Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance. Klein, P., Weingarz, N., & Bergmann, R. In IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning, volume 1325, of Communications in Computer and Information Science, 2020. Springer International Publishing.. The original publication is available at www.springerlink.com![Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance [pdf]](https://bibbase.org/img/filetypes/pdf.svg "pdf")
Paper doi abstract bibtex 45 downloads The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..
@inproceedings{klein_SiameseNN4PredM_2020,
title = {{Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance}},
author = {Patrick Klein and Niklas Weingarz and Ralph Bergmann},
year = 2020,
year = 2020,
booktitle = {{IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning}},
publisher = {Springer International Publishing.},
series = {Communications in Computer and Information Science},
volume = 1325,
doi = {10.1007/978-3-030-66770-2\_6},
isbn = {978-3-030-66769-6},
url = {http://www.wi2.uni-trier.de/shared/publications/2020_ECML-IoTStreams_SiameseNeuralNetwork-for-PredictiveMaintenance_Preprint.pdf},
note = {The original publication is available at www.springerlink.com},
%pages = {1--16},
keywords = {{Siamese Neural Network, Predictive Maintenance, Expert Knowledge, 2D Convolution, Time Series Similarity}},
abstract = {The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..}
}
Downloads: 45
{"_id":"kSm8TjvThPLzwtcJM","bibbaseid":"klein-weingarz-bergmann-enhancingsiameseneuralnetworksthroughexpertknowledgeforpredictivemaintenance-2020","authorIDs":["3FKNrvYDeY7gXnjbC","9Hjm5fp7EXedfiskD"],"author_short":["Klein, P.","Weingarz, N.","Bergmann, R."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","title":"Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance","author":[{"firstnames":["Patrick"],"propositions":[],"lastnames":["Klein"],"suffixes":[]},{"firstnames":["Niklas"],"propositions":[],"lastnames":["Weingarz"],"suffixes":[]},{"firstnames":["Ralph"],"propositions":[],"lastnames":["Bergmann"],"suffixes":[]}],"year":"2020","booktitle":"IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning","publisher":"Springer International Publishing.","series":"Communications in Computer and Information Science","volume":"1325","doi":"10.1007/978-3-030-66770-2_6","isbn":"978-3-030-66769-6","url":"http://www.wi2.uni-trier.de/shared/publications/2020_ECML-IoTStreams_SiameseNeuralNetwork-for-PredictiveMaintenance_Preprint.pdf","note":"The original publication is available at www.springerlink.com","keywords":"Siamese Neural Network, Predictive Maintenance, Expert Knowledge, 2D Convolution, Time Series Similarity","abstract":"The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..","bibtex":"@inproceedings{klein_SiameseNN4PredM_2020,\n\ttitle = {{Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance}},\n\tauthor = {Patrick Klein and Niklas Weingarz and Ralph Bergmann},\n\tyear = 2020,\n\tyear = 2020,\n\tbooktitle = {{IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning}},\n\tpublisher = {Springer International Publishing.},\n\tseries = {Communications in Computer and Information Science},\n\tvolume = 1325,\n\tdoi = {10.1007/978-3-030-66770-2\\_6},\n\tisbn = {978-3-030-66769-6},\n\turl = {http://www.wi2.uni-trier.de/shared/publications/2020_ECML-IoTStreams_SiameseNeuralNetwork-for-PredictiveMaintenance_Preprint.pdf},\n\tnote = {The original publication is available at www.springerlink.com},\n\t%pages = {1--16},\n\tkeywords = {{Siamese Neural Network, Predictive Maintenance, Expert Knowledge, 2D Convolution, Time Series Similarity}},\n\tabstract = {The data provided by cyber-physical production systems (CPPSs) to monitor their condition via data-driven predictive maintenance is often high dimensional and only a few fault and failure (FaF) examples are available. These FaFs can usually be detected in a (small) localized subset of data streams, whereas the use of all data streams induces noise that could negatively affect the training and prediction performance. In addition, a CPPS often consists of multiple similar units that generate comparable data streams and show similar failure modes. However, existing approaches for learning a similarity measure generally do not consider these two aspects. For this reason, we propose two approaches for integrating expert knowledge about class or failure mode dependent attributes into siamese neural networks (SNN). Additionally, we present an attribute-wise encoding of time series based on 2D convolutions. This enables that learned knowledge in the form of filters is shared between similar data streams, which would not be possible with conventional 1D convolutions due to their spatial focus. We evaluate our approaches against state-of-the-art time series similarity measures such as dynamic time warping, NeuralWarp, as well as a feature-based representation approach. Our results show that the integration of expert knowledge is advantageous and combined with the novel SNN architecture it is possible to achieve the best performance compared to the other investigated methods..}\n}\n","author_short":["Klein, P.","Weingarz, N.","Bergmann, R."],"key":"klein_SiameseNN4PredM_2020","id":"klein_SiameseNN4PredM_2020","bibbaseid":"klein-weingarz-bergmann-enhancingsiameseneuralnetworksthroughexpertknowledgeforpredictivemaintenance-2020","role":"author","urls":{"Paper":"http://www.wi2.uni-trier.de/shared/publications/2020_ECML-IoTStreams_SiameseNeuralNetwork-for-PredictiveMaintenance_Preprint.pdf"},"keyword":["Siamese Neural Network","Predictive Maintenance","Expert Knowledge","2D Convolution","Time Series Similarity"],"metadata":{"authorlinks":{"bergmann, r":"https://bibbase.org/show?bib=https%3A%2F%2Fweb.wi2.uni-trier.de%2Fpublications%2FWI2Publikationen_IoT.bib&group0=year&css=https%3A%2F%2Fweb.wi2.uni-trier.de%2Fpublications%2FWI2.css"}},"downloads":45},"bibtype":"inproceedings","biburl":"https://web.wi2.uni-trier.de/publications/WI2Publikationen.bib","creationDate":"2020-09-15T13:37:52.469Z","downloads":45,"keywords":["siamese neural network","predictive maintenance","expert knowledge","2d convolution","time series similarity"],"search_terms":["enhancing","siamese","neural","networks","through","expert","knowledge","predictive","maintenance","klein","weingarz","bergmann"],"title":"Enhancing Siamese Neural Networks through Expert Knowledge for Predictive Maintenance","year":2020,"dataSources":["HPSXnFwLpXtbkmbx7","CnwPa99ZchEF4SZgh","vtdjwAo6eNiqRLfnG","nZxfXH3fRFhwWejKL","MSp3DzP4ToPojqkFy","Td7BJ334QwxWK4vLW"]}