Adaptive Selection of Deep Learning Models on Embedded Systems

TL;DR

This paper introduces an adaptive method that uses machine learning to select the most suitable deep neural network model for image classification on embedded devices, balancing accuracy and inference time.

Contribution

It presents a novel adaptive scheme that predicts and selects the optimal DNN model for each input on embedded systems, improving efficiency and accuracy.

Findings

Achieved 7.52% higher inference accuracy

Reduced inference time by 1.8x

Validated on Jetson TX2 with ImageNet dataset

Abstract

The recent ground-breaking advances in deep learning networks ( DNNs ) make them attractive for embedded systems. However, it can take a long time for DNNs to make an inference on resource-limited embedded devices. Offloading the computation into the cloud is often infeasible due to privacy concerns, high latency, or the lack of connectivity. As such, there is a critical need to find a way to effectively execute the DNN models locally on the devices. This paper presents an adaptive scheme to determine which DNN model to use for a given input, by considering the desired accuracy and inference time. Our approach employs machine learning to develop a predictive model to quickly select a pre-trained DNN to use for a given input and the optimization constraint. We achieve this by first training off-line a predictive model, and then use the learnt model to select a DNN model to use for new, unseen inputs. We apply our approach to the image classification task and evaluate it on a Jetson TX2 embedded deep learning platform using the ImageNet ILSVRC 2012 validation dataset. We consider a range of influential DNN models. Experimental results show that our approach achieves a 7.52% improvement in inference accuracy, and a 1.8x reduction in inference time over the most-capable single DNN model.