Federated Learning over Images: Vertical Decompositions and Pre-Trained Backbones Are Difficult to Beat

Erdong Hu, Yuxin Tang, Anastasios Kyrillidis, Chris Jermaine

ICCV 2023 pp. 19385-19396

doi:10.1109/ICCV51070.2023.01776 /iccv/2023/hu2023iccv-federated/

Abstract

We carefully evaluate a number of algorithms for learning in a federated environment, and test their utility for a variety of image classification tasks. We consider many issues that have not been adequately considered before: whether learning over data sets that do not have diverse sets of images affects the results; whether to use a pre-trained feature extraction "backbone"; how to evaluate learner performance (we argue that classification accuracy is not enough), among others. Overall, across a wide variety of settings, we find that vertically decomposing a neural network seems to give the best results, and outperforms more standard reconciliation-used methods.

PDF ICCV Semantic Scholar

Cite

Text

Hu et al. "Federated Learning over Images: Vertical Decompositions and Pre-Trained Backbones Are Difficult to Beat." International Conference on Computer Vision, 2023. doi:10.1109/ICCV51070.2023.01776

Markdown

[Hu et al. "Federated Learning over Images: Vertical Decompositions and Pre-Trained Backbones Are Difficult to Beat." International Conference on Computer Vision, 2023.](https://mlanthology.org/iccv/2023/hu2023iccv-federated/) doi:10.1109/ICCV51070.2023.01776

BibTeX

@inproceedings{hu2023iccv-federated,
  title     = {{Federated Learning over Images: Vertical Decompositions and Pre-Trained Backbones Are Difficult to Beat}},
  author    = {Hu, Erdong and Tang, Yuxin and Kyrillidis, Anastasios and Jermaine, Chris},
  booktitle = {International Conference on Computer Vision},
  year      = {2023},
  pages     = {19385-19396},
  doi       = {10.1109/ICCV51070.2023.01776},
  url       = {https://mlanthology.org/iccv/2023/hu2023iccv-federated/}
}