Publications

Full Papers

Quantifying Data Leakage in Failure Prediction Tasks. Grillmeyer, Daniel; Hadry, Marius; Lesch, Veronika; Borst, Vanessa; Leppich, Robert; Bauer, Andr’e}; Kounev, Samuel; in Proceedings of the 16th ACM/SPEC International Conference on Performance Engineering (2025). 334–346. Association for Computing Machinery, Toronto ON, Canada.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
- [ DOI ]
With the ever increasing importance of cloud computing and a strong focus on reliable data centers, a high amount of research has been done on failure prediction for hard disk drives. The collection of monitoring data, such as SMART statistics (Self-Monitoring, Analysis, and Reporting Technology) from operational HDDs, enables operators to obtain predictions about the expected remaining useful life. Numerous methods for HDD failure prediction have been published in recent years, and their evaluation has shown decent results. However, a naive splitting into training and test sets can lead to data leakage and, thus, over-optimistic results that cannot be achieved on the data of scientific interest. In this paper, we propose a novel data leakage measure for quantifying the amount of data leakage in training and test datasets. Further, we define four splitting techniques and evaluate our measure in terms of the performance optimism of classification models with respect to these different splitting strategies. Our results consistently show that splitting techniques prone to data leakage induce an overestimation of predictive performance. Overall, we were able to show the usefulness of the defined data leakage measure, as well as its connection with different splitting techniques and the performance optimism of prediction models.

@inproceedings{10.1145/3676151.3719368, abstract = {With the ever increasing importance of cloud computing and a strong focus on reliable data centers, a high amount of research has been done on failure prediction for hard disk drives. The collection of monitoring data, such as SMART statistics (Self-Monitoring, Analysis, and Reporting Technology) from operational HDDs, enables operators to obtain predictions about the expected remaining useful life. Numerous methods for HDD failure prediction have been published in recent years, and their evaluation has shown decent results. However, a naive splitting into training and test sets can lead to data leakage and, thus, over-optimistic results that cannot be achieved on the data of scientific interest. In this paper, we propose a novel data leakage measure for quantifying the amount of data leakage in training and test datasets. Further, we define four splitting techniques and evaluate our measure in terms of the performance optimism of classification models with respect to these different splitting strategies. Our results consistently show that splitting techniques prone to data leakage induce an overestimation of predictive performance. Overall, we were able to show the usefulness of the defined data leakage measure, as well as its connection with different splitting techniques and the performance optimism of prediction models.}, address = {New York, NY, USA}, author = {Grillmeyer, Daniel and Hadry, Marius and Lesch, Veronika and Borst, Vanessa and Leppich, Robert and Bauer, Andr\'{e} and Kounev, Samuel}, booktitle = {Proceedings of the 16th ACM/SPEC International Conference on Performance Engineering}, keywords = {descartes}, pages = {334–346}, publisher = {Association for Computing Machinery}, series = {ICPE '25}, title = {Quantifying Data Leakage in Failure Prediction Tasks}, year = 2025 }
MACRO: A Multi-Head Attentional Convolutional Recurrent Network for the Classification of Co-Occurring Diseases in 12-Lead ECGs. Borst, Vanessa; Leppich, Robert; Kounev, Samuel; in 2024 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI) (2024). IEEE.

Acceptance Rate: 24.4%
- [ BibTeX ]
- [ URL ]
@inproceedings{borst2024MACRO, author = {Borst, Vanessa and Leppich, Robert and Kounev, Samuel}, booktitle = {2024 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI)}, keywords = {t_interdisciplinary}, note = {Acceptance Rate: 24.4%}, organization = {IEEE}, title = {{MACRO}: A Multi-Head Attentional Convolutional Recurrent Network for the Classification of Co-Occurring Diseases in 12-Lead {ECGs}}, year = 2024 }
An Empirical Study of Container Image Configurations and Their Impact on Start Times. Straesser, Martin; Bauer, André; Leppich, Robert; Herbst, Nikolas; Chard, Kyle; Foster, Ian; Kounev, Samuel; in 2023 IEEE/ACM 23rd International Symposium on Cluster, Cloud and Internet Computing (CCGrid) (2023). 94–105. IEEE.
- [ BibTeX ]
- [ URL ]
- [ DOI ]
- [ Download ]
@inproceedings{straesser2023starttime, author = {Straesser, Martin and Bauer, André and Leppich, Robert and Herbst, Nikolas and Chard, Kyle and Foster, Ian and Kounev, Samuel}, booktitle = {2023 IEEE/ACM 23rd International Symposium on Cluster, Cloud and Internet Computing (CCGrid)}, keywords = {empirical_study}, pages = {94-105}, publisher = {IEEE}, title = {An Empirical Study of Container Image Configurations and Their Impact on Start Times}, year = 2023 }
Retrospektive Analyse ausgew"a}hlter Belastungs-und Beanspruchungsparameter einer professionellen Fuss}ball-mannschaft vor und w"a}hrend coronabedingter Kontakt-beschr"a}nkungen in der Spielsaison 2019/20. Kunz, Philipp; D"u}king, Peter; Leppich, Robert; Lachmann, Jonas; Sperlich, Billy; in Leipziger Sportwissenschaftliche Beitr"a}ge Jahrgang 63 (2022 Heft 1 (2022). 75. Lehmanns Media.
- [ BibTeX ]
@article{kunz2022retrospektive, author = {Kunz, Philipp and D{\"u}king, Peter and Leppich, Robert and Lachmann, Jonas and Sperlich, Billy}, journal = {Leipziger Sportwissenschaftliche Beitr{\"a}ge Jahrgang 63 (2022 Heft 1}, keywords = {t_interdisciplinary}, pages = 75, publisher = {Lehmanns Media}, title = {Retrospektive Analyse ausgew{\"a}hlter Belastungs-und Beanspruchungsparameter einer professionellen Fu{\ss}ball-mannschaft vor und w{\"a}hrend coronabedingter Kontakt-beschr{\"a}nkungen in der Spielsaison 2019/20}, year = 2022 }

Journal and Magazine Articles

Telling Fortunes? Evaluation of Traffic Forecasting Models using Traffic and Context Features. Hadry, Marius; Bauer, André; Leppich, Robert; Lesch, Veronika; Kounev, Samuel; in Applied intelligence (2025). 55(755) Springer.
- [ BibTeX ]
- [ URL ]
- [ DOI ]
@article{hadry2025telling, author = {Hadry, Marius and Bauer, André and Leppich, Robert and Lesch, Veronika and Kounev, Samuel}, journal = {Applied intelligence}, keywords = {t_journalmagazine}, month = {06}, number = 755, publisher = {Springer}, title = {Telling Fortunes? Evaluation of Traffic Forecasting Models using Traffic and Context Features}, volume = 55, year = 2025 }
Thinking in Categories: A Survey on Assessing the Quality for Time Series Synthesis. Stenger, Michael; Bauer, Andr\’{e}; Prantl, Thomas; Leppich, Robert; Hudson, Nathaniel; Chard, Kyle; Foster, Ian; Kounev, Samuel; in ACM J. Data and Information Quality (2024). 16(2) Association for Computing Machinery, New York, NY, USA.
- [ Abstract ]
- [ BibTeX ]
- [ URL ]
Time series data are widely used and provide a wealth of information for countless applications. However, some applications are faced with a limited amount of data, or the data cannot be used due to confidentiality concerns. To overcome these obstacles, time series can be generated synthetically. For example, electrocardiograms can be synthesized to make them available for building models to predict conditions such as cardiac arrhythmia without leaking patient information. Although many different approaches to time series synthesis have been proposed, evaluating the quality of synthetic time series data poses unique challenges and remains an open problem, as there is a lack of a clear definition of what constitutes a “good” synthesis. To this end, we present a comprehensive literature survey to identify different aspects of synthesis quality and their relationships. Based on this, we propose a definition of synthesis quality and a systematic evaluation procedure for assessing it. With this work, we aim to provide a common language and criteria for evaluating synthetic time series data. Our goal is to promote more rigorous and reproducible research in time series synthesis by enabling researchers and practitioners to generate high-quality synthetic time series data.

@article{10.1145/3666006, abstract = {Time series data are widely used and provide a wealth of information for countless applications. However, some applications are faced with a limited amount of data, or the data cannot be used due to confidentiality concerns. To overcome these obstacles, time series can be generated synthetically. For example, electrocardiograms can be synthesized to make them available for building models to predict conditions such as cardiac arrhythmia without leaking patient information. Although many different approaches to time series synthesis have been proposed, evaluating the quality of synthetic time series data poses unique challenges and remains an open problem, as there is a lack of a clear definition of what constitutes a “good” synthesis. To this end, we present a comprehensive literature survey to identify different aspects of synthesis quality and their relationships. Based on this, we propose a definition of synthesis quality and a systematic evaluation procedure for assessing it. With this work, we aim to provide a common language and criteria for evaluating synthetic time series data. Our goal is to promote more rigorous and reproducible research in time series synthesis by enabling researchers and practitioners to generate high-quality synthetic time series data.}, address = {New York, NY, USA}, author = {Stenger, Michael and Bauer, Andr\'{e} and Prantl, Thomas and Leppich, Robert and Hudson, Nathaniel and Chard, Kyle and Foster, Ian and Kounev, Samuel}, journal = {ACM J. Data and Information Quality}, keywords = {t_journalmagazine}, month = {06}, number = 2, publisher = {Association for Computing Machinery}, title = {Thinking in Categories: A Survey on Assessing the Quality for Time Series Synthesis}, volume = 16, year = 2024 }
Evaluation is key: a survey on evaluation measures for synthetic time series. Stenger, Michael; Leppich, Robert; Foster, Ian; Kounev, Samuel; Bauer, Andr{\’e}; in Journal of Big Data (2024). 11 66.
- [ BibTeX ]
@article{Stenger2024, author = {Stenger, Michael and Leppich, Robert and Foster, Ian and Kounev, Samuel and Bauer, Andr{\'e}}, journal = {Journal of Big Data}, keywords = {t_journalmagazine}, month = {05}, pages = 66, title = {Evaluation is key: a survey on evaluation measures for synthetic time series}, volume = 11, year = 2024 }
Comprehensive Exploration of Synthetic Data Generation: A Survey. Bauer, André; and Trapp, Simon; and Stenger, Michael; and Leppich, Robert; and Kounev, Samuel; and Leznik, Mark; and Chard, Kyle; and Foster, Ian; in arXiv preprint (2024). 103.

https://doi.org/10.48550/arXiv.2401.02524
- [ BibTeX ]
@article{bauer2024comprehensive, author = {Bauer, André and and Trapp, Simon and and Stenger, Michael and and Leppich, Robert and and Kounev, Samuel and and Leznik, Mark and and Chard, Kyle and and Foster, Ian}, journal = {arXiv preprint}, keywords = {t_journalmagazine}, note = {https://doi.org/10.48550/arXiv.2401.02524}, pages = 103, title = {Comprehensive Exploration of Synthetic Data Generation: A Survey}, year = 2024 }
Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery. Müller, Konstantin; Leppich, Robert; Geiß, Christian; Borst, Vanessa; Pelizari, Patrick Aravena; Kounev, Samuel; in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (2023). 1–14.
- [ Abstract ]
- [ BibTeX ]
In recent history, normalized digital surface models (nDSMs) have been constantly gaining importance as a means to solve large-scale geographic problems. High-resolution surface models are precious, as they can provide detailed information for a specific area. However, measurements with a high resolution are time-consuming and costly. Only a few approaches exist to create high-resolution normalized digital surface models for extensive areas. This paper explores approaches to extract high-resolution nDSMs from low-resolution Sentinel-2 data, allowing us to derive large-scale models. We thereby utilize the advantages of Sentinel 2 being open access, having global coverage, and providing steady updates through a high repetition rate. Several deep-learning models are trained to overcome the gap in producing high-resolution surface maps from low-resolution input data. With U-Net as a base architecture, we extend the capabilities of our model by integrating tailored multiscale encoders with differently sized kernels in the convolution as well as conformed self-attention inside the skip connection gates. Using pixel-wise regression, our U-Net base models can achieve a mean height error of approximately two meters. Moreover, through our enhancements to the model architecture, we reduce the model error by more than seven percent.

@article{10189905, abstract = {In recent history, normalized digital surface models (nDSMs) have been constantly gaining importance as a means to solve large-scale geographic problems. High-resolution surface models are precious, as they can provide detailed information for a specific area. However, measurements with a high resolution are time-consuming and costly. Only a few approaches exist to create high-resolution normalized digital surface models for extensive areas. This paper explores approaches to extract high-resolution nDSMs from low-resolution Sentinel-2 data, allowing us to derive large-scale models. We thereby utilize the advantages of Sentinel 2 being open access, having global coverage, and providing steady updates through a high repetition rate. Several deep-learning models are trained to overcome the gap in producing high-resolution surface maps from low-resolution input data. With U-Net as a base architecture, we extend the capabilities of our model by integrating tailored multiscale encoders with differently sized kernels in the convolution as well as conformed self-attention inside the skip connection gates. Using pixel-wise regression, our U-Net base models can achieve a mean height error of approximately two meters. Moreover, through our enhancements to the model architecture, we reduce the model error by more than seven percent.}, author = {Müller, Konstantin and Leppich, Robert and Geiß, Christian and Borst, Vanessa and Pelizari, Patrick Aravena and Kounev, Samuel}, journal = {IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing}, keywords = {t_journalmagazine}, pages = {1-14}, title = {Deep Neural Network Regression for Normalized Digital Surface Model Generation with Sentinel-2 Imagery}, year = 2023 }
Telescope: An Automated Hybrid Forecasting Approach on a Level-Playing Field. Bauer, André; Leznik, Mark; Stenger, Michael; Leppich, Robert; Herbst, Nikolas; Kounev, Samuel; Foster, Ian; (2023).
- [ BibTeX ]
- [ Download ]
@misc{bauer2023telescope, author = {Bauer, André and Leznik, Mark and Stenger, Michael and Leppich, Robert and Herbst, Nikolas and Kounev, Samuel and Foster, Ian}, keywords = {t_journalmagazine}, title = {Telescope: An Automated Hybrid Forecasting Approach on a Level-Playing Field}, year = 2023 }
Accuracy and Systematic Biases of Heart Rate Measurements by Consumer-Grade Fitness Trackers in Postoperative Patients: Prospective Clinical Trial. Helmer, Philipp; Hottenrott, Sebastian; Rodemers, Philipp; Leppich, Robert; Helwich, Maja; Pryss, R{\"u}diger; Kranke, Peter; Meybohm, Patrick; Winkler, Bernd E; Sammeth, Michael; in Journal of Medical Internet Research (2022). 24(12) e42359. JMIR Publications Toronto, Canada.
- [ BibTeX ]
@article{helmer2022accuracy, author = {Helmer, Philipp and Hottenrott, Sebastian and Rodemers, Philipp and Leppich, Robert and Helwich, Maja and Pryss, R{\"u}diger and Kranke, Peter and Meybohm, Patrick and Winkler, Bernd E and Sammeth, Michael}, journal = {Journal of Medical Internet Research}, keywords = {t_journalmagazine}, number = 12, pages = {e42359}, publisher = {JMIR Publications Toronto, Canada}, title = {Accuracy and Systematic Biases of Heart Rate Measurements by Consumer-Grade Fitness Trackers in Postoperative Patients: Prospective Clinical Trial}, volume = 24, year = 2022 }
Retrospective Analysis of Training Intensity Distribution Based on Race Pace Versus Physiological Benchmarks in Highly Trained Sprint Kayakers. Matzka, Manuel; Leppich, Robert; Sperlich, Billy; Zinner, Christoph; in Sports Medicine-Open (2022). 8(1) 1–12. SpringerOpen.
- [ BibTeX ]
- [ Download ]
@article{matzka2022retrospective, author = {Matzka, Manuel and Leppich, Robert and Sperlich, Billy and Zinner, Christoph}, journal = {Sports Medicine-Open}, keywords = {t_journalmagazine}, number = 1, pages = {1–12}, publisher = {SpringerOpen}, title = {Retrospective Analysis of Training Intensity Distribution Based on Race Pace Versus Physiological Benchmarks in Highly Trained Sprint Kayakers}, volume = 8, year = 2022 }
The Relationship Between the Distribution of Training Intensity and Performance of Kayak and Canoe Sprinters: A Retrospective Observational Analysis of One Season of Competition. Matzka, Manuel; Leppich, Robert; Holmberg, Hans-Christer; Sperlich, Billy; Zinner, Christoph; in Frontiers in Sports and Active Living (2022). 3 788108. Frontiers Media SA.
- [ BibTeX ]
@article{matzka2022relationship, author = {Matzka, Manuel and Leppich, Robert and Holmberg, Hans-Christer and Sperlich, Billy and Zinner, Christoph}, journal = {Frontiers in Sports and Active Living}, keywords = {t_journalmagazine}, pages = 788108, publisher = {Frontiers Media SA}, title = {The Relationship Between the Distribution of Training Intensity and Performance of Kayak and Canoe Sprinters: A Retrospective Observational Analysis of One Season of Competition}, volume = 3, year = 2022 }
Establishing a cardiac training group for patients with heart failure: the “HIP-in-W{\"u}rzburg” study. G{\"u}der, G{\"u}lmisal; Wilkesmann, Joana; Scholz, Nina; Leppich, Robert; D{\"u}king, Peter; Sperlich, Billy; Rost, Christian; Frantz, Stefan; Morbach, Caroline; Sahiti, Floran; others; in Clinical Research in Cardiology (2021). 1–10. Springer.
- [ BibTeX ]
@article{guder2021establishing, author = {G{\"u}der, G{\"u}lmisal and Wilkesmann, Joana and Scholz, Nina and Leppich, Robert and D{\"u}king, Peter and Sperlich, Billy and Rost, Christian and Frantz, Stefan and Morbach, Caroline and Sahiti, Floran and others}, journal = {Clinical Research in Cardiology}, keywords = {t_journalmagazine}, pages = {1–10}, publisher = {Springer}, title = {Establishing a cardiac training group for patients with heart failure: the “HIP-in-W{\"u}rzburg” study}, year = 2021 }
The integration of training and off-training activities substantially alters training volume and load analysis in elite rowers. Treff, Gunnar; Leppich, Robert; Winkert, Kay; Steinacker, J{\"u}rgen M; Mayer, Benjamin; Sperlich, Billy; in Scientific reports (2021). 11(1) 1–10. Nature Publishing Group.
- [ BibTeX ]
@article{treff2021integration, author = {Treff, Gunnar and Leppich, Robert and Winkert, Kay and Steinacker, J{\"u}rgen M and Mayer, Benjamin and Sperlich, Billy}, journal = {Scientific reports}, keywords = {t_journalmagazine}, number = 1, pages = {1–10}, publisher = {Nature Publishing Group}, title = {The integration of training and off-training activities substantially alters training volume and load analysis in elite rowers}, volume = 11, year = 2021 }
Intra-individual physiological response of recreational runners to different training mesocycles: a randomized cross-over study. D{\"u}king, Peter; Holmberg, Hans-Christer; Kunz, Philipp; Leppich, Robert; Sperlich, Billy; in European Journal of Applied Physiology (2020). 1–9. Springer.
- [ BibTeX ]
@article{duking2020intra, author = {D{\"u}king, Peter and Holmberg, Hans-Christer and Kunz, Philipp and Leppich, Robert and Sperlich, Billy}, journal = {European Journal of Applied Physiology}, keywords = {t_journalmagazine}, pages = {1–9}, publisher = {Springer}, title = {Intra-individual physiological response of recreational runners to different training mesocycles: a randomized cross-over study}, year = 2020 }
The impact of the German strategy for containment of Coronavirus SARS-CoV-2 on the training characteristics, physical activity, sleep of highly trained kayakers and canoeists: A retrospective observational study. Zinner, Christoph; Matzka, Manuel; Leppich, Robert; Kounev, Samuel; Holmberg, HC; Sperlich, Billy; in Frontiers in Sports and Active Living (2020). 2 127. Frontiers.
- [ BibTeX ]
@article{zinner2020impact, author = {Zinner, Christoph and Matzka, Manuel and Leppich, Robert and Kounev, Samuel and Holmberg, HC and Sperlich, Billy}, journal = {Frontiers in Sports and Active Living}, keywords = {covid19}, pages = 127, publisher = {Frontiers}, title = {The impact of the German strategy for containment of Coronavirus SARS-CoV-2 on the training characteristics, physical activity, sleep of highly trained kayakers and canoeists: A retrospective observational study}, volume = 2, year = 2020 }
A 4-Week intervention involving mobile-based daily 6-minute micro-sessions of functional high-intensity circuit training improves strength and quality of life, but not cardio-respiratory fitness of young untrained adults. Sperlich, Billy; Hahn, Lea-Sofie; Edel, Antonia; Behr, Tino; Helmprobst, Julian; Leppich, Robert; Wallmann-Sperlich, Birgit; Holmberg, Hans-Christer; in Frontiers in physiology (2018). 9 423. Frontiers.
- [ BibTeX ]
- [ Download ]
@article{sperlich20184, author = {Sperlich, Billy and Hahn, Lea-Sofie and Edel, Antonia and Behr, Tino and Helmprobst, Julian and Leppich, Robert and Wallmann-Sperlich, Birgit and Holmberg, Hans-Christer}, journal = {Frontiers in physiology}, keywords = {t_journalmagazine}, pages = 423, publisher = {Frontiers}, title = {A 4-Week intervention involving mobile-based daily 6-minute micro-sessions of functional high-intensity circuit training improves strength and quality of life, but not cardio-respiratory fitness of young untrained adults}, volume = 9, year = 2018 }

Talks

Artificial Intelligence in Medicine - From Data Collection to Prediction. Züfle, Marwin; Leppich, Robert; (2019, November).
- [ BibTeX ]
@misc{talk-ZuLe, author = {Z{\"u}fle, Marwin and Leppich, Robert}, howpublished = {Invited Talk at Klinikum N{\"u}rnberg}, keywords = {Sports}, month = 11, title = {{Artificial Intelligence in Medicine - From Data Collection to Prediction}}, year = 2019 }

Bildnachweise