Philipp Jaß
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Philipp Jaß has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Carsten Thomas
School of Engineering–Energy and Information, HTW Berlin, 10313 Berlin, Germany
2 shared publications Carsten Thomas
HTW Berlin, Berlin, Deutschland
1 shared publicationHamzih Abukhashab
HTW Berlin, Berlin, Deutschland
1 shared publicationHarald Schülzke
Samoconsult GmbH, Berlin, Deutschland
1 shared publicationInes Fey
Samoconsult GmbH, Berlin, Deutschland
1 shared publicationMichael Weber
DResearch Fahrzeugelektronik GmbH, Berlin, Deutschland
1 shared publication
Latest Publications
- 16 November 2025
- Published by Springer Nature
- p. 53-57
Using N-Version Architectures for Railway Segmentation with Deep Neural Networks
- 16 November 2025
- Published by Springer Nature
- p. 53-57
Using N-Version Architectures for Railway Segmentation with Deep Neural Networks
- Open Access
- 26 May 2025
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 7 (2) , 49
Using N-Version Architectures for Railway Segmentation with Deep Neural Networks
Autonomous trains require reliable and accurate environmental perception to take over safety-critical tasks from the driver. This paper investigates the application of N-version architectures to rail track detection using Deep Neural Networks (DNNs) as a means to improve the safety of machine learning (ML)-enabled perception systems. We combine three different neural network architectures (WCID, VGG16-UNet, MobileNet–SegNet) in a 3M1I configuration. In this configuration, we apply two fusion methods to increase accuracy and to enable error detection: Maximum Confidence Voting (MCV), combining the DNN predictions at the image level, and Pixel Majority Voting (PMV), a novel approach for combining the predictions at the pixel level. In addition, we implement a new method for evaluating and combining prediction confidence values in the N-version architecture during runtime. We adjust the overall prediction confidence according to the conformity of all individual predictions, which is not possible with an individual network. Our results show that the N-version architecture not only enables a detection of erroneous predictions by utilizing those adjusted confidence values, but it can also partially improve the predictions by using the PMV combination algorithm. This work emphasizes the importance of model diversity and appropriate thresholds for an accurate assessment of prediction safety. These approaches can significantly improve the practical applicability of ML-based systems in safety-critical domains such as rail transportation.
- 26 May 2025
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 7 (2) , 49
Open AccessUsing N-Version Architectures for Railway Segmentation with Deep Neural Networks
Autonomous trains require reliable and accurate environmental perception to take over safety-critical tasks from the driver. This paper investigates the application of N-version architectures to rail track detection using Deep Neural Networks (DNNs) as a means to improve the safety of machine learning (ML)-enabled perception systems. We combine three different neural network architectures (WCID, VGG16-UNet, MobileNet–SegNet) in a 3M1I configuration. In this configuration, we apply two fusion methods to increase accuracy and to enable error detection: Maximum Confidence Voting (MCV), combining the DNN predictions at the image level, and Pixel Majority Voting (PMV), a novel approach for combining the predictions at the pixel level. In addition, we implement a new method for evaluating and combining prediction confidence values in the N-version architecture during runtime. We adjust the overall prediction confidence according to the conformity of all individual predictions, which is not possible with an individual network. Our results show that the N-version architecture not only enables a detection of erroneous predictions by utilizing those adjusted confidence values, but it can also partially improve the predictions by using the PMV combination algorithm. This work emphasizes the importance of model diversity and appropriate thresholds for an accurate assessment of prediction safety. These approaches can significantly improve the practical applicability of ML-based systems in safety-critical domains such as rail transportation.
- 25 July 2023
- Published by Springer Nature in Datenschutz und Datensicherheit - DuD
- Vol. 47 (8) , 483-486
CertML: Initial Steps Towards Using N-Version Neural Networks for Improving AI Safety
- 25 July 2023
- Published by Springer Nature in Datenschutz und Datensicherheit - DuD
- Vol. 47 (8) , 483-486
CertML: Initial Steps Towards Using N-Version Neural Networks for Improving AI Safety