Efficient-Scalable-Machine-Learning / beyond-weights

Implementation of pure synaptic-delay training of spiking neural networks (based on SLAYER)

Beyond Weights

This repository contains an implementation of the paper Beyond Weights: Deep learning in Spiking Neural Networks with pure synaptic-delay training.

These mastodon posts provide a short summary of this work.

This code is a forked version of the SLAYER repository with appropriate changes.

Usage

The environment.yml file provides a snapshot of the conda environment that can be used with conda env create -f environment.yml. Install this package with python setup.py install Note that you might have to install the pip packages listed there manually.

Experiments with MNIST and Fashion MNIST are on ./example/experiments, to run simply "python file_name.py"

Citation

Edoardo W. Grappolini and Anand Subramoney. Beyond weights: Deep learning in spiking neural networks with pure synaptic-delay training. In International Conference on Neuromorphic Systems (ICONS ’23), Santa Fe, NM, USA. ACM, June 2023.

@inproceedings{grappolini2023weights,
      title={Beyond Weights: Deep learning in Spiking Neural Networks with pure synaptic-delay training}, 
      author={Edoardo W. Grappolini and Anand Subramoney},
      booktitle = {International Conference on Neuromorphic Systems (ICONS '23), Santa Fe, NM, USA},
      year={2023},
      month = jun,
      publisher = {ACM},
}