Computer Science > Emerging Technologies
[Submitted on 3 Aug 2018 (v1), revised 12 Aug 2018 (this version, v2), latest version 24 Oct 2018 (v3)]
Title:Photonic Spiking Neural Networks - From Devices to Systems
View PDFAbstract:Spiking Neural Networks offer an event-driven and biologically realistic alternative to standard Artificial Neural Networks based on analog information processing which make them more suitable for energy-efficient hardware implementations of the functional units of the brain, namely, neurons and synapses. Despite extensive efforts to replicate the energy efficiency of the brain in standard von-Neumann architecture, the massive parallelism has remained elusive. This has led researchers to venture towards beyond von-Neumann computing or `in-memory' computing paradigms based on CMOS and post-CMOS technologies. However, implementations of such platforms in the electrical domain faces potential limitations of switching speed and interconnect losses. Integrated Photonics offers an welcome alternative to standard `microelectronic' platforms and have recently shown promise as a viable technology for spike-based neural processing systems. However, non-volatility has been identified as an important component of large-scale neuromorphic systems. Although the recent demonstrations of ultra-fast computing with phase-change materials (PCMs) show promise, the jump from standalone computing devices to parallel computing architecture is challenging. In this work, we utilize the optical properties of the PCM, Ge$_2$Sb$_2$Te$_5$ (GST), to propose an all-Photonic Spiking Neural Network, comprising of a non-volatile synaptic array integrated seamlessly with previously explored `integrate-and-fire' neurons to realize an `in-memory' computing platform leveraging the inherent parallelism of wavelength-division-multiplexing (WDM). The proposed design not only bridges the gap between isolated computing devices and parallel large-scale implementation, but also paves the way for ultra-fast computing and localized on-chip learning.
Submission history
From: Indranil Chakraborty [view email][v1] Fri, 3 Aug 2018 16:03:17 UTC (3,540 KB)
[v2] Sun, 12 Aug 2018 23:19:34 UTC (3,473 KB)
[v3] Wed, 24 Oct 2018 01:28:41 UTC (4,005 KB)
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.