Carsten Knoll
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Carsten Knoll has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Klaus Röbenack
TUD Dresden University of Technology
7 shared publications Klaus Röbenack
Institut für Regelungs‐ und Steuerungstheorie Fakultät Elektrotechnik und Informationstechnik Technische Universität Dresden 01062 Dresden
6 shared publications- Julius Fiedler
Institute of Control Theory TUD Dresden University of Technology Dresden Germany
4 shared publications - Steffen Seitz
TUD Dresden University of Technology
4 shared publications Chenzi Huang
Technische Universität Dresden, Fakultät Elektrotechnik und Informationstechnik, Institut für Regelungs- und Steuerungstheorie, 01062 Dresden Germany
2 shared publications- Haadia Amjad
Faculty of Electrical and Computer Engineering, Chair of Fundamentals of Electrical Engineering, TUD | Dresden University of Technology, Dresden, Germany
2 shared publications
Latest Publications
- Open Access
- 17 November 2026
- Published by Springer Nature
- p. 39-52
How a Model’s Architecture and Performance Influences Explainability: A Study of Grad-CAM and CRAFT in CNN-Based Image Classification
- 17 November 2026
- Published by Springer Nature
- p. 39-52
Open AccessHow a Model’s Architecture and Performance Influences Explainability: A Study of Grad-CAM and CRAFT in CNN-Based Image Classification
- Open Access
LLM-Supported Formal Knowledge Representation for Enhancing Control Engineering Content with an Interactive Semantic Layer
The rapid growth of research output in control engineering calls for new approaches to structure and formalize domain knowledge. This paper briefly describes an LLM-supported method for semi-automated generation of formal knowledge representations that combine human readability with machine interpretability and increased expressiveness. Based on the Imperative Representation of Knowledge (PyIRK) framework, we demonstrate how language models can assist in transforming natural-language descriptions and mathematical definitions (available as LaTeX source code) into a formalized knowledge graph. As a first application we present the generation of an ``interactive semantic layer'' to enhance the source documents in order to facilitate knowledge transfer. From our perspective this contributes to the vision of easily accessible, collaborative, and verifiable knowledge bases for the control engineering domain.
Open AccessLLM-Supported Formal Knowledge Representation for Enhancing Control Engineering Content with an Interactive Semantic Layer
The rapid growth of research output in control engineering calls for new approaches to structure and formalize domain knowledge. This paper briefly describes an LLM-supported method for semi-automated generation of formal knowledge representations that combine human readability with machine interpretability and increased expressiveness. Based on the Imperative Representation of Knowledge (PyIRK) framework, we demonstrate how language models can assist in transforming natural-language descriptions and mathematical definitions (available as LaTeX source code) into a formalized knowledge graph. As a first application we present the generation of an ``interactive semantic layer'' to enhance the source documents in order to facilitate knowledge transfer. From our perspective this contributes to the vision of easily accessible, collaborative, and verifiable knowledge bases for the control engineering domain.
- Open Access
- 12 October 2025
- Published by Springer Nature
- p. 250-263
XAIEV – A Framework for the Evaluation of XAI-Algorithms for Image Classification
Convolutional Neural Networks (CNNs), such as VGG and ResNet, have been widely used for image classification for several years. Numerous explainable AI (XAI) algorithms, including Grad-CAM and XRAI, have been proposed to enhance interpretability in this domain. However, a persistent challenge lies in quantitatively comparing different XAI algorithms, variants of the same algorithm, or combinations of CNN models and XAI algorithms. In this work, we introduce XAIEV – a versatile framework for computationally evaluating the quality of saliency-map-based XAI algorithms. This framework includes (A) a benchmark dataset (traffic sign recognition) with a known ground truth and (B) a software toolbox designed to facilitate the evaluation pipeline. The pipeline consists of four steps: (1) model training, (2) applying XAI algorithms to generate weighted saliency maps, (3) generating new test images with varying percentages of “important” pixels removed or retained, and (4) statistically evaluating accuracy changes on these test images and comparison to the ground truth. Based on this statistical evaluation, we define an Accuracy-Sensitivity Quotient (ASQ) as a novel quality metric for XAI algorithms applied to image classification. Using the XAIEV framework, we compare various combinations of CNN architectures (“SimpleCNN” (custom model), VGG, ResNet, ConvNext) with multiple XAI algorithms (Grad-CAM, XRAI, LIME, PRISM). Our numerical results reveal that the performance of XAI algorithms is highly dependent on the underlying CNN model.
- 12 October 2025
- Published by Springer Nature
- p. 250-263
Open AccessXAIEV – A Framework for the Evaluation of XAI-Algorithms for Image Classification
Convolutional Neural Networks (CNNs), such as VGG and ResNet, have been widely used for image classification for several years. Numerous explainable AI (XAI) algorithms, including Grad-CAM and XRAI, have been proposed to enhance interpretability in this domain. However, a persistent challenge lies in quantitatively comparing different XAI algorithms, variants of the same algorithm, or combinations of CNN models and XAI algorithms. In this work, we introduce XAIEV – a versatile framework for computationally evaluating the quality of saliency-map-based XAI algorithms. This framework includes (A) a benchmark dataset (traffic sign recognition) with a known ground truth and (B) a software toolbox designed to facilitate the evaluation pipeline. The pipeline consists of four steps: (1) model training, (2) applying XAI algorithms to generate weighted saliency maps, (3) generating new test images with varying percentages of “important” pixels removed or retained, and (4) statistically evaluating accuracy changes on these test images and comparison to the ground truth. Based on this statistical evaluation, we define an Accuracy-Sensitivity Quotient (ASQ) as a novel quality metric for XAI algorithms applied to image classification. Using the XAIEV framework, we compare various combinations of CNN architectures (“SimpleCNN” (custom model), VGG, ResNet, ConvNext) with multiple XAI algorithms (Grad-CAM, XRAI, LIME, PRISM). Our numerical results reveal that the performance of XAI algorithms is highly dependent on the underlying CNN model.
- Open Access
- 21 March 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (6) , 2648
Do Humans and Convolutional Neural Networks Attend to Similar Areas during Scene Classification: Effects of Task and Image Type
Deep neural networks are powerful image classifiers but do they attend to similar image areas as humans? While previous studies have investigated how this similarity is shaped by technological factors, little is known about the role of factors that affect human attention. Therefore, we investigated the interactive effects of task and image characteristics. We varied the intentionality of the tasks used to elicit human attention maps (i.e., spontaneous gaze, gaze-pointing, manual area selection). Moreover, we varied the type of image to be categorized (i.e., singular objects, indoor scenes consisting of object arrangements, landscapes without distinct objects). The human attention maps generated in this way were compared to the attention maps of a convolutional neural network (CNN) as revealed by a method of explainable artificial intelligence (Grad-CAM). The influence of human tasks strongly depended on image type: for objects, human manual selection produced attention maps that were most similar to CNN, while the specific eye movement task had little impact. For indoor scenes, spontaneous gaze produced the least similarity, while for landscapes, similarity was equally low across all human tasks. Our results highlight the importance of taking human factors into account when comparing the attention of humans and CNN.
- 21 March 2024
- Published by MDPI AG in Applied Sciences
- Vol. 14 (6) , 2648
Open AccessDo Humans and Convolutional Neural Networks Attend to Similar Areas during Scene Classification: Effects of Task and Image Type
Deep neural networks are powerful image classifiers but do they attend to similar image areas as humans? While previous studies have investigated how this similarity is shaped by technological factors, little is known about the role of factors that affect human attention. Therefore, we investigated the interactive effects of task and image characteristics. We varied the intentionality of the tasks used to elicit human attention maps (i.e., spontaneous gaze, gaze-pointing, manual area selection). Moreover, we varied the type of image to be categorized (i.e., singular objects, indoor scenes consisting of object arrangements, landscapes without distinct objects). The human attention maps generated in this way were compared to the attention maps of a convolutional neural network (CNN) as revealed by a method of explainable artificial intelligence (Grad-CAM). The influence of human tasks strongly depended on image type: for objects, human manual selection produced attention maps that were most similar to CNN, while the specific eye movement task had little impact. For indoor scenes, spontaneous gaze produced the least similarity, while for landscapes, similarity was equally low across all human tasks. Our results highlight the importance of taking human factors into account when comparing the attention of humans and CNN.
- Open Access
Imperative Formal Knowledge Representation for Control Engineering: Examples from Lyapunov Theory
In this paper, we introduce a novel method to formally represent elements of control engineering knowledge in a suitable data structure. To this end, we first briefly review existing representation methods (RDF, OWL, Wikidata, ORKG). Based on this, we introduce our own approach: The Python-based imperative representation of knowledge (PyIRK) and its application to formulate the Ontology of Control Systems Engineering (OCSE). One of its main features is the possibility to represent the actual content of definitions and theorems as nodes and edges of a knowledge graph, which is demonstrated by selected theorems from Lyapunov’s theory. While the approach is still experimental, the current result already allows the application of methods of automated quality assurance and a SPARQL-based semantic search mechanism. The feature set of the framework is demonstrated by various examples. The paper concludes with a discussion of the limitations and directions for further development.
Open AccessImperative Formal Knowledge Representation for Control Engineering: Examples from Lyapunov Theory
In this paper, we introduce a novel method to formally represent elements of control engineering knowledge in a suitable data structure. To this end, we first briefly review existing representation methods (RDF, OWL, Wikidata, ORKG). Based on this, we introduce our own approach: The Python-based imperative representation of knowledge (PyIRK) and its application to formulate the Ontology of Control Systems Engineering (OCSE). One of its main features is the possibility to represent the actual content of definitions and theorems as nodes and edges of a knowledge graph, which is demonstrated by selected theorems from Lyapunov’s theory. While the approach is still experimental, the current result already allows the application of methods of automated quality assurance and a SPARQL-based semantic search mechanism. The feature set of the framework is demonstrated by various examples. The paper concludes with a discussion of the limitations and directions for further development.
- Open Access
- 6 March 2024
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 41 (4) , 2371-2393
Interpretability is in the Eye of the Beholder: Human Versus Artificial Classification of Image Segments Generated by Humans Versus XAI
- 6 March 2024
- Published by Taylor & Francis in International Journal of Human–Computer Interaction
- Vol. 41 (4) , 2371-2393
Open AccessInterpretability is in the Eye of the Beholder: Human Versus Artificial Classification of Image Segments Generated by Humans Versus XAI
- Open Access
- 3 January 2024
- Published by MDPI AG in Preprints.org
Formal Knowledge Representation Applied to Lyapunov Theory
This paper is motivated by three main assumptions: (1) The currently dominant forms of knowledge representation (natural language text, formulas, graphics) are suboptimal to facilitate knowledge transfer between control engineering sub-fields and into potential application domains. (2) Formal knowledge representation (in general) is a promising supplementary approach, but (3) the established technologies so far have failed to be significantly adopted for formal representation of control engineering knowledge. Thus, after briefly reviewing existing representation methods, we introduce our own approach: The Python based imperative representation of knowledge (PyIRK) and its application to formulate the Ontology of Control Systems Engineering (OCSE). One of its main features is the possibility to represent the actual content of definitions and theorems as nodes and edges of a knowledge graph, which is demonstrated by selected elements from Lyapunov theory. While the approach is still experimental, the formalization already allows to apply methods of automated quality assurance and a semantic search mechanism. The current feature set of the framework is demonstrated on various examples before directions for further developments are discussed.
- 3 January 2024
- Published by MDPI AG in Preprints.org
Open AccessFormal Knowledge Representation Applied to Lyapunov Theory
This paper is motivated by three main assumptions: (1) The currently dominant forms of knowledge representation (natural language text, formulas, graphics) are suboptimal to facilitate knowledge transfer between control engineering sub-fields and into potential application domains. (2) Formal knowledge representation (in general) is a promising supplementary approach, but (3) the established technologies so far have failed to be significantly adopted for formal representation of control engineering knowledge. Thus, after briefly reviewing existing representation methods, we introduce our own approach: The Python based imperative representation of knowledge (PyIRK) and its application to formulate the Ontology of Control Systems Engineering (OCSE). One of its main features is the possibility to represent the actual content of definitions and theorems as nodes and edges of a knowledge graph, which is demonstrated by selected elements from Lyapunov theory. While the approach is still experimental, the formalization already allows to apply methods of automated quality assurance and a semantic search mechanism. The current feature set of the framework is demonstrated on various examples before directions for further developments are discussed.
- Open Access
- 10 September 2023
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 23 (2)
Catalog of Dynamical System Models with Semantic Metadata
This contribution presents the concept and implementation of a catalog (i.e., a structured collection) of dynamical system models, relevant to control theory. The key features of the catalog are: (a) a combination of human‐readable model documentation with executable code; (b) semantic metadata for each model plus a web platform to provide SPARQL search interface; and (c) model data and supporting software based on open source repositories and open for extension and improvement.
- 10 September 2023
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 23 (2)
Open AccessCatalog of Dynamical System Models with Semantic Metadata
This contribution presents the concept and implementation of a catalog (i.e., a structured collection) of dynamical system models, relevant to control theory. The key features of the catalog are: (a) a combination of human‐readable model documentation with executable code; (b) semantic metadata for each model plus a web platform to provide SPARQL search interface; and (c) model data and supporting software based on open source repositories and open for extension and improvement.
- Open Access
- 25 January 2021
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 20 (1) , e202000324
Trajectory Planning for Closed Kinematic Chains Applied to Cooperative Motions in Health Care
Physical interaction of humans can be a challenging. For some practically relevant situations a reasonable model is given by a closed kinematic chain of a planar rigid body mechanism. From a control perspective a cooperative motion, e. g. assisted standing‐up, can be seen as an optimal control problem (OCP) with input and state restrictions. Due to the algebraic constraints and the comparably high number of joints involved, proper formulation of such a problem is not trivial. In this contribution we compare two different modeling approaches: discretization of the full dynamical problem and iterative solution of consecutive stationary problems. The latter turns out to be significantly faster while still providing – in some sense – optimal solutions.
- 25 January 2021
- Published by Wiley in Proceedings in Applied Mathematics and Mechanics
- Vol. 20 (1) , e202000324
Open AccessTrajectory Planning for Closed Kinematic Chains Applied to Cooperative Motions in Health Care
Physical interaction of humans can be a challenging. For some practically relevant situations a reasonable model is given by a closed kinematic chain of a planar rigid body mechanism. From a control perspective a cooperative motion, e. g. assisted standing‐up, can be seen as an optimal control problem (OCP) with input and state restrictions. Due to the algebraic constraints and the comparably high number of joints involved, proper formulation of such a problem is not trivial. In this contribution we compare two different modeling approaches: discretization of the full dynamical problem and iterative solution of consecutive stationary problems. The latter turns out to be significantly faster while still providing – in some sense – optimal solutions.
- 31 July 2020
- Published by Walter de Gruyter GmbH in at - Automatisierungstechnik
- Vol. 68 (8) , 612-624
Bestärkendes Lernen mittels Offline-Trajektorienplanung basierend auf iterativ approximierten Modellen
In diesem Beitrag nutzen wir Künstliche Neuronale Netze (KNN) zur Approximation der Dynamik nichtlinearer (mechanischer) Systeme. Diese iterativ approximierten neuronalen Systemmodelle werden in einer Offline-Trajektorienplanung verwendet, um eine optimale Rückführung zu bestimmen, welche auf das reale System angewandt wird. Dieser Ansatz des modellbasierten bestärkenden Lernens (engl. model-based reinforcement learning (RL)) wird am Aufschwingen des Einfachwagenpendels zunächst simulativ evaluiert und zeigt gegenüber modellfreien RL-Ansätzen eine signifikante Verbesserung der Dateneffizienz. Weiterhin zeigen wir Experimentalergebnisse an einem Versuchsstand, wobei der vorgestellte Algorithmus innerhalb weniger Versuche in der Lage ist, eine für das System optimale Rückführung hinreichend gut zu approximieren. In this paper we use artificial neural networks (ANN) to approximate the dynamics of nonlinear (mechanical) systems. These iteratively approximated neural system models are used in an offline trajectory planning to calculate an optimal feedback law that is applied to the real system. This model-based reinforcement learning (RL) approach is evaluated on the swing-up manoeuvre of the cart-pole system and shows a significant performance gain in terms of data efficiency compared to model-free RL approaches. Furthermore, we show experimental results on a test bench. The proposed algorithm is capable of approximating an optimal feedback law for the system after only a few trials.
- 31 July 2020
- Published by Walter de Gruyter GmbH in at - Automatisierungstechnik
- Vol. 68 (8) , 612-624
Bestärkendes Lernen mittels Offline-Trajektorienplanung basierend auf iterativ approximierten Modellen
In diesem Beitrag nutzen wir Künstliche Neuronale Netze (KNN) zur Approximation der Dynamik nichtlinearer (mechanischer) Systeme. Diese iterativ approximierten neuronalen Systemmodelle werden in einer Offline-Trajektorienplanung verwendet, um eine optimale Rückführung zu bestimmen, welche auf das reale System angewandt wird. Dieser Ansatz des modellbasierten bestärkenden Lernens (engl. model-based reinforcement learning (RL)) wird am Aufschwingen des Einfachwagenpendels zunächst simulativ evaluiert und zeigt gegenüber modellfreien RL-Ansätzen eine signifikante Verbesserung der Dateneffizienz. Weiterhin zeigen wir Experimentalergebnisse an einem Versuchsstand, wobei der vorgestellte Algorithmus innerhalb weniger Versuche in der Lage ist, eine für das System optimale Rückführung hinreichend gut zu approximieren. In this paper we use artificial neural networks (ANN) to approximate the dynamics of nonlinear (mechanical) systems. These iteratively approximated neural system models are used in an offline trajectory planning to calculate an optimal feedback law that is applied to the real system. This model-based reinforcement learning (RL) approach is evaluated on the swing-up manoeuvre of the cart-pole system and shows a significant performance gain in terms of data efficiency compared to model-free RL approaches. Furthermore, we show experimental results on a test bench. The proposed algorithm is capable of approximating an optimal feedback law for the system after only a few trials.