DistrEE: Distributed Early Exit of Deep Neural Network Inference on Edge Devices
This work addresses the problem of resource-efficient AI inference for edge computing applications like autonomous vehicles and smart homes, but it is incremental as it combines existing techniques of early exit and distributed inference.
The paper tackles the challenge of achieving accurate and efficient distributed deep neural network inference on edge devices by proposing DistrEE, a framework that integrates early exit and distributed inference to trade off latency and accuracy, with simulation results showing it enables efficient collaborative inference.
Distributed DNN inference is becoming increasingly important as the demand for intelligent services at the network edge grows. By leveraging the power of distributed computing, edge devices can perform complicated and resource-hungry inference tasks previously only possible on powerful servers, enabling new applications in areas such as autonomous vehicles, industrial automation, and smart homes. However, it is challenging to achieve accurate and efficient distributed edge inference due to the fluctuating nature of the actual resources of the devices and the processing difficulty of the input data. In this work, we propose DistrEE, a distributed DNN inference framework that can exit model inference early to meet specific quality of service requirements. In particular, the framework firstly integrates model early exit and distributed inference for multi-node collaborative inferencing scenarios. Furthermore, it designs an early exit policy to control when the model inference terminates. Extensive simulation results demonstrate that DistrEE can efficiently realize efficient collaborative inference, achieving an effective trade-off between inference latency and accuracy.