The centralized training of deep learning models not only incurs high communication cost of data transfer into the cloud systems but it also raises the privacy-protection concerns of data providers. Federated learning is the new approach whereby centralized deep learning models are distributed across the devices and systems near the data sources which perform initial training and report the updated model attributes to the centralized cloud servers. These centralized servers perform secure and privacy preserving attributes-aggregation and update the global learning model. Federated learning ensures communication efficiency because the raw data never leaves the premises of data providers. In addition, secure federated aggregation benefits in terms of privacy-preservation where neither the centralized servers nor other data providers in the federated learning systems can discriminate the actual data. Still, federated learning systems face key challenges that may hinder their massive adoption. These challenges include privacy, latency, decentralization, asynchronization, personalization, fairness, and bandwidth-optimization, to name a few. This book aims at congregating researchers and practitioners to share their research in showing how federated learning can transform next-generation artificial intelligence applications, and propose solutions to address key federated learning challenges. We are interested in both survey and original works in unexplored and/or emerging topics in the broad area of federated learning systems, architectures, applications, and algorithms, and in novel findings and/or new insights that build on existing works. Our topics of interest include but not limited to:
Differential Privacy Techniques
Latency-minimal Federated Learning Applications
Bandwidth-Optimization Techniques for Efficient Data Communication
Local and Global Model Personalization
Decentralized Model Training
Fine-grained Federated Learning
Incentive Mechanisms for Large-scale Data Providers
Trust Models in Federated Learning Systems
Reputation Models in Federated Learning Systems
Active Monitoring for Secure and Quality Model Aggregation
Heterogeneity-Awareness Across Federated Learning Systems
Context-Awareness for Data Collection, Model Training, and Aggregation
Model Compression
Adaptive Model Aggregation
Fairness (Algorithmic, Systematic)