Iterative free-energy optimization for recurrent neural networks (INFERNO).
The intra-parietal lobe coupled with the Basal Ganglia forms a working memory that demonstrates strong planning capabilities for generating robust yet flexible neuronal sequences. Neurocomputational models however, often fails to control long range neural synchrony in recurrent spiking networks due...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Public Library of Science (PLoS)
2017-01-01
|
| Series: | PLoS ONE |
| Online Access: | http://europepmc.org/articles/PMC5345841?pdf=render |
| _version_ | 1819211023050080256 |
|---|---|
| author | Alexandre Pitti Philippe Gaussier Mathias Quoy |
| author_facet | Alexandre Pitti Philippe Gaussier Mathias Quoy |
| author_sort | Alexandre Pitti |
| collection | DOAJ |
| description | The intra-parietal lobe coupled with the Basal Ganglia forms a working memory that demonstrates strong planning capabilities for generating robust yet flexible neuronal sequences. Neurocomputational models however, often fails to control long range neural synchrony in recurrent spiking networks due to spontaneous activity. As a novel framework based on the free-energy principle, we propose to see the problem of spikes' synchrony as an optimization problem of the neurons sub-threshold activity for the generation of long neuronal chains. Using a stochastic gradient descent, a reinforcement signal (presumably dopaminergic) evaluates the quality of one input vector to move the recurrent neural network to a desired activity; depending on the error made, this input vector is strengthened to hill-climb the gradient or elicited to search for another solution. This vector can be learned then by one associative memory as a model of the basal-ganglia to control the recurrent neural network. Experiments on habit learning and on sequence retrieving demonstrate the capabilities of the dual system to generate very long and precise spatio-temporal sequences, above two hundred iterations. Its features are applied then to the sequential planning of arm movements. In line with neurobiological theories, we discuss its relevance for modeling the cortico-basal working memory to initiate flexible goal-directed neuronal chains of causation and its relation to novel architectures such as Deep Networks, Neural Turing Machines and the Free-Energy Principle. |
| first_indexed | 2024-12-23T06:20:28Z |
| format | Article |
| id | doaj.art-8bcad1f098d648ab96662c821764940a |
| institution | Directory Open Access Journal |
| issn | 1932-6203 |
| language | English |
| last_indexed | 2024-12-23T06:20:28Z |
| publishDate | 2017-01-01 |
| publisher | Public Library of Science (PLoS) |
| record_format | Article |
| series | PLoS ONE |
| spelling | doaj.art-8bcad1f098d648ab96662c821764940a2022-12-21T17:57:13ZengPublic Library of Science (PLoS)PLoS ONE1932-62032017-01-01123e017368410.1371/journal.pone.0173684Iterative free-energy optimization for recurrent neural networks (INFERNO).Alexandre PittiPhilippe GaussierMathias QuoyThe intra-parietal lobe coupled with the Basal Ganglia forms a working memory that demonstrates strong planning capabilities for generating robust yet flexible neuronal sequences. Neurocomputational models however, often fails to control long range neural synchrony in recurrent spiking networks due to spontaneous activity. As a novel framework based on the free-energy principle, we propose to see the problem of spikes' synchrony as an optimization problem of the neurons sub-threshold activity for the generation of long neuronal chains. Using a stochastic gradient descent, a reinforcement signal (presumably dopaminergic) evaluates the quality of one input vector to move the recurrent neural network to a desired activity; depending on the error made, this input vector is strengthened to hill-climb the gradient or elicited to search for another solution. This vector can be learned then by one associative memory as a model of the basal-ganglia to control the recurrent neural network. Experiments on habit learning and on sequence retrieving demonstrate the capabilities of the dual system to generate very long and precise spatio-temporal sequences, above two hundred iterations. Its features are applied then to the sequential planning of arm movements. In line with neurobiological theories, we discuss its relevance for modeling the cortico-basal working memory to initiate flexible goal-directed neuronal chains of causation and its relation to novel architectures such as Deep Networks, Neural Turing Machines and the Free-Energy Principle.http://europepmc.org/articles/PMC5345841?pdf=render |
| spellingShingle | Alexandre Pitti Philippe Gaussier Mathias Quoy Iterative free-energy optimization for recurrent neural networks (INFERNO). PLoS ONE |
| title | Iterative free-energy optimization for recurrent neural networks (INFERNO). |
| title_full | Iterative free-energy optimization for recurrent neural networks (INFERNO). |
| title_fullStr | Iterative free-energy optimization for recurrent neural networks (INFERNO). |
| title_full_unstemmed | Iterative free-energy optimization for recurrent neural networks (INFERNO). |
| title_short | Iterative free-energy optimization for recurrent neural networks (INFERNO). |
| title_sort | iterative free energy optimization for recurrent neural networks inferno |
| url | http://europepmc.org/articles/PMC5345841?pdf=render |
| work_keys_str_mv | AT alexandrepitti iterativefreeenergyoptimizationforrecurrentneuralnetworksinferno AT philippegaussier iterativefreeenergyoptimizationforrecurrentneuralnetworksinferno AT mathiasquoy iterativefreeenergyoptimizationforrecurrentneuralnetworksinferno |