Differentiable Implicit Layers

Andreas Look; Simona Doneva; Melih Kandemir; Rainer Gemulla; Jan Peters

Differentiable Implicit Layers

Andreas Look, Simona Doneva, Melih Kandemir, Rainer Gemulla, Jan Peters

Research output: Contribution to conference without publisher/journal › Paper › Research › peer-review

10 Downloads (Pure)

Abstract

In this paper, we introduce an efficient backpropagation scheme for non-constrained implicit functions. These functions are parametrized by a set of learnable weights and may optionally depend on some input; making them perfectly suitable as a learnable layer in a neural network. We demonstrate our scheme on different applications: (i) neural ODEs with the implicit Euler method, and (ii) system identification in model predictive control.

Original language	English
Publication date	14. Oct 2020
Publication status	Published - 14. Oct 2020
Event	Workshop on machine learning for engineering modeling, simulation and design @ NeurIPS 2020 - Duration: 12. Dec 2020 → 12. Dec 2020

Conference

Conference	Workshop on machine learning for engineering modeling, simulation and design @ NeurIPS 2020
Period	12/12/2020 → 12/12/2020

Documents & Links

Open access versionFinal published version, 1.1 MB

Cite this

@conference{16358c3756bc4b278596bee5b027e327,

title = "Differentiable Implicit Layers",

abstract = " In this paper, we introduce an efficient backpropagation scheme for non-constrained implicit functions. These functions are parametrized by a set of learnable weights and may optionally depend on some input; making them perfectly suitable as a learnable layer in a neural network. We demonstrate our scheme on different applications: (i) neural ODEs with the implicit Euler method, and (ii) system identification in model predictive control. ",

keywords = "cs.LG, cs.NE",

author = "Andreas Look and Simona Doneva and Melih Kandemir and Rainer Gemulla and Jan Peters",

year = "2020",

month = oct,

day = "14",

language = "English",

note = "Workshop on machine learning for engineering modeling, simulation and design @ NeurIPS 2020 ; Conference date: 12-12-2020 Through 12-12-2020",

}

TY - CONF

T1 - Differentiable Implicit Layers

AU - Look, Andreas

AU - Doneva, Simona

AU - Kandemir, Melih

AU - Gemulla, Rainer

AU - Peters, Jan

PY - 2020/10/14

Y1 - 2020/10/14

N2 - In this paper, we introduce an efficient backpropagation scheme for non-constrained implicit functions. These functions are parametrized by a set of learnable weights and may optionally depend on some input; making them perfectly suitable as a learnable layer in a neural network. We demonstrate our scheme on different applications: (i) neural ODEs with the implicit Euler method, and (ii) system identification in model predictive control.

AB - In this paper, we introduce an efficient backpropagation scheme for non-constrained implicit functions. These functions are parametrized by a set of learnable weights and may optionally depend on some input; making them perfectly suitable as a learnable layer in a neural network. We demonstrate our scheme on different applications: (i) neural ODEs with the implicit Euler method, and (ii) system identification in model predictive control.

KW - cs.LG

KW - cs.NE

M3 - Paper

T2 - Workshop on machine learning for engineering modeling, simulation and design @ NeurIPS 2020

Y2 - 12 December 2020 through 12 December 2020

ER -

Differentiable Implicit Layers

Abstract

Conference

Documents & Links

Fingerprint

Cite this