@article{Heusinger, author = {Heusinger, Moritz}, title = {Dimensionality reduction in the context of dynamic social media data streams}, series = {Evolving Systems}, journal = {Evolving Systems}, issn = {1868-6486}, abstract = {In recent years social media became an important part of everyday life for many people. A big challenge of social media is, to find posts, that are interesting for the user. Many social networks like Twitter handle this problem with so-called hashtags. A user can label his own Tweet (post) with a hashtag, while other users can search for posts containing a specified hashtag. But what about finding posts which are not labeled by the creator? We provide a way of completing hashtags for unlabeled posts using classification on a novel real-world Twitter data stream. New posts will be created every second, thus this context fits perfectly for non-stationary data analysis. Our goal is to show, how labels (hashtags) of social media posts can be predicted by stream classifiers. In particular, we employ random projection (RP) as a preprocessing step in calculating streaming models. Also, we provide a novel real-world data set for streaming analysis called NSDQ with a comprehensive data description. We show that this dataset is a real challenge for state-of-the-art stream classifiers. While RP has been widely used and evaluated in stationary data analysis scenarios, non-stationary environments are not well analyzed. In this paper, we provide a use case of RP on real-world streaming data, especially on NSDQ dataset. We discuss why RP can be used in this scenario and how it can handle stream-specific situations like concept drift. We also provide experiments with RP on streaming data, using state-of-the-art stream classifiers like adaptive random forest and concept drift detectors. Additionally, we experimentally evaluate an online principal component analysis (PCA) approach in the same fashion as we do for RP. To obtain higher dimensional synthetic streams, we use random Fourier features (RFF) in an online manner which allows us, to increase the number of dimensions of low dimensional streams.}, language = {en} } @article{Heusinger, author = {Heusinger, Moritz}, title = {Passive concept drift handling via variations of learning vector quantization}, series = {Neural Computing and Applications}, volume = {34}, journal = {Neural Computing and Applications}, issn = {1433-3058}, pages = {89 -- 100}, abstract = {Concept drift is a change of the underlying data distribution which occurs especially with streaming data. Besides other challenges in the field of streaming data classification, concept drift has to be addressed to obtain reliable predictions. Robust Soft Learning Vector Quantization as well as Generalized Learning Vector Quantization has already shown good performance in traditional settings and is modified in this work to handle streaming data. Further, momentum-based stochastic gradient descent techniques are applied to tackle concept drift passively due to increased learning capabilities. The proposed work is tested against common benchmark algorithms and streaming data in the field and achieved promising results.}, language = {en} } @article{RaabHeusingerSchleif, author = {Raab, Christoph and Heusinger, Moritz and Schleif, Frank-Michael}, title = {Reactive Soft Prototype Computing for Concept Drift Streams}, series = {Neurocomputing}, volume = {416}, journal = {Neurocomputing}, doi = {10.1016/j.neucom.2019.11.111}, pages = {340 -- 351}, language = {en} } @inproceedings{HeusingerSchleif, author = {Heusinger, Moritz and Schleif, Frank-Michael}, title = {Classification in Non-stationary Environments Using Coresets over Sliding Windows}, series = {Advances in Computational Intelligence - 16th International Work-Conference on Artificial Neural Networks, IWANN 2021, Virtual Event, June 16-18, 2021, Proceedings, Part I}, booktitle = {Advances in Computational Intelligence - 16th International Work-Conference on Artificial Neural Networks, IWANN 2021, Virtual Event, June 16-18, 2021, Proceedings, Part I}, editor = {Rojas, Ignacio and Joya, Gonzalo and Catal{\`a}, Andreu}, doi = {10.1007/978-3-030-85030-2_11}, pages = {126 -- 137}, language = {en} } @inproceedings{HeusingerSchleif, author = {Heusinger, Moritz and Schleif, Frank-Michael}, title = {A Streaming Approach to the Core Vector Machine}, series = {Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Zakopane, Poland, June 19-23, 2022, Proceedings, Part II}, booktitle = {Artificial Intelligence and Soft Computing - 21st International Conference, ICAISC 2022, Zakopane, Poland, June 19-23, 2022, Proceedings, Part II}, editor = {Rutkowski, Leszek and Scherer, Rafal and Korytkowski, Marcin and Pedrycz, Witold and Tadeusiewicz, Ryszard and Zurada, Jacek M.}, doi = {10.1007/978-3-031-23480-4_8}, pages = {91 -- 101}, language = {en} } @inproceedings{HeusingerSchleif, author = {Heusinger, Moritz and Schleif, Frank-Michael}, title = {Random Projection in the Presence of Concept Drift in Supervised Environments}, series = {Artificial Intelligence and Soft Computing - 19th International Conference, ICAISC 2020, Zakopane, Poland, October 12-14, 2020, Proceedings, Part I}, booktitle = {Artificial Intelligence and Soft Computing - 19th International Conference, ICAISC 2020, Zakopane, Poland, October 12-14, 2020, Proceedings, Part I}, editor = {Rutkowski, Leszek and Scherer, Rafal and Korytkowski, Marcin and Pedrycz, Witold and Tadeusiewicz, Ryszard and Zurada, Jacek M.}, doi = {10.1007/978-3-030-61401-0_48}, pages = {514 -- 524}, language = {en} } @inproceedings{HeusingerRaabSchleif, author = {Heusinger, Moritz and Raab, Christoph and Schleif, Frank-Michael}, title = {Passive Concept Drift Handling via Momentum Based Robust Soft Learning Vector Quantization}, series = {Advances in Self-Organizing Maps, Learning Vector Quantization, Clustering and Data Visualization - Proceedings of the 13th International Workshop, WSOM+ 2019, Barcelona, Spain, June 26-28, 2019}, booktitle = {Advances in Self-Organizing Maps, Learning Vector Quantization, Clustering and Data Visualization - Proceedings of the 13th International Workshop, WSOM+ 2019, Barcelona, Spain, June 26-28, 2019}, editor = {Vellido, Alfredo and Gibert, Karina and Angulo, Cecilio and David Mart{\i}́n-Guerrero, Jos{\´e}}, doi = {10.1007/978-3-030-19642-4_20}, pages = {200 -- 209}, language = {en} } @inproceedings{RaabHeusingerSchleif, author = {Raab, Christoph and Heusinger, Moritz and Schleif, Frank-Michael}, title = {Reactive Soft Prototype Computing for frequent reoccurring Concept Drift}, series = {27th European Symposium on Artificial Neural Networks, ESANN 2019, Bruges, Belgium, April 24-26, 2019}, booktitle = {27th European Symposium on Artificial Neural Networks, ESANN 2019, Bruges, Belgium, April 24-26, 2019}, language = {en} } @inproceedings{HeusingerSchleif, author = {Heusinger, Moritz and Schleif, Frank-Michael}, title = {Random Projection in supervised non-stationary environments}, series = {28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2020, Bruges, Belgium, October 2-4, 2020}, booktitle = {28th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2020, Bruges, Belgium, October 2-4, 2020}, pages = {405 -- 410}, language = {en} } @inproceedings{HeusingerSchleif, author = {Heusinger, Moritz and Schleif, Frank-Michael}, title = {Reactive Concept Drift Detection Using Coresets Over Sliding Windows}, series = {2020 IEEE Symposium Series on Computational Intelligence, SSCI 2020, Canberra, Australia, December 1-4, 2020}, booktitle = {2020 IEEE Symposium Series on Computational Intelligence, SSCI 2020, Canberra, Australia, December 1-4, 2020}, doi = {10.1109/SSCI47803.2020.9308521}, pages = {1350 -- 1355}, language = {en} } @inproceedings{HeusingerRaabRossietal., author = {Heusinger, Moritz and Raab, Christoph and Rossi, Fabrice and Schleif, Frank-Michael}, title = {Federated Learning - Methods, Applications and beyond}, series = {29th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2021, Online event (Bruges, Belgium), October 6-8, 2021}, booktitle = {29th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning, ESANN 2021, Online event (Bruges, Belgium), October 6-8, 2021}, doi = {10.14428/esann/2021.ES2021-4}, language = {en} } @inproceedings{HeusingerRaabSchleif, author = {Heusinger, Moritz and Raab, Christoph and Schleif, Frank-Michael}, title = {Analyzing Dynamic Social Media Data via Random Projection - A New Challenge for Stream Classifiers}, series = {2020 IEEE Conference on Evolving and Adaptive Intelligent Systems, EAIS 2020, Bari, Italy, May 27-29, 2020}, booktitle = {2020 IEEE Conference on Evolving and Adaptive Intelligent Systems, EAIS 2020, Bari, Italy, May 27-29, 2020}, doi = {10.1109/EAIS48028.2020.9122780}, pages = {1 -- 8}, language = {en} } @article{ZieglerDobhanHeusinger, author = {Ziegler, Cedric C. and Dobhan, Alexander and Heusinger, Moritz}, title = {Applications of Neural Networks in Engineer-to-order Environment}, series = {Procedia CIRP}, volume = {112}, journal = {Procedia CIRP}, publisher = {Elsevier}, doi = {10.1016/j.procir.2022.09.052}, pages = {140 -- 145}, abstract = {Machine tool selection and quotation costing have a low level of automation in today's engineer-to-order environments. The decision-making process is based on imprecise human judgment even if all final product characteristics are known. To improve precision and save time, we propose to apply artificial neural networks that are trained on data of already produced products. In particular, we address the selection of several grinding wheels to produce a milling cutter, which is a multiple-criteria decision, using multi-label classification. The quotation costing is a single output regression problem. In both tasks, our results on real-world data show high accuracy.}, language = {en} } @article{MillerBorysenkoHeusingeretal., author = {Miller, Eddi and Borysenko, Vladyslav and Heusinger, Moritz and Niedner, Niklas and Engelmann, Bastian and Schmitt, Jan}, title = {Enhanced Changeover Detection in Industry 4.0 Environments with Machine Learning}, series = {Sensors}, volume = {21}, journal = {Sensors}, number = {17}, pages = {5896 -- 5896}, language = {en} } @article{HofmannEberhardtHeusingeretal., author = {Hofmann, Jan and Eberhardt, Lars and Heusinger, Moritz and Dobhan, Alexander and Engelmann, Bastian and Schleif, Frank-Michael}, title = {Optimierung von Prozessen und Werkzeugmaschinen durch Bereitstellung, Analyse und Soll-Ist-Vergleich von Produktionsdaten}, series = {FHWS Science Journal}, volume = {5}, journal = {FHWS Science Journal}, number = {2}, issn = {2196-6095}, url = {http://nbn-resolving.de/urn:nbn:de:bvb:863-opus-20012}, pages = {135 -- 142}, abstract = {Mit einem Umsatz von 103 Milliarden Euro ist die Metallindustrie eine der gr{\"o}ßten deutschen Industriebranchen. Diese ist von volatilen Marktbedingungen und hohem Wettbewerb gepr{\"a}gt [1][2]. Kleine und mittlere produzierende Unternehmen (sogenannte KMU) sehen zunehmend gravierende Probleme bei der Einhaltung von Lieferterminen bedingt durch hohe Durchlaufzeiten in der Produktion [3]. Neben kaufm{\"a}nnischen Planungssystemen zur Erstellung von Produktionspl{\"a}nen nutzen Unternehmen als Planungsgrundlage weiterhin Excel mit 31 \% und manuelle Prozesse mit 10 \% [4]. Gleiches gilt f{\"u}r Produktwechselvorg{\"a}nge auf Maschinen (R{\"u}sten). Aufgrund dieser Aspekte ist es notwendig, die Rentabilit{\"a}t der KMU in der Metallindustrie zu steigern. Das wird durch effiziente Produktionsplanung und -steuerung, sowie der daraus resultierenden hohen Reaktionsf{\"a}higkeit und Flexibilit{\"a}t realisiert. Daher ist die Produktionsplanung auf die Markt- und Kundenanforderungen und die Anlageneffektivit{\"a}t auf ein hohes und stabiles Niveau auszurichten [5]. Hier bietet die Erfassung von Echtzeitdaten eine ad{\"a}quate Reaktion auf die genannten Anforderungen. Ebenfalls liefert sie großes Potenzial f{\"u}r die Produktionsplanung und -steuerung, um die Disposition und Koordination von Arbeitsauftr{\"a}gen zu optimieren. Weiterhin werden St{\"o}rgr{\"o}ßen oder unvorhergesehene Planungsabweichungen reduziert [4][6]. Zus{\"a}tzlich ist eine erh{\"o}hte Transparenz und Verbesserung menschlicher Entscheidungsprozesse notwendig. Dies kann durch datengetriebene Methoden unterst{\"u}tzt und sichergestellt werden [7]. Ein Ansatz zur Optimierung des Produktionsergebnisses kann durch die Erh{\"o}hung der Anlagenproduktivit{\"a}t selbst realisiert werden. Dazu muss die Verf{\"u}gbarkeit der Anlagen durch Lokalisierung und Reduzierung von Verlusten erh{\"o}ht werden. Die Umr{\"u}stungsprozesse tragen stark negativ zur Verf{\"u}gbarkeit einer Produktion bei. Eine Steigerung der Gesamtanlageneffektivit{\"a}t (overall equipment effectiveness oder kurz OEE) in einer Fertigungsumgebung ist jedoch m{\"o}glich durch eine intelligente Nutzung von Sensordaten mit Techniken wie z. B. Machine Learning (ML).}, language = {de} } @inproceedings{HartmannDobhanEngelmannetal., author = {Hartmann, J{\"u}rgen and Dobhan, Alexander and Engelmann, Bastian and Eberhardt, Lars and Heusinger, Moritz and Raab, C and Schleif, Frank-Michael and T{\"u}rk, M.}, title = {Optimierung von Prozessen und Werkzeugmaschinen durch Bereitstellung, Analyse und Soll-Ist-Vergleich von Produktionsdaten: Digitalkonferenz}, language = {en} } @article{EngelmannSchmittHeusingeretal., author = {Engelmann, Bastian and Schmitt, Anna-Maria and Heusinger, Moritz and Borysenko, Vladyslav and Niedner, Niklas and Schmitt, Jan}, title = {Detecting Changeover Events on Manufacturing Machines with Machine Learning and NC data}, series = {Applied Artificial Intelligence}, journal = {Applied Artificial Intelligence}, publisher = {Taylor \& Francis}, language = {en} }