Federated Learning is a novel machine learning paradigm where training happens in a distributed way on heterogeneous clients. Recently it has been applied to computer vision problems, where it proved to be a very valuable tool to deal with the huge amount of data and with privacy issues. We considered Federated Learning settings both for image classification and semantic segmentation.
Key research topics include:
- We propose a novel semantic segmentation task in which the clients’ data is unlabeled and the server accesses a source labeled dataset for pre-training only. We tackled it with the combination of self-supervision, ad-hoc regularization techniques and a novel federated clustered aggregation scheme
- We introduce a novel federated learning setting where the continual learning of multiple tasks happens at each client with different orderings and in asynchronous time slots. We tackle this novel task using prototype-based learning, a representation loss, fractal pre-training, and a modified aggregation policy.
Recent publications:
When Cars meet Drones: Hyperbolic Federated Learning for Source-Free Domain Adaptation in Adverse Weather Proceedings Article
In: Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025.
Asynchronous Federated Continual Learning Proceedings Article
In: CVPR FedVision Workshop, 2023.
Learning Across Domains and Devices: Style-Driven Source-Free Domain Adaptation in Clustered Federated Learning Proceedings Article
In: IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) 2023, 2023.
Federated Learning in Computer Vision Journal Article
In: IEEE Access, 2023.