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/

Zendy Plus

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

Zendy Tools

Zendy Open

Federated Learning (FL) plays a critical role in distributed systems. Inthese systems, data privacy and confidentiality hold paramount importance,particularly within edge-based data processing systems such as IoT devicesdeployed in smart homes. FL emerges as a privacy-enforcing sub-domain ofmachine learning that enables model training on client devices, eliminating thenecessity to share private data with a central server. While existing researchhas predominantly addressed challenges pertaining to data heterogeneity, thereremains a current gap in addressing issues such as varying device capabilitiesand efficient communication. These unaddressed issues raise a number ofimplications in resource-constrained environments. In particular, the practicalimplementation of FL-based IoT or edge systems is extremely inefficient. Inthis paper, we propose "Resource-Efficient Federated Training Framework forHeterogeneous and Resource-Constrained Environments (REFT)," a novel approachspecifically devised to address these challenges in resource-limited devices.Our proposed method uses Variable Pruning to optimize resource utilization byadapting pruning strategies to the computational capabilities of each client.Furthermore, our proposed REFT technique employs knowledge distillation tominimize the need for continuous bidirectional client-server communication.This achieves a significant reduction in communication bandwidth, therebyenhancing the overall resource efficiency. We conduct experiments for an imageclassification task, and the results demonstrate the effectiveness of ourapproach in resource-limited settings. Our technique not only preserves dataprivacy and performance standards but also accommodates heterogeneous modelarchitectures, facilitating the participation of a broader array of diverseclient devices in the training process, all while consuming minimal bandwidth.

REFT: Resource-Efficient Federated Training Framework for Heterogeneous and Resource-Constrained Environments