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

Generative Adversarial Networks (GANs) are typically trained to synthesizedata, from images and more recently tabular data, under the assumption ofdirectly accessible training data. Recently, federated learning (FL) is anemerging paradigm that features decentralized learning on client's local datawith a privacy-preserving capability. And, while learning GANs to synthesizeimages on FL systems has just been demonstrated, it is unknown if GANs fortabular data can be learned from decentralized data sources. Moreover, itremains unclear which distributed architecture suits them best. Different fromimage GANs, state-of-the-art tabular GANs require prior knowledge on the datadistribution of each (discrete and continuous) column to agree on a commonencoding -- risking privacy guarantees. In this paper, we propose Fed-TGAN, thefirst Federated learning framework for Tabular GANs. To effectively learn acomplex tabular GAN on non-identical participants, Fed-TGAN designs two novelfeatures: (i) a privacy-preserving multi-source feature encoding for modelinitialization; and (ii) table similarity aware weighting strategies toaggregate local models for countering data skew. We extensively evaluate theproposed Fed-TGAN against variants of decentralized learning architectures onfour widely used datasets. Results show that Fed-TGAN accelerates training timeper epoch up to 200% compared to the alternative architectures, for both IIDand Non-IID data. Overall, Fed-TGAN not only stabilizes the training loss, butalso achieves better similarity between generated and original data.

Fed-TGAN: Federated Learning Framework for Synthesizing Tabular Data