anfiteatro do Instituto de Sistemas e Robótica, Polo de Coimbra.
O Prof. Christoph Lampert é docente no "Institute of Science and Technology-Austria" e coordenador do grupo de Visão por Computer e Reconhecimento de Padrões no mesmo Instituto.

PALESTRA : CLASSIFIER ADAPTATION AT PREDICTION TIME
ORADOR : PROF. CHRISTOPH LAMPERT (IST-Austria)
Local: Anf. do Instituto de Sistemas e Robótica, piso 0, DEEC
Hora : 16h00

Com os melhores cumprimentos

Jorge Batista

Jorge Batista, PhD
Associate Professor w/ tenure
DEEC-FCTUC
University of Coimbra
Coimbra, PORTUGAL

------------------------------------------------------------------------------------------------

Prof. Christoph Lampert (Short-CV)

Christoph Lampert received the PhD degree in mathematics from the
University of Bonn in 2003. In 2010 he joined the Institute of Science
and Technology Austria (IST Austria) first as an Assistant Professor and
since 2015 as a Professor. His research on computer vision and machine
learning won several international and national awards, including the
best paper prize of CVPR 2008. In 2012 he was awarded an ERC Starting
Grant by the European Research Council. He is an Editor of the
International Journal of Computer Vision (IJCV), Action Editor of the
Journal for Machine Learning Research (JMLR), and Associate Editor in
Chief of the IEEE Transactions on Pattern Analysis and Machine
Intelligence (TPAMI).


"Classifier Adaptation at Prediction Time"

In the era of "big data" and a large commercial interest in computer
vision, it is only a matter of time until we will buy commercial object
recognition systems in pre-trained form instead of training them
ourselves. This, however, poses a problem of domain adaptation: the data
distribution in which a customer plans to use the system will almost
certainly differ from the data distribution that the vendor used during
training. Two relevant effects are a change of the class ratios and the
fact that the image sequences that needs to be classified in real
applications are typically not i.i.d. In my talk I will introduce simple
probabilistic technique that can adapt the object recognition system to
the test time distribution without having to change the underlying
pre-trained classifiers. I will also introduce a framework for creating
realistically distributed image sequences that offer a way to benchmark
such adaptive recognition systems. Our results show that the above
"problem" of domain adaptation can actually be a blessing in disguise:
with proper adaptation the error rates on realistic image sequences are
typically lower than on standard i.i.d. test sets.