BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260523T010244EDT-8860CxHUBJ@132.216.98.100
DTSTAMP:20260523T050244Z
DESCRIPTION:Areas of Attention for Image Captioning.\n\nWe propose “Areas o
 f Attention”\, a novel attention-based model for automatic image captionin
 g.  Our approachmodels the dependencies between image regions\, caption wo
 rds\, and the state of an RNN language model\, using three pairwise intera
 ctions. In contrast to previous attention-based approaches that associate 
 image regions only to the RNN state\, our method allows a direct associati
 on between caption words and image regions. During training these associat
 ions are inferred from image-level captions\, akin to weakly-supervised ob
 ject detector training. These associ- ations help to improve captioning by
  localizing the corresponding regions during testing. We also propose and 
 com- pare different ways of generating attention areas: CNN activation gri
 ds\, object proposals\, and spatial transformers nets applied in a convolu
 tional fashion. Spatial transformers give the best results. They allow for
  image specific at- tention areas\, and can be trained jointly with the re
 st of the network. Our attention mechanism and spatial transformer attenti
 on areas together yield state-of-the-art results on the MSCOCO dataset.\n	
 \n	Bio: Marco Pedersoli obtained his Ph.D. from the Autonomous University O
 f Barcelona (june 2012\, with distinction and best thesis award) on Hierar
 chical Multi-resolution Detection of objects in images. He completed a pos
 tdoctoral fellowship at the KU Leuven\, where he has developed several inn
 ovative approaches for object detection\, action classification and pose e
 stimation based on weakly-supervised methods. In September 2015 he moved t
 o INRIA-Grenoble\, where he has developed new techniques for the automatic
  description and understanding of images. From February 2017 he is assista
 nt professor at the École de technologie supérieure of Montreal. He has pu
 blished more than 20 articles in peer-reviewed internationals journals and
  top peer-reviewed conferences in computer vision.\n
DTSTART:20170526T174500Z
DTEND:20170526T193000Z
LOCATION:Room 5340\, CA\, QC\, Montreal\, Universite de Montreal\, 2920\, c
 h. de la Tour
SUMMARY:Marco Pedersoli\, ETS
URL:https://www.mcgill.ca/mathstat/channels/event/marco-pedersoli-ets-26825
 9
END:VEVENT
END:VCALENDAR