About: Neural Power Units

Facets (new session)
Description
Metadata
Settings
- owl:sameAs
- Inference Rule:

About: Neural Power Units Goto Sponge NotDistinct Permalink

An Entity of Type : schema:ScholarlyArticle, within Data Space : wasabi.inria.fr associated with source document(s)

Attributes	Values
type	Academic Article research paper schema:ScholarlyArticle
isDefinedBy	Covid-on-the-Web dataset
has title	Neural Power Units
Creator	Heim, Niklas Pevný, Tomáš
Source	ArXiv
abstract	Conventional Neural Networks can approximate simple arithmetic operations, but fail to generalize beyond the range of numbers that were seen during training. Neural Arithmetic Units aim to overcome this difficulty, but current arithmetic units are either limited to operate on positive numbers or can only represent a subset of arithmetic operations. We introduce the Neural Power Unit (NPU) that operates on the full domain of real numbers and is capable of learning arbitrary power functions in a single layer. The NPU thus fixes the shortcomings of existing arithmetic units and extends their expressivity. We achieve this by using complex arithmetic without requiring a conversion of the network to complex numbers. A simplification of the unit to the RealNPU yields a highly interpretable model. We show that the NPUs outperform their competitors in terms of accuracy and sparsity on artificial arithmetic datasets, and that the RealNPU can discover the governing equations of a dynamical systems only from data.
has issue date	2020-06-02(xsd:dateTime)
has license	arxiv
sha1sum (hex)	0bfd4ed399054eae26c3cdaabc0aed80ca95e125
resource representing a document's title	Neural Power Units
resource representing a document's body	covid:0bfd4ed399054eae26c3cdaabc0aed80ca95e125#body_text
is schema:about of	named entity 'Unit' named entity 'yields' named entity 'governing' named entity 'limited' named entity 'artificial' named entity 'datasets' named entity 'complex' named entity 'Power' named entity 'HIGHLY' named entity 'DATA' named entity 'DOMAIN' named entity 'TRAINING' covid:arg/0bfd4ed399054eae26c3cdaabc0aed80ca95e125 named entity 'operates' named entity 'Neural Networks' named entity 'generalize' named entity 'arithmetic' named entity 'numbers' named entity 'domain' named entity 'introduce' named entity 'overcome' named entity 'fail' named entity 'current' named entity 'existing' named entity 'terms' named entity 'complex arithmetic' named entity 'power functions' named entity 'complex numbers' named entity 'arithmetic operations' named entity 'arithmetic units' named entity 'learning rate' named entity 'fractional powers' named entity 'complex logarithm' named entity 'square-root' named entity 'symbolic integration' named entity 'power functions' named entity 'control systems' named entity 'power functions' named entity '4.2' named entity 'scientific applications' named entity 'neural network' named entity 'L1 regularization' named entity 'NALU' named entity 'NALU' named entity 'logarithm' named entity 'arithmetic' named entity 'multiplication' named entity 'intelligent agent' named entity 'multiplication' named entity 'inductive bias' named entity 'inductive bias' named entity 'Arithmetic Units' named entity 'black-box models' named entity 'SIR model' named entity 'NALU' named entity 'multiplication' named entity 'machine learning' named entity 'arithmetic' named entity 'NALU' named entity 'multiplication' named entity 'NALU' named entity 'matrices' named entity 'NALU' named entity 'multiplication' named entity 'logarithm' named entity 'sparsity' named entity 'Sec. 4.3' named entity 'NALU' named entity 'epidemiological SIR model' named entity 'subset' named entity 'parameter space' named entity 'complex arithmetic' named entity 'simple algebra' named entity 'real networks' named entity 'NALU' named entity 'arithmetics' named entity 'arithmetic' named entity 'fractional powers' named entity 'human intelligence' named entity 'SIR model' named entity 'COVID-19 outbreak' named entity 'sparsity' named entity 'arithmetic units' named entity 'basis functions' named entity 'arithmetic operations' named entity 'square-root'

Faceted Search & Find service v1.13.91 as of Mar 24 2020

Alternative Linked Data Documents: Sponger | ODE Content Formats:

RDF

ODATA

Microdata

About

OpenLink Virtuoso version 07.20.3229 as of Jul 10 2020, on Linux (x86_64-pc-linux-gnu), Single-Server Edition (94 GB total memory)
Data on this page belongs to its respective rights holders.
Virtuoso Faceted Browser Copyright © 2009-2024 OpenLink Software