About

We live in a data-driven era where AI and ML are integrated into every aspect of life and industry when making decisions. Recent AI/ML applications, scenarios, and use cases data sources come from large-scale distributed and diverse sources, i.e., in terms of capacity and data heterogeneity. Such an approach empowers applications to discover unique insights, which can be intelligently utilized to provide better services and user experience. Yet it imposes serious debate on data and client privacy, specifically on data protection regulations and restrictions such as EU GDPR. Moreover, collecting, aggregating, and integrating heterogeneous data dispersed over various data sources and securely managing and processing the data are non-trivial tasks. The challenges are not only due to transporting high-volume, high-velocity, high-veracity, cybersecurity attacks, and heterogeneous data across organizations. There is also a challenge with domain-specific language models to get enough training data since it is usually private or sensitive, with complicated administrative procedures surrounding it. Such private data include users’ financial transactions, patients’ health data, or camera footage on the street.

In this context, Federated learning (FL) has emerged as a prospective solution that facilitates distributed collaborative learning without disclosing original training data. The idea behind FL is to train the ML model collaboratively among distributed actors without sharing their data and violating the privacy accord. FL locates ML services and operations closer to the clients, facilitating leveraging available resources on the network’s edge. Hence, FL has become a critical enabling technology for future intelligent applications in domains such as autonomous driving, smart manufacturing, and healthcare. This development will lead to an overall advancement of FL and its impact on the community, noting that FL has gained significant attention within the machine learning community in recent years.

The FLTA2023 aims to provide a global forum for disseminating the latest scientific research and industry results in all aspects of federated learning. FLTA2023 also aims to bring together researchers, practitioners, and edge intelligence advocators in sharing and presenting their perspectives on the effective management of FL deployment architectures. The symposium will address the theoretical foundations of the field, as well as applications, datasets, benchmarking, software, hardware, and systems. Also, to create an annual forum for researchers and practitioners who share an interest in FL. FLTA offers an opportunity to showcase the latest advances in this area and discuss and identify future directions and challenges in FL systems. FLTA2023 will also provide ample opportunities for networking, sharing knowledge, and collaborating with others in the metaverse community.

Topics of interest:

• Federated Learning frameworks
• Federated Learning Aggregation Algorithms
• Federated Learning Applications
• Federated Learning Deployment Architectures
• Privacy-Preserving FL Techniques
• Federated Learning Communication-efficiency
• Federated Learning modelling and simulation tools
• Federated Learning datasets and benchmarking
• Federated Learning Associated Technologies