Florian Kalinke
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Florian Kalinke has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Edouard Fouché
Siemens AG, Munich, Germany
5 shared publications - Klemens Böhm
Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany
4 shared publications - Marco Heyden
Karlsruhe Institute of Technology, Karlsruhe, Germany
2 shared publications - Tobias Fuchs
Karlsruhe Institute of Technology Karlsruhe Germany
2 shared publications Haiko Thiessen
Karlsruhe Institute of Technology (KIT)
1 shared publicationKlemens Böhm
1 shared publication
Latest Publications
- Open Access
- 15 July 2025
- Published by Springer Nature in Machine Learning
- Vol. 114 (8) , 1-26
Robust partial-label learning by leveraging class activation values
Real-world training data is often noisy; for example, human annotators assign conflicting class labels to the same instances. Partial-label learning (PLL) is a weakly supervised learning paradigm that allows training classifiers in this context without manual data cleaning. While state-of-the-art methods have good predictive performance, their predictions are sensitive to high noise levels, out-of-distribution data, and adversarial perturbations. We propose a novel PLL method based on subjective logic, which explicitly represents uncertainty by leveraging the magnitudes of the underlying neural network’s class activation values. Thereby, we effectively incorporate prior knowledge about the class labels by using a novel label weight re-distribution strategy that we prove to be optimal. We empirically show that our method yields more robust predictions in terms of predictive performance under high PLL noise levels, handling out-of-distribution examples, and handling adversarial perturbations on the test instances.
- 15 July 2025
- Published by Springer Nature in Machine Learning
- Vol. 114 (8) , 1-26
Open AccessRobust partial-label learning by leveraging class activation values
Real-world training data is often noisy; for example, human annotators assign conflicting class labels to the same instances. Partial-label learning (PLL) is a weakly supervised learning paradigm that allows training classifiers in this context without manual data cleaning. While state-of-the-art methods have good predictive performance, their predictions are sensitive to high noise levels, out-of-distribution data, and adversarial perturbations. We propose a novel PLL method based on subjective logic, which explicitly represents uncertainty by leveraging the magnitudes of the underlying neural network’s class activation values. Thereby, we effectively incorporate prior knowledge about the class labels by using a novel label weight re-distribution strategy that we prove to be optimal. We empirically show that our method yields more robust predictions in terms of predictive performance under high PLL noise levels, handling out-of-distribution examples, and handling adversarial perturbations on the test instances.
- Open Access
Robust Partial-Label Learning by Leveraging Class Activation Values
Real-world training data is often noisy; for example, human annotators assign conflicting class labels to the same instances. Partial-label learning (PLL) is a weakly supervised learning paradigm that allows training classifiers in this context without manual data cleaning. While state-of-the-art methods have good predictive performance, their predictions are sensitive to high noise levels, out-of-distribution data, and adversarial perturbations. We propose a novel PLL method based on subjective logic, which explicitly represents uncertainty by leveraging the magnitudes of the underlying neural network's class activation values. Thereby, we effectively incorporate prior knowledge about the class labels by using a novel label weight re-distribution strategy that we prove to be optimal. We empirically show that our method yields more robust predictions in terms of predictive performance under high PLL noise levels, handling out-of-distribution examples, and handling adversarial perturbations on the test instances.
Open AccessRobust Partial-Label Learning by Leveraging Class Activation Values
Real-world training data is often noisy; for example, human annotators assign conflicting class labels to the same instances. Partial-label learning (PLL) is a weakly supervised learning paradigm that allows training classifiers in this context without manual data cleaning. While state-of-the-art methods have good predictive performance, their predictions are sensitive to high noise levels, out-of-distribution data, and adversarial perturbations. We propose a novel PLL method based on subjective logic, which explicitly represents uncertainty by leveraging the magnitudes of the underlying neural network's class activation values. Thereby, we effectively incorporate prior knowledge about the class labels by using a novel label weight re-distribution strategy that we prove to be optimal. We empirically show that our method yields more robust predictions in terms of predictive performance under high PLL noise levels, handling out-of-distribution examples, and handling adversarial perturbations on the test instances.
- Open Access
- 9 January 2024
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 38 (3) , 1334-1363
Adaptive Bernstein change detector for high-dimensional data streams
Change detection is of fundamental importance when analyzing data streams. Detecting changes both quickly and accurately enables monitoring and prediction systems to react, e.g., by issuing an alarm or by updating a learning algorithm. However, detecting changes is challenging when observations are high-dimensional. In high-dimensional data, change detectors should not only be able to identify when changes happen, but also in which subspace they occur. Ideally, one should also quantify how severe they are. Our approach, ABCD, has these properties. ABCD learns an encoder-decoder model and monitors its accuracy over a window of adaptive size. ABCD derives a change score based on Bernstein’s inequality to detect deviations in terms of accuracy, which indicate changes. Our experiments demonstrate that ABCD outperforms its best competitor by up to 20% in F1-score on average. It can also accurately estimate changes’ subspace, together with a severity measure that correlates with the ground truth.
- 9 January 2024
- Published by Springer Nature in Data Mining and Knowledge Discovery
- Vol. 38 (3) , 1334-1363
Open AccessAdaptive Bernstein change detector for high-dimensional data streams
Change detection is of fundamental importance when analyzing data streams. Detecting changes both quickly and accurately enables monitoring and prediction systems to react, e.g., by issuing an alarm or by updating a learning algorithm. However, detecting changes is challenging when observations are high-dimensional. In high-dimensional data, change detectors should not only be able to identify when changes happen, but also in which subspace they occur. Ideally, one should also quantify how severe they are. Our approach, ABCD, has these properties. ABCD learns an encoder-decoder model and monitors its accuracy over a window of adaptive size. ABCD derives a change score based on Bernstein’s inequality to detect deviations in terms of accuracy, which indicate changes. Our experiments demonstrate that ABCD outperforms its best competitor by up to 20% in F1-score on average. It can also accurately estimate changes’ subspace, together with a severity measure that correlates with the ground truth.
- 8 October 2023
- Published by Springer Nature
- p. 688-702
Multi-kernel Times Series Outlier Detection
- Open Access
Maximum Mean Discrepancy on Exponential Windows for Online Change Detection
Detecting changes is of fundamental importance when analyzing data streams and has many applications, e.g., predictive maintenance, fraud detection, or medicine. A principled approach to detect changes is to compare the distributions of observations within the stream to each other via hypothesis testing. Maximum mean discrepancy (MMD; also called energy distance) is a well-known (semi-)metric on the space of probability distributions. MMD gives rise to powerful non-parametric two-sample tests on kernel-enriched domains under mild conditions, which makes its deployment for change detection desirable. However, the classic MMD estimators suffer quadratic complexity, which prohibits their application in the online change detection setting. We propose a general-purpose change detection algorithm, Maximum Mean Discrepancy on Exponential Windows (MMDEW), which leverages the MMD two-sample test, facilitates its efficient online computation on any kernel-enriched domain, and is able to detect any disparity between distributions. Our experiments and analysis show that (1) MMDEW achieves better detection quality than state-of-the-art competitors and that (2) the algorithm has polylogarithmic runtime and logarithmic memory requirements, which allow its deployment to the streaming setting.
Open AccessMaximum Mean Discrepancy on Exponential Windows for Online Change Detection
Detecting changes is of fundamental importance when analyzing data streams and has many applications, e.g., predictive maintenance, fraud detection, or medicine. A principled approach to detect changes is to compare the distributions of observations within the stream to each other via hypothesis testing. Maximum mean discrepancy (MMD; also called energy distance) is a well-known (semi-)metric on the space of probability distributions. MMD gives rise to powerful non-parametric two-sample tests on kernel-enriched domains under mild conditions, which makes its deployment for change detection desirable. However, the classic MMD estimators suffer quadratic complexity, which prohibits their application in the online change detection setting. We propose a general-purpose change detection algorithm, Maximum Mean Discrepancy on Exponential Windows (MMDEW), which leverages the MMD two-sample test, facilitates its efficient online computation on any kernel-enriched domain, and is able to detect any disparity between distributions. Our experiments and analysis show that (1) MMDEW achieves better detection quality than state-of-the-art competitors and that (2) the algorithm has polylogarithmic runtime and logarithmic memory requirements, which allow its deployment to the streaming setting.
- Open Access
- 22 June 2021
- Published by Association for Computing Machinery (ACM)
- p. 239-243
An Evaluation of NILM Approaches on Industrial Energy-Consumption Data
Load disaggregation methods infer the energy consumption of individual appliances from their aggregated consumption. This facilitates energy savings and efficient energy management. However, most existing work on load disaggregation has only considered household settings. This may be due to companies preferring to not share their data, rendering such data hardly available. This article makes three contributions: First, we compare data describing the energy consumption of two facilities and of households. Second, we study the performance of seven prominent load disaggregation algorithms on industrial data and compare it to the one on household data. Our results indicate a performance gap on individual appliances. Third, we publish a tool converting an industrial data set to a standard format for load disaggregation, to facilitate further research and benchmarking in the field.
- 22 June 2021
- Published by Association for Computing Machinery (ACM)
- p. 239-243
Open AccessAn Evaluation of NILM Approaches on Industrial Energy-Consumption Data
Load disaggregation methods infer the energy consumption of individual appliances from their aggregated consumption. This facilitates energy savings and efficient energy management. However, most existing work on load disaggregation has only considered household settings. This may be due to companies preferring to not share their data, rendering such data hardly available. This article makes three contributions: First, we compare data describing the energy consumption of two facilities and of households. Second, we study the performance of seven prominent load disaggregation algorithms on industrial data and compare it to the one on household data. Our results indicate a performance gap on individual appliances. Third, we publish a tool converting an industrial data set to a standard format for load disaggregation, to facilitate further research and benchmarking in the field.
- 1 March 2021
- Published by Elsevier in Information Systems
- Vol. 97
Efficient subspace search in data streams
- 1 March 2021
- Published by Elsevier in Information Systems
- Vol. 97
Efficient subspace search in data streams