Kathleen E. Hamilton and Catherine D. Schuman

July, 2018

ICONS: International Conference on Neuromorphic Systems

https://ornlcda.github.io/icons2018/

View Article

Abstract

Graph algorithms are a new class of applications for neuromorphic hardware. Rather than adapting deep learning and standard neural network approaches to a low-precision spiking environment, we look at how graph algorithms can be redesigned to incorporate and extract information generated by spiking neurons. While fully connected spin glass implementations of spiking label propagation have shown promising results on graphs with dense communities, identifying sparse communities remains difficult. This work focuses on steps towards adaptive spike-based implementations of label propagation, utilizing sparse embeddings and synaptic plasticity. Sparser embeddings reduce the number of inhibitory connections and synaptic plasticity is used to simultaneously amplify spike responses between neurons in the same community, while impeding spike responses across different communities. We present results on identifying communities in sparse graphs, focusing on graphs with very sparse communities.

Citation Information

Text

author          K. E. Hamilton and C. D. Schuman
title           Towards Adaptive Spiking Label Propagation
booktitle       International Conference on Neuromorphic Computing Systems
publisher       ACM
address         Knoxville, TN
month           July
year            2018

Bibtex

@INPROCEEDINGS{hs:18:tas,
    author = "K. E. Hamilton and C. D. Schuman",
    title = "Towards Adaptive Spiking Label Propagation",
    booktitle = "International Conference on Neuromorphic Computing Systems",
    publisher = "ACM",
    address = "Knoxville, TN",
    month = "July",
    year = "2018"
}