Datum
2023-03-13Schlagwort
004 Informatik Maschinelles LernenNeuronales NetzDatenReflexionsspektroskopieDatenverarbeitungMetadata
Zur Langanzeige
Konferenzveröffentlichung
Utilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Features
Zusammenfassung
Continuous 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.
Zitierform
In: Behn, Carsten; International Academy, Research, and Industry Association (Hrsg.): ICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systems. March13th-17th, 2023, Barcelona; Spain. : 2023-03-13, S. 49-53; isbn:978-1-68558-053-7Zitieren
@inproceedings{doi:10.17170/kobra-2024042910095,
author={Magnussen, Birk Martin and Stern, Claudius and Sick, Bernhard},
title={Utilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Features},
booktitle={ICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systems. March13th-17th, 2023, Barcelona; Spain},
month={03},
year={2023}
}
0500 Oax 0501 Text $btxt$2rdacontent 0502 Computermedien $bc$2rdacarrier 1100 2023$n2023 1500 1/eng 2050 ##0##http://hdl.handle.net/123456789/15737 3000 Magnussen, Birk Martin 3010 Stern, Claudius 3010 Sick, Bernhard 4000 Utilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Features / Magnussen, Birk Martin 4030 4060 Online-Ressource 4085 ##0##=u http://nbn-resolving.de/http://hdl.handle.net/123456789/15737=x R 4204 \$dKonferenzveröffentlichung 4170 5550 {{Maschinelles Lernen}} 5550 {{Neuronales Netz}} 5550 {{Daten}} 5550 {{Reflexionsspektroskopie}} 5550 {{Datenverarbeitung}} 7136 ##0##http://hdl.handle.net/123456789/15737
2024-04-30T14:42:19Z 2024-04-30T14:42:19Z 2023-03-13 doi:10.17170/kobra-2024042910095 http://hdl.handle.net/123456789/15737 eng Urheberrechtlich geschützt https://rightsstatements.org/page/InC/1.0/ machine learning neural nets continuous kernel irregularly sampled data reflection spectroscopy 004 Utilizing Continuous Kernels for Processing Irregularly and Inconsistently Sampled Data With Position-Dependent Features Konferenzveröffentlichung Continuous 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. open access Magnussen, Birk Martin Stern, Claudius Sick, Bernhard Maschinelles Lernen Neuronales Netz Daten Reflexionsspektroskopie Datenverarbeitung publishedVersion 2023-03-13 Barcelona; Online ICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systems. March13th-17th, 2023, Barcelona; Spain Behn, Carsten International Academy, Research, and Industry Association isbn:978-1-68558-053-7 49-53 false ICAS 2023. The Nineteenth International Conference on Autonomic and Autonomous Systems
Die folgenden Lizenzbestimmungen sind mit dieser Ressource verbunden:
:Urheberrechtlich geschützt