Utilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Features

Simple item page
dc.date.accessioned2024-04-30T14:42:19Z
dc.date.available2024-04-30T14:42:19Z
dc.date.issued2023-03-13
dc.identifierdoi:10.17170/kobra-2024042910095
dc.identifier.urihttp://hdl.handle.net/123456789/15737
dc.language.isoeng
dc.rightsUrheberrechtlich geschützt
dc.rights.urihttps://rightsstatements.org/page/InC/1.0/
dc.subjectmachine learningeng
dc.subjectneural netseng
dc.subjectcontinuous kerneleng
dc.subjectirregularly sampled dataeng
dc.subjectreflection spectroscopyeng
dc.subject.ddc004
dc.subject.swdMaschinelles Lernenger
dc.subject.swdNeuronales Netzger
dc.subject.swdDatenger
dc.subject.swdReflexionsspektroskopieger
dc.subject.swdDatenverarbeitungger
dc.titleUtilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Featureseng
dc.typeKonferenzveröffentlichung
dc.type.versionpublishedVersion
dcterms.abstractContinuous Kernels have been a recent development in convolutional neural networks. Such kernels are used to process data sampled at different resolutions as well as irregularly and inconsistently sampled data. Convolutional neural networks have the property of translational invariance (e.g., features are detected regardless of their position in the measurement domain), which is unsuitable for certain types of data, where the position of detected features is relevant. However, the capabilities of continuous kernels to process irregularly sampled data are still desired. This article introduces a novel method utilizing continuous kernels for detecting global features at absolute positions in the data domain. Through a use case in processing multiple spatially resolved reflection spectroscopy data, which is sampled irregularly and inconsistently, we show that the proposed method is capable of processing such data natively without additional preprocessing as is needed using comparable methods. In addition, we show that the proposed method is able to achieve a higher prediction accuracy than a comparable network on a dataset with position-dependent features. Furthermore, a higher robustness to missing data compared to a benchmark network using data interpolation is observed, which allows the network to adapt to sensors with individual failed components without the need for retraining.eng
dcterms.accessRightsopen access
dcterms.creatorMagnussen, Birk Martin
dcterms.creatorStern, Claudius
dcterms.creatorSick, Bernhard
dcterms.eventICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systemseng
dcterms.event.date2023-03-13
dcterms.event.placeBarcelona; Onlineger
dcterms.source.collectionICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systems. March13th-17th, 2023, Barcelona; Spaineng
dcterms.source.editorBehn, Carsten
dcterms.source.editorInternational Academy, Research, and Industry Association
dcterms.source.identifierisbn:978-1-68558-053-7
dcterms.source.pageinfo49-53
kup.iskupfalse

Files

Original bundle

Now showing 1 - 1 of 1
Thumbnail Image
Name:
MagnussenSternSickUtilizingContinuousKernels.pdf
Size:
701.57 KB
Format:
Adobe Portable Document Format
Description:
Download

License bundle

Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
3.03 KB
Format:
Item-specific license agreed upon to submission
Description:
Download

Collections

Publikationen