spiking-models-for-TSC

Code to accompany our paper: Reservoir based Spiking Models for Univariate Time Series Classification

Accepted in Frontiers in Computational Neuroscience journal, 2023 (open access).

Two models are presented in the paper:

• Spiking Legendre Reservoir Computing (SLRC) model
• Legendre Spiking Neural Network (LSNN) model

Repository description:

• slrc-model directory contains all the code for SLRC model.
• lsnn-model directory contains all the code for LSNN model.
• energy-consumption-analysis directory contains all the code for measuring energy consumption on Loihi-1 and CPU.

To run the slrc-model and lsnn-model directories codes, you would be requried to download the datasets (mentioned in the paper) from https://timeseriesclassification.com/ website, along with setting up the environment by install Nengo, NengoLoihi, and PyTorch libraries.

To run the code in energy-consumption-analysis, you don't need to download datasets, but have access to Loihi-1 on INRC and of course an Intel CPU machine. You would still need to install Nengo and NengoLoihi libraries along with pyJoules library.