Profiling AI Models: Towards Efficient Computation Offloading in Heterogeneous Edge AI Systems
This work addresses the problem of resource management for Edge AI systems, crucial for future 6G networks, but it appears incremental as it builds on existing profiling concepts.
The paper tackles the challenge of inefficient computation offloading in heterogeneous Edge AI systems due to limited resources and unrealistic assumptions of homogeneity, proposing a profiling approach for AI models that shows promise in optimizing resource allocation and enhancing performance based on initial experiments with over 3,000 runs.
The rapid growth of end-user AI applications, such as computer vision and generative AI, has led to immense data and processing demands often exceeding user devices' capabilities. Edge AI addresses this by offloading computation to the network edge, crucial for future services in 6G networks. However, it faces challenges such as limited resources during simultaneous offloads and the unrealistic assumption of homogeneous system architecture. To address these, we propose a research roadmap focused on profiling AI models, capturing data about model types, hyperparameters, and underlying hardware to predict resource utilisation and task completion time. Initial experiments with over 3,000 runs show promise in optimising resource allocation and enhancing Edge AI performance.