Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network
The brain needs to predict how the body reacts to motor commands, but how a network of spiking neurons can learn non-linear body dynamics using local, online and stable learning rules is unclear. Here, we present a supervised learning scheme for the feedforward and recurrent connections in a network...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
eLife Sciences Publications Ltd
2017-11-01
|
| Series: | eLife |
| Subjects: | |
| Online Access: | https://elifesciences.org/articles/28295 |
| _version_ | 1828166242641379328 |
|---|---|
| author | Aditya Gilra Wulfram Gerstner |
| author_facet | Aditya Gilra Wulfram Gerstner |
| author_sort | Aditya Gilra |
| collection | DOAJ |
| description | The brain needs to predict how the body reacts to motor commands, but how a network of spiking neurons can learn non-linear body dynamics using local, online and stable learning rules is unclear. Here, we present a supervised learning scheme for the feedforward and recurrent connections in a network of heterogeneous spiking neurons. The error in the output is fed back through fixed random connections with a negative gain, causing the network to follow the desired dynamics. The rule for Feedback-based Online Local Learning Of Weights (FOLLOW) is local in the sense that weight changes depend on the presynaptic activity and the error signal projected onto the postsynaptic neuron. We provide examples of learning linear, non-linear and chaotic dynamics, as well as the dynamics of a two-link arm. Under reasonable approximations, we show, using the Lyapunov method, that FOLLOW learning is uniformly stable, with the error going to zero asymptotically. |
| first_indexed | 2024-04-12T01:58:31Z |
| format | Article |
| id | doaj.art-d6dccb2a6c32487686bdad3e84e72480 |
| institution | Directory Open Access Journal |
| issn | 2050-084X |
| language | English |
| last_indexed | 2024-04-12T01:58:31Z |
| publishDate | 2017-11-01 |
| publisher | eLife Sciences Publications Ltd |
| record_format | Article |
| series | eLife |
| spelling | doaj.art-d6dccb2a6c32487686bdad3e84e724802022-12-22T03:52:45ZengeLife Sciences Publications LtdeLife2050-084X2017-11-01610.7554/eLife.28295Predicting non-linear dynamics by stable local learning in a recurrent spiking neural networkAditya Gilra0https://orcid.org/0000-0002-8628-1864Wulfram Gerstner1Brain-Mind Institute, School of Life Sciences, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; School of Computer and Communication Sciences, École Polytechnique Fédérale de Lausanne, Lausanne, SwitzerlandBrain-Mind Institute, School of Life Sciences, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerland; School of Computer and Communication Sciences, École Polytechnique Fédérale de Lausanne, Lausanne, SwitzerlandThe brain needs to predict how the body reacts to motor commands, but how a network of spiking neurons can learn non-linear body dynamics using local, online and stable learning rules is unclear. Here, we present a supervised learning scheme for the feedforward and recurrent connections in a network of heterogeneous spiking neurons. The error in the output is fed back through fixed random connections with a negative gain, causing the network to follow the desired dynamics. The rule for Feedback-based Online Local Learning Of Weights (FOLLOW) is local in the sense that weight changes depend on the presynaptic activity and the error signal projected onto the postsynaptic neuron. We provide examples of learning linear, non-linear and chaotic dynamics, as well as the dynamics of a two-link arm. Under reasonable approximations, we show, using the Lyapunov method, that FOLLOW learning is uniformly stable, with the error going to zero asymptotically.https://elifesciences.org/articles/28295learningmotor controlrecurrent neural networksplasticityfeedbackstability |
| spellingShingle | Aditya Gilra Wulfram Gerstner Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network eLife learning motor control recurrent neural networks plasticity feedback stability |
| title | Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network |
| title_full | Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network |
| title_fullStr | Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network |
| title_full_unstemmed | Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network |
| title_short | Predicting non-linear dynamics by stable local learning in a recurrent spiking neural network |
| title_sort | predicting non linear dynamics by stable local learning in a recurrent spiking neural network |
| topic | learning motor control recurrent neural networks plasticity feedback stability |
| url | https://elifesciences.org/articles/28295 |
| work_keys_str_mv | AT adityagilra predictingnonlineardynamicsbystablelocallearninginarecurrentspikingneuralnetwork AT wulframgerstner predictingnonlineardynamicsbystablelocallearninginarecurrentspikingneuralnetwork |