English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In the medical field, multi-center collaborations are often sought to yieldmore generalizable findings by leveraging the heterogeneity of patient andclinical data. However, recent privacy regulations hinder the possibility toshare data, and consequently, to come up with machine learning-based solutionsthat support diagnosis and prognosis. Federated learning (FL) aims atsidestepping this limitation by bringing AI-based solutions to data owners andonly sharing local AI models, or parts thereof, that need then to beaggregated. However, most of the existing federated learning solutions arestill at their infancy and show several shortcomings, from the lack of areliable and effective aggregation scheme able to retain the knowledge learnedlocally to weak privacy preservation as real data may be reconstructed frommodel updates. Furthermore, the majority of these approaches, especially thosedealing with medical data, relies on a centralized distributed learningstrategy that poses robustness, scalability and trust issues. In this paper wepresent a federated and decentralized learning strategy, FedER, that,exploiting experience replay and generative adversarial concepts, effectivelyintegrates features from local nodes, providing models able to generalizeacross multiple datasets while maintaining privacy. FedER is tested on twotasks -- tuberculosis and melanoma classification -- using multiple datasets inorder to simulate realistic non-i.i.d. medical data scenarios. Results showthat our approach achieves performance comparable to standard (non-federated)learning and significantly outperforms state-of-the-art federated methods intheir centralized (thus, more favourable) formulation. Code is available athttps://github.com/perceivelab/FedER

FedER: Federated Learning through Experience Replay and Privacy-Preserving Data Synthesis