2022
Caldarola, Debora; Caputo, Barbara; Ciccone, Marco
Improving generalization in federated learning by seeking flat minima Proceedings Article
In: European Conference on Computer Vision, pp. 654–672, Springer 2022.
BibTeX | Tags: European Conference on Computer Vision, Federated Learning
@inproceedings{caldarola2022improving,
title = {Improving generalization in federated learning by seeking flat minima},
author = {Debora Caldarola and Barbara Caputo and Marco Ciccone},
year = {2022},
date = {2022-01-01},
urldate = {2022-01-01},
booktitle = {European Conference on Computer Vision},
pages = {654--672},
organization = {Springer},
keywords = {European Conference on Computer Vision, Federated Learning},
pubstate = {published},
tppubtype = {inproceedings}
}
2020
D’Innocente, Antonio; Borlino, Francesco Cappio; Bucci, Silvia; Caputo, Barbara; Tommasi, Tatiana
One-Shot Unsupervised Cross-Domain Detection Proceedings Article
In: European Conference on Computer Vision — ECCV 2020, pp. 732–748, Springer International Publishing, Cham, 2020, ISBN: 978-3-030-58517-4.
Abstract | BibTeX | Tags: European Conference on Computer Vision
@inproceedings{10.1007/978-3-030-58517-4_43,
title = {One-Shot Unsupervised Cross-Domain Detection},
author = {Antonio D'Innocente and Francesco Cappio Borlino and Silvia Bucci and Barbara Caputo and Tatiana Tommasi},
isbn = {978-3-030-58517-4},
year = {2020},
date = {2020-01-01},
booktitle = {European Conference on Computer Vision -- ECCV 2020},
pages = {732--748},
publisher = {Springer International Publishing},
address = {Cham},
abstract = {Despite impressive progress in object detection over the last years, it is still an open challenge to reliably detect objects across visual domains. All current approaches access a sizable amount of target data at training time. This is a heavy assumption, as often it is not possible to anticipate the domain where a detector will be used, nor to access it in advance for data acquisition. Consider for instance the task of monitoring image feeds from social media: as every image is uploaded by a different user it belongs to a different target domain that is impossible to foresee during training. Our work addresses this setting, presenting an object detection algorithm able to perform unsupervised adaptation across domains by using only one target sample, seen at test time. We introduce a multi-task architecture that one-shot adapts to any incoming sample by iteratively solving a self-supervised task on it. We further enhance this auxiliary adaptation with cross-task pseudo-labeling. A thorough benchmark analysis against the most recent cross-domain detection methods and a detailed ablation study show the advantage of our approach.},
keywords = {European Conference on Computer Vision},
pubstate = {published},
tppubtype = {inproceedings}
}
Bucci, Silvia; Loghmani, Mohammad Reza; Tommasi, Tatiana
On the Effectiveness of Image Rotation for Open Set Domain Adaptation Proceedings Article
In: European Conference on Computer Vision, pp. 422–438, Springer 2020.
BibTeX | Tags: Domain adaptation, European Conference on Computer Vision
@inproceedings{bucci2020effectiveness,
title = {On the Effectiveness of Image Rotation for Open Set Domain Adaptation},
author = {Silvia Bucci and Mohammad Reza Loghmani and Tatiana Tommasi},
year = {2020},
date = {2020-01-01},
booktitle = {European Conference on Computer Vision},
pages = {422--438},
organization = {Springer},
keywords = {Domain adaptation, European Conference on Computer Vision},
pubstate = {published},
tppubtype = {inproceedings}
}