The Internet of Intelligent Things (IoIT) represents a significant evolution in combining Internet of Things devices with embedded AI algorithms. However, this emerging field faces critical challenges such as energy supply and computing resource constraints. Notably, existing research primarily focuses on centralized systems, which can become bottlenecks and introduce security risks. In contrast, a decentralized system allows devices to self-organize and make independent decisions, enhancing efficiency and responsiveness.
To overcome current limitations, a new heterogeneous, decentralized peer-to-peer mesh network model is proposed. This innovative approach utilizes federated learning for distributed training across nodes, enabling optimal task allocation and efficient routing paths. Additionally, employing multi-objective optimization helps to achieve a balance between conflicting performance goals such as reliability, energy efficiency, and latency. This model has the potential to significantly improve the deployment of machine learning and deep learning models in resource-constrained environments, thus impacting the future of IoIT.
👉 Pročitaj original: arXiv AI Papers
