Coupled Oscillatory Recurrent Neural Network (coRNN): An accurate and (gradient) stable architecture for learning long time dependencies
dc.contributor.author
Rusch, T. Konstantin
dc.contributor.author
Mishra, Siddhartha
dc.date.accessioned
2020-10-22T13:50:59Z
dc.date.available
2020-10-22T09:18:15Z
dc.date.available
2020-10-22T13:50:59Z
dc.date.issued
2020-10
dc.identifier.uri
http://hdl.handle.net/20.500.11850/447238
dc.description.abstract
Circuits of biological neurons, such as in the functional parts of the brain can be modeled as networks of coupled oscillators. Inspired by the ability of these systems to express a rich set of outputs while keeping (gradients of) state variables bounded, we propose a novel architecture for recurrent neural networks. Our proposed RNN is based on a time-discretization of a system of second-order ordinary differential equations, modeling networks of controlled nonlinear oscillators. We prove precise bounds on the gradients of the hidden states, leading to the mitigation of the exploding and vanishing gradient problem for this RNN. Experiments show that the proposed RNN is comparable in performance to the state of the art on a variety of benchmarks, demonstrating the potential of this architecture to provide stable and accurate RNNs for processing complex sequential data.
en_US
dc.language.iso
en
en_US
dc.publisher
Seminar for Applied Mathematics, ETH Zurich
en_US
dc.subject
RNNs
en_US
dc.subject
Oscillators
en_US
dc.subject
Gradient stability
en_US
dc.subject
Long-term dependencies
en_US
dc.title
Coupled Oscillatory Recurrent Neural Network (coRNN): An accurate and (gradient) stable architecture for learning long time dependencies
en_US
dc.type
Report
ethz.journal.title
SAM Research Report
ethz.journal.volume
2020-63
en_US
ethz.size
19 p.
en_US
ethz.publication.place
Zurich
en_US
ethz.publication.status
published
en_US
ethz.leitzahl
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02000 - Dep. Mathematik / Dep. of Mathematics::02501 - Seminar für Angewandte Mathematik / Seminar for Applied Mathematics::03851 - Mishra, Siddhartha / Mishra, Siddhartha
en_US
ethz.leitzahl.certified
ETH Zürich::00002 - ETH Zürich::00012 - Lehre und Forschung::00007 - Departemente::02000 - Dep. Mathematik / Dep. of Mathematics::02501 - Seminar für Angewandte Mathematik / Seminar for Applied Mathematics::03851 - Mishra, Siddhartha / Mishra, Siddhartha
en_US
ethz.identifier.url
https://math.ethz.ch/sam/research/reports.html?id=936
ethz.date.deposited
2020-10-22T09:18:24Z
ethz.source
FORM
ethz.eth
yes
en_US
ethz.identifier.internal
https://math.ethz.ch/sam/research/reports.html?id=936
en_US
ethz.availability
Metadata only
en_US
ethz.rosetta.installDate
2020-10-22T13:51:10Z
ethz.rosetta.lastUpdated
2020-10-22T13:51:10Z
ethz.rosetta.versionExported
true
ethz.COinS
ctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.atitle=Coupled%20Oscillatory%20Recurrent%20Neural%20Network%20(coRNN):%20An%20accurate%20and%20(gradient)%20stable%20architecture%20for%20learning%20long%20time%20dependencies&rft.jtitle=SAM%20Research%20Report&rft.date=2020-10&rft.volume=2020-63&rft.au=Rusch,%20T.%20Konstantin&Mishra,%20Siddhartha&rft.genre=report&
Files in this item
Files | Size | Format | Open in viewer |
---|---|---|---|
There are no files associated with this item. |
Publication type
-
Report [6965]