Zobrazeno 1 - 4
of 4
pro vyhledávání: '"Arthur Douillard"'
Publikováno v:
CVPR
CVPR, Jun 2021, Nashville, United States
CVPR, Jun 2021, Nashville, United States
Deep learning approaches are nowadays ubiquitously used to tackle computer vision tasks such as semantic segmentation, requiring large datasets and substantial computational power. Continual learning for semantic segmentation (CSS) is an emerging tre
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::455e808538af4df1c53354b3b0258d71
https://hal.archives-ouvertes.fr/hal-03503831
https://hal.archives-ouvertes.fr/hal-03503831
Publikováno v:
CVPR Workshops
CVPR Workshop
CVPR Workshop, Jun 2021, Nashville, United States
CVPR Workshop
CVPR Workshop, Jun 2021, Nashville, United States
International audience; Continual learning aims to learn tasks sequentially, with (often severe) constraints on the storage of old learning samples, without suffering from catastrophic forgetting. In this work, we propose prescient continual learning
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7a1c7643833226cb588f8ad19e3557be
https://doi.org/10.1109/cvprw53098.2021.00387
https://doi.org/10.1109/cvprw53098.2021.00387
Autor:
Arthur Douillard, Timothée LESORT
Publikováno v:
CVPR Workshop
CVPR Workshop, Jun 2021, Nashville, United States
HAL
CVPR Workshop, Jun 2021, Nashville, United States
HAL
Continual learning is a machine learning sub-field specialized in settings with non-iid data. Hence, the training data distribution is not static and drifts through time. Those drifts might cause interferences in the trained model and knowledge learn
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::95a9e9ffc4db5647f2afe66c2139c3b6
https://hal.archives-ouvertes.fr/hal-03503833
https://hal.archives-ouvertes.fr/hal-03503833
Developing deep networks that analyze fashion garments has many real-world applications. Among all fashion attributes, color is one of the most important yet challenging to detect. Existing approaches are classification-based and thus cannot go beyon
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_dedup___::7313e47082751ce1b31d3c4bc0024144
http://arxiv.org/abs/2010.02849
http://arxiv.org/abs/2010.02849