Abstract
Labeling image collections is a tedious task, especially when multiple labels have to be chosen for each image. In this paper we introduce a new framework that extends state of the art models in word prediction to incorporate information from unlabeled examples, using manifold regularization. To the best of our knowledge this is the first semi-supervised multi-task model used in vision problems. The new model can be solved using gradient descent and is fast and efficient. We show remarkable improvements for cases with few labeled examples for challenging multi-task learning problems in vision (predicting words for images and attributes for objects).
| Original language | English (US) |
|---|---|
| Title of host publication | 2009 IEEE 12th International Conference on Computer Vision, ICCV 2009 |
| Pages | 956-962 |
| Number of pages | 7 |
| DOIs | |
| State | Published - Dec 1 2009 |
| Event | 12th International Conference on Computer Vision, ICCV 2009 - Kyoto, Japan Duration: Sep 29 2009 → Oct 2 2009 |
Publication series
| Name | Proceedings of the IEEE International Conference on Computer Vision |
|---|
Other
| Other | 12th International Conference on Computer Vision, ICCV 2009 |
|---|---|
| Country | Japan |
| City | Kyoto |
| Period | 9/29/09 → 10/2/09 |
Fingerprint
ASJC Scopus subject areas
- Software
- Computer Vision and Pattern Recognition
Cite this
Unlabeled data improves word prediction. / Loeff, Nicolas; Farhadi, Ali; Endres, Ian; Forsyth, David Alexander.
2009 IEEE 12th International Conference on Computer Vision, ICCV 2009. 2009. p. 956-962 5459347 (Proceedings of the IEEE International Conference on Computer Vision).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Unlabeled data improves word prediction
AU - Loeff, Nicolas
AU - Farhadi, Ali
AU - Endres, Ian
AU - Forsyth, David Alexander
PY - 2009/12/1
Y1 - 2009/12/1
N2 - Labeling image collections is a tedious task, especially when multiple labels have to be chosen for each image. In this paper we introduce a new framework that extends state of the art models in word prediction to incorporate information from unlabeled examples, using manifold regularization. To the best of our knowledge this is the first semi-supervised multi-task model used in vision problems. The new model can be solved using gradient descent and is fast and efficient. We show remarkable improvements for cases with few labeled examples for challenging multi-task learning problems in vision (predicting words for images and attributes for objects).
AB - Labeling image collections is a tedious task, especially when multiple labels have to be chosen for each image. In this paper we introduce a new framework that extends state of the art models in word prediction to incorporate information from unlabeled examples, using manifold regularization. To the best of our knowledge this is the first semi-supervised multi-task model used in vision problems. The new model can be solved using gradient descent and is fast and efficient. We show remarkable improvements for cases with few labeled examples for challenging multi-task learning problems in vision (predicting words for images and attributes for objects).
UR - http://www.scopus.com/inward/record.url?scp=77953189631&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=77953189631&partnerID=8YFLogxK
U2 - 10.1109/ICCV.2009.5459347
DO - 10.1109/ICCV.2009.5459347
M3 - Conference contribution
AN - SCOPUS:77953189631
SN - 9781424444205
T3 - Proceedings of the IEEE International Conference on Computer Vision
SP - 956
EP - 962
BT - 2009 IEEE 12th International Conference on Computer Vision, ICCV 2009
ER -