University of Surrey

Test tubes in the lab Research in the ATI Dance Research

Optimal Supervised Learning in Spiking Neural Networks for Precise Temporal Encoding

Gardner, B and Grüning, A (2016) Optimal Supervised Learning in Spiking Neural Networks for Precise Temporal Encoding arXiv.

[img]
Preview
Text
1601.03649v1.pdf - ["content_typename_Submitted version (pre-print)" not defined]
Available under License : See the attached licence file.

Download (732kB) | Preview
[img] Text (deleted)
1601.03649v1.pdf - ["content_typename_Submitted version (pre-print)" not defined]
Restricted to Repository staff only
Available under License : See the attached licence file.

Download (732kB)
[img]
Preview
PDF (licence)
SRI_deposit_agreement.pdf
Available under License : See the attached licence file.

Download (33kB) | Preview

Abstract

Precise spike timing as a means to encode information in neural networks is biologically supported, and is advantageous over frequency-based codes by processing input features on a much shorter time-scale. For these reasons, much recent attention has been focused on the development of supervised learning rules for spiking neural networks that utilise a temporal coding scheme. However, despite significant progress in this area, there still lack rules that are theoretically justified, and yet can be considered biologically relevant. Here we examine the general conditions under which optimal synaptic plasticity takes place to support the supervised learning of a precise temporal code. As part of our analysis we introduce two analytically derived learning rules, one of which relies on an instantaneous error signal to optimise synaptic weights in a network (INST rule), and the other one relying on a filtered error signal to minimise the variance of synaptic weight modifications (FILT rule). We test the optimality of the solutions provided by each rule with respect to their temporal encoding precision, and then measure the maximum number of input patterns they can learn to memorise using the precise timings of individual spikes. Our results demonstrate the optimality of the FILT rule in most cases, underpinned by the rule's error-filtering mechanism which provides smooth convergence during learning. We also find the FILT rule to be most efficient at performing input pattern memorisations, and most noticeably when patterns are identified using spikes with sub-millisecond temporal precision. In comparison with existing work, we determine the performance of the FILT rule to be consistent with that of the highly efficient E-learning Chronotron rule, but with the distinct advantage that our FILT rule is also implementable as an online method for increased biological realism.

Item Type: Article
Subjects : Computer Science
Divisions : Faculty of Engineering and Physical Sciences
Authors :
AuthorsEmailORCID
Gardner, BUNSPECIFIEDUNSPECIFIED
Grüning, AUNSPECIFIEDUNSPECIFIED
Date : 14 January 2016
Uncontrolled Keywords : cs.NE, cs.NE, q-bio.NC
Related URLs :
Additional Information : This is the arXiv version of the paper.
Depositing User : Symplectic Elements
Date Deposited : 27 Jan 2016 09:55
Last Modified : 27 Jan 2016 10:11
URI: http://epubs.surrey.ac.uk/id/eprint/809802

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year


Information about this web site

© The University of Surrey, Guildford, Surrey, GU2 7XH, United Kingdom.
+44 (0)1483 300800