Temporal Patience: Efficient Adaptive Deep Learning for Embedded Radar Data Processing

TL;DR

This paper introduces adaptive early exit neural networks that leverage temporal correlations in streaming radar data to reduce computational costs on embedded devices, enabling real-time processing for smart applications.

Contribution

The paper proposes novel techniques for early exit neural networks that utilize temporal correlation in radar data to improve efficiency on resource-constrained embedded platforms.

Findings

Up to 26% reduction in operations per inference.

Works on commodity hardware with traditional optimizations.

Facilitates real-time radar data processing in smart devices.

Abstract

Radar sensors offer power-efficient solutions for always-on smart devices, but processing the data streams on resource-constrained embedded platforms remains challenging. This paper presents novel techniques that leverage the temporal correlation present in streaming radar data to enhance the efficiency of Early Exit Neural Networks for Deep Learning inference on embedded devices. These networks add additional classifier branches between the architecture's hidden layers that allow for an early termination of the inference if their result is deemed sufficient enough by an at-runtime decision mechanism. Our methods enable more informed decisions on when to terminate the inference, reducing computational costs while maintaining a minimal loss of accuracy. Our results demonstrate that our techniques save up to 26% of operations per inference over a Single Exit Network and 12% over a confidence-based Early Exit version. Our proposed techniques work on commodity hardware and can be combined with traditional optimizations, making them accessible for resource-constrained embedded platforms commonly used in smart devices. Such efficiency gains enable real-time radar data processing on resource-constrained platforms, allowing for new applications in the context of smart homes, Internet-of-Things, and human-computer interaction.