Stochastic Variational Learning in Recurrent Spiking Networks
The ability to learn and perform statistical inference with biologically plausible recurrent network of spiking neurons is an important step towards understanding perception and reasoning. <br/>Here we derive and investigate a new learning rule for recurrent spiking networks with hidden neuro...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2014-04-01
|
Series: | Frontiers in Computational Neuroscience |
Subjects: | |
Online Access: | http://journal.frontiersin.org/Journal/10.3389/fncom.2014.00038/full |
id |
doaj-11c6423bfbc342768f5892eafed81c92 |
---|---|
record_format |
Article |
spelling |
doaj-11c6423bfbc342768f5892eafed81c922020-11-25T01:00:38ZengFrontiers Media S.A.Frontiers in Computational Neuroscience1662-51882014-04-01810.3389/fncom.2014.0003875811Stochastic Variational Learning in Recurrent Spiking NetworksDanilo eJimenez Rezende0Danilo eJimenez Rezende1Wulfram eGerstner2Wulfram eGerstner3Ecole Polytechnique Federale de LausanneEcole Polytechnique Federale de LausanneEcole Polytechnique Federale de LausanneEcole Polytechnique Federale de LausanneThe ability to learn and perform statistical inference with biologically plausible recurrent network of spiking neurons is an important step towards understanding perception and reasoning. <br/>Here we derive and investigate a new learning rule for recurrent spiking networks with hidden neurons, combining principles from variational learning and reinforcement learning. <br/>Our network defines a generative model over <br/>spike train histories and the derived learning rule has the form of a <br/>local Spike Timing Dependent Plasticity rule modulated by global factors (neuromodulators) conveying information about ``novelty on a statistically rigorous ground.<br/>Simulations show that our model is able to learn both<br/>stationary and non-stationary patterns of spike trains.<br/>We also propose one experiment that could potentially be performed with animals in order to test the dynamics of the predicted novelty signal.http://journal.frontiersin.org/Journal/10.3389/fncom.2014.00038/fullAction PotentialsSynapsesspiking neuronsneural networksvariational learning |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Danilo eJimenez Rezende Danilo eJimenez Rezende Wulfram eGerstner Wulfram eGerstner |
spellingShingle |
Danilo eJimenez Rezende Danilo eJimenez Rezende Wulfram eGerstner Wulfram eGerstner Stochastic Variational Learning in Recurrent Spiking Networks Frontiers in Computational Neuroscience Action Potentials Synapses spiking neurons neural networks variational learning |
author_facet |
Danilo eJimenez Rezende Danilo eJimenez Rezende Wulfram eGerstner Wulfram eGerstner |
author_sort |
Danilo eJimenez Rezende |
title |
Stochastic Variational Learning in Recurrent Spiking Networks |
title_short |
Stochastic Variational Learning in Recurrent Spiking Networks |
title_full |
Stochastic Variational Learning in Recurrent Spiking Networks |
title_fullStr |
Stochastic Variational Learning in Recurrent Spiking Networks |
title_full_unstemmed |
Stochastic Variational Learning in Recurrent Spiking Networks |
title_sort |
stochastic variational learning in recurrent spiking networks |
publisher |
Frontiers Media S.A. |
series |
Frontiers in Computational Neuroscience |
issn |
1662-5188 |
publishDate |
2014-04-01 |
description |
The ability to learn and perform statistical inference with biologically plausible recurrent network of spiking neurons is an important step towards understanding perception and reasoning. <br/>Here we derive and investigate a new learning rule for recurrent spiking networks with hidden neurons, combining principles from variational learning and reinforcement learning. <br/>Our network defines a generative model over <br/>spike train histories and the derived learning rule has the form of a <br/>local Spike Timing Dependent Plasticity rule modulated by global factors (neuromodulators) conveying information about ``novelty on a statistically rigorous ground.<br/>Simulations show that our model is able to learn both<br/>stationary and non-stationary patterns of spike trains.<br/>We also propose one experiment that could potentially be performed with animals in order to test the dynamics of the predicted novelty signal. |
topic |
Action Potentials Synapses spiking neurons neural networks variational learning |
url |
http://journal.frontiersin.org/Journal/10.3389/fncom.2014.00038/full |
work_keys_str_mv |
AT daniloejimenezrezende stochasticvariationallearninginrecurrentspikingnetworks AT daniloejimenezrezende stochasticvariationallearninginrecurrentspikingnetworks AT wulframegerstner stochasticvariationallearninginrecurrentspikingnetworks AT wulframegerstner stochasticvariationallearninginrecurrentspikingnetworks |
_version_ |
1725212723959037952 |