Skip to the content.

Call for papers

Scope of the workshop

The use of deep learning-based systems is growing fast in applications in both science and real-world scenarios. To further improve the performance of existing models, an increasing amount of data is needed, which also needs to be heterogeneous to develop better and more generalizable methods. However, huge datasets mean huge training in terms of time and computational burden. Furthermore, the collection of large-scale, diverse centralized datasets could be practically impossible, especially in some fields, e.g., medical domain, in which privacy and data ownership issues come into play.

To overcome these problems, Federated Learning (FL) has been proposed to develop better AI models without compromising users’ privacy. Different institutions can train more robust models, benefiting from the federation, without sharing raw training data. Particularly in medical fields, the heterogeneity of the data is very pronounced, both in terms of the number of samples and domain shifts among the various contributing institutions, due, for example, to the variability in equipment and image acquisition protocols.

This workshop aims at bringing together researchers and practitioners with common interest in FL for visual tasks, with a particular focus on medical imaging, to address the open questions and challenges of this research area.

Topics

The workshop aims to attract novel and original contributions exploring federated and collaborative learning with its challenges and peculiarities. Expected submissions should cover, but are not limited to, the following topics:

Workshop Proceedings

Workshop proceedings will be published by Springer LNCS.