Memory-Efficient Deep Learning on a SpiNNaker 2 Prototype
- Submitting institution
-
The University of Manchester
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 82385020
- Type
- D - Journal article
- DOI
-
10.3389/fnins.2018.00840
- Title of journal
- Frontiers in Neuroscience
- Article number
- 00840
- First page
- -
- Volume
- 12
- Issue
- NOV
- ISSN
- 1662-4548
- Open access status
- Compliant
- Month of publication
- November
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
10
- Research group(s)
-
A - Computer Science
- Citation count
- 10
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- "This paper describes collaborative work in the EU Human Brain Project between the SpiNNaker2 designers (Manchester and TU Dresden) and theoretical neuroscientists (TU Graz).
Shows that synaptic rewiring reduces the number of connections in an artificial neural network by 100x, greatly reducing memory and energy requirements to levels available in small edge computing devices without compromising performance.
Enabled the Saxony Government investment of EUR8,000,000 into SpiNNaker2.
Announcement (Sept 2019) via EU Human Brain Project Twitter account said: ""The new computer will constitute a significant leap forward in Neuromorphic Computing as the second generation of the pioneering SpiNNaker system""."
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
