HyperSNN: A new efficient and robust deep learning model for resource constrained control applications

TL;DR

HyperSNN combines spiking neural networks with hyperdimensional computing to create an energy-efficient, robust control model suitable for resource-constrained edge devices, achieving comparable accuracy to traditional methods with significantly lower energy use.

Contribution

This paper introduces HyperSNN, a novel control model that integrates SNNs and hyperdimensional computing, reducing energy consumption and increasing robustness for edge computing applications.

Findings

HyperSNN achieves similar control accuracy to traditional methods.

Energy consumption is reduced to 1.36%-9.96% of conventional models.

HyperSNN demonstrates increased robustness in control tasks.

Abstract

In light of the increasing adoption of edge computing in areas such as intelligent furniture, robotics, and smart homes, this paper introduces HyperSNN, an innovative method for control tasks that uses spiking neural networks (SNNs) in combination with hyperdimensional computing. HyperSNN substitutes expensive 32-bit floating point multiplications with 8-bit integer additions, resulting in reduced energy consumption while enhancing robustness and potentially improving accuracy. Our model was tested on AI Gym benchmarks, including Cartpole, Acrobot, MountainCar, and Lunar Lander. HyperSNN achieves control accuracies that are on par with conventional machine learning methods but with only 1.36% to 9.96% of the energy expenditure. Furthermore, our experiments showed increased robustness when using HyperSNN. We believe that HyperSNN is especially suitable for interactive, mobile, and wearable devices, promoting energy-efficient and robust system design. Furthermore, it paves the way for the practical implementation of complex algorithms like model predictive control (MPC) in real-world industrial scenarios.