HyperSpike: HyperDimensional Computing for More Efficient and Robust Spiking Neural Networks
Justin Morris1,2,a, Hin Wai Lui3,b, Kenneth Stewart3,c, Behnam Khaleghi1,d, Anthony Thomas1,e, Thiago Marback1,f, Baris Aksanli2,g, Emre Neftci3,h and Tajana Rosing1,i
1University of California San Diego, La Jolla, CA 92093, USA
2San Diego State University, San Diego, CA 92182, USA
3University of California Irvine, Irvine, CA 92697, USA
ajustinmorris@ucsd.edu
bhwluiu@uci.edu
ckennetms@uci.edu
dbkhalegh@ucd.edu
eahthomas@ucd.edu
ftmarback@ucd.edu
gbaksanli@sdsu.edu
heneftci@uci.edu
itajana@ucsd.edu
ABSTRACT
Today’s Machine Learning(ML) systems, especially those running in server farms running workloads such as Deep Neural Networks, which require billions of parameters and many hours to train a model, consume a significant amount of energy. To combat this, researchers have been focusing on new emerging neuromorphic computing models. Two of those models are Hyperdimensional Computing (HDC) and Spiking Neural Networks (SNNs), both with their own benefits. HDC has various desirable properties that other Machine Learning (ML) algorithms lack such as: robustness to noise in the system, simple operations, and high parallelism. SNNs are able to process event based signal data in an efficient manner. In this paper, we create HyperSpike, which utilizes a single, randomly initialized and untrained SNN layer as feature extractor connected to a trained HDC classifier. HDC is used to enable more efficient classification as well as provide robustness to errors. We experimentally show that HyperSpike is on average 31:5× more robust to errors than traditional SNNs. We also implement HyperSpike in hardware, and show that it is 10× faster and 2:6× more energy efficient over traditional SNN networks run on Intel’s Loihi [1].
