Optimization and Federated Learning for Edge Computing with Resource Constraints

Explore cutting-edge advancements in resource optimization and federated learning for edge computing environments with limited resources in this 29-minute keynote presentation. Learn how to implement machine learning-based solutions for optimal resource allocation using Coupled Long Short-Term Memory (CLSTM) networks, which significantly reduce computational time while maintaining near-optimal results. Discover innovative federated learning techniques including model pruning and split learning designed to enhance learning performance under bandwidth and computational constraints—critical challenges in modern wireless and mobile edge networks. Master resource allocation through CLSTM networks and understand how to solve optimization problems using machine learning with real-world efficiency. Examine federated learning across decentralized systems, model pruning to reduce communication overhead, and split learning for balancing computation and communication. Apply these concepts to network management, edge AI, and privacy-preserving learning scenarios. Review real-world experiments and numerical results that demonstrate superior performance over conventional methods, providing practical insights for implementing these advanced techniques in resource-constrained edge computing environments.