Abstract
The breakthrough of Artificial Intelligence (AI) has revolutionized Machine Learning (ML), particularly in the form of Transformer models such as Chat Generative Pre-training Transformer (ChatGPT), achieving state-of-the-art (SoTA) performance in various do mains, including Edge Computing. This paper introduces the TANDEM Smart Resource Monitoring and Management approach, which addresses the challenges of resource monitoring and management at the edge. TANDEM leverages AI/ML mechanisms to enable efficient service distribution, user-centered operation, and diverse edge applications support. It proposes novel forms of dynamic resource monitoring and management, utilizing Trans former models for accurate system usage prediction. Furthermore, TANDEM provides an AutoML framework and its AutoTinyML extension, which enhances IoT applications with powerful ML services. The proposed architecture and approach contribute to the intersection of AutoML and Transformer models with Edge Computing. Extensive experimental evaluations and analysis demonstrate the effectiveness and potential of our approach.
