Spiking Approximations of the MaxPooling Operation in Deep SNNs

Accepted in IJCNN 2022, Padua, Italy

• Link to the pre-print paper on Arxiv with Technical Appendix / Supplementary Material
• Link to the camera-ready paper on IEEE Xplore

Short Abstract:

MaxPooling in Spiking Neural Networks (SNNs) is not trivial; it isn't a regular max() operation. This paper proposes two different methods of implementing MaxPooling in SNNs - called AVAM and MJOP Spiking-MaxPooling, both of which are entirely deployable on Loihi neuromorphic hardware.

Misc

A tutorial on the introduction to MJOP and AVAM methods can be found here. Feel free to use the code in this repository.