Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors

Daniel R. Mendat; Sang Chin; Steve Furber; Andreas G. Andreou

doi:10.1109/LASCAS.2016.7451094

Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors

Daniel R. Mendat, Sang Chin, Steve Furber, Andreas G. Andreou

Advanced Processor Technology

Johns Hopkins University

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

Abstract

We present a bio-inspired, hardware/software architecture to perform Markov Chain Monte Carlo sampling on probabilistic graphical models using energy aware hardware. We have developed algorithms and programming data flows for two recently developed multiprocessor architectures, the SpiNNaker and Parallella. We employ a neurally inspired sampling algorithm that abstracts the functionality of neurons in a biological network and exploits the neural dynamics to implement the sampling process. This algorithm maps nicely on the two hardware systems. Speedups as high as 1000 fold are achieved when performing inference using this approach, compared to algorithms running on traditional engineering workstations.

Original language	English
Title of host publication	LASCAS 2016 - 7th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference
Publisher	IEEE
Pages	399-402
Number of pages	4
ISBN (Print)	9781467378352
DOIs	https://doi.org/10.1109/LASCAS.2016.7451094
Publication status	Published - 14 Apr 2016
Event	7th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2016 - Florianopolis, Brazil Duration: 27 Feb 2016 → 1 Mar 2016

Conference

Conference	7th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2016
Country/Territory	Brazil
City	Florianopolis
Period	27/02/16 → 1/03/16

Access to Document

10.1109/LASCAS.2016.7451094

Cite this

@inproceedings{1549033b9869423ca51bb759b1afa660,

title = "Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors",

abstract = "We present a bio-inspired, hardware/software architecture to perform Markov Chain Monte Carlo sampling on probabilistic graphical models using energy aware hardware. We have developed algorithms and programming data flows for two recently developed multiprocessor architectures, the SpiNNaker and Parallella. We employ a neurally inspired sampling algorithm that abstracts the functionality of neurons in a biological network and exploits the neural dynamics to implement the sampling process. This algorithm maps nicely on the two hardware systems. Speedups as high as 1000 fold are achieved when performing inference using this approach, compared to algorithms running on traditional engineering workstations.",

author = "Mendat, {Daniel R.} and Sang Chin and Steve Furber and Andreou, {Andreas G.}",

year = "2016",

month = apr,

day = "14",

doi = "10.1109/LASCAS.2016.7451094",

language = "English",

isbn = "9781467378352",

pages = "399--402",

booktitle = "LASCAS 2016 - 7th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference",

publisher = "IEEE",

address = "United States",

note = "7th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2016 ; Conference date: 27-02-2016 Through 01-03-2016",

}

Mendat, DR, Chin, S, Furber, S & Andreou, AG 2016, Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors. in LASCAS 2016 - 7th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference., 7451094, IEEE, pp. 399-402, 7th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2016, Florianopolis, Brazil, 27/02/16. https://doi.org/10.1109/LASCAS.2016.7451094

TY - GEN

T1 - Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors

AU - Mendat, Daniel R.

AU - Chin, Sang

AU - Furber, Steve

AU - Andreou, Andreas G.

PY - 2016/4/14

Y1 - 2016/4/14

N2 - We present a bio-inspired, hardware/software architecture to perform Markov Chain Monte Carlo sampling on probabilistic graphical models using energy aware hardware. We have developed algorithms and programming data flows for two recently developed multiprocessor architectures, the SpiNNaker and Parallella. We employ a neurally inspired sampling algorithm that abstracts the functionality of neurons in a biological network and exploits the neural dynamics to implement the sampling process. This algorithm maps nicely on the two hardware systems. Speedups as high as 1000 fold are achieved when performing inference using this approach, compared to algorithms running on traditional engineering workstations.

AB - We present a bio-inspired, hardware/software architecture to perform Markov Chain Monte Carlo sampling on probabilistic graphical models using energy aware hardware. We have developed algorithms and programming data flows for two recently developed multiprocessor architectures, the SpiNNaker and Parallella. We employ a neurally inspired sampling algorithm that abstracts the functionality of neurons in a biological network and exploits the neural dynamics to implement the sampling process. This algorithm maps nicely on the two hardware systems. Speedups as high as 1000 fold are achieved when performing inference using this approach, compared to algorithms running on traditional engineering workstations.

UR - http://www.scopus.com/inward/record.url?scp=84980425638&partnerID=8YFLogxK

U2 - 10.1109/LASCAS.2016.7451094

DO - 10.1109/LASCAS.2016.7451094

M3 - Conference contribution

SN - 9781467378352

SP - 399

EP - 402

BT - LASCAS 2016 - 7th IEEE Latin American Symposium on Circuits and Systems, R9 IEEE CASS Flagship Conference

PB - IEEE

T2 - 7th IEEE Latin American Symposium on Circuits and Systems, LASCAS 2016

Y2 - 27 February 2016 through 1 March 2016

ER -

Neuromorphic sampling on the SpiNNaker and parallella chip multiprocessors

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this