Jens Ulrik Hansen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Jens Ulrik Hansen has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
Zoé Christoff
3 shared publications- Thomas Bolander
Technical University of Denmark
2 shared publications Carlo Proietti
University of Liverpool
1 shared publication- Christopher H. Gyldenkærne
Roskilde University
1 shared publication - Ekaterina Pashevich
University of Copenhagen
1 shared publication - Erik J. Olsson
Department of Philosophy, Lund University, Sweden
1 shared publication
Latest Publications
- Open Access
- 30 May 2026
- Published by Springer Nature in Minds and Machines
- Vol. 36 (2)
On the Conceptualization of Synthetic Data: Why Not All Synthetic Data are Useful and Some are
Synthetic data are increasingly used across the sciences and industry and are widely regarded as a solution to data problems such as privacy, bias mitigation, and data insufficiency. At the same time, the “promises” of synthetic data have faced substantial scholarly critique. We argue that much of the disagreement surrounding the usefulness of synthetic data stems from a lack of clarity about what synthetic data actually are and what they can do. In this paper, we examine several conceptualizations of synthetic data and propose a refined concept of synthesized data that more accurately captures the phenomenon commonly referred to as synthetic data. We argue that synthetic data cannot be conceptualized in contrast with other types of data. Moreover, there is no meaningful difference between synthetic data and data. We further prove this argument through an Iceberg model of the process of data synthetization and utilization, which addresses the “submerged” data risks such as bias, privacy, and validity, thus ensuring the usefulness of synthesized data. Our approach is grounded in a relational view of data, which links their intended purpose with their method of generation. We further elaborate this connection through concrete examples. In conclusion, we demonstrate the complexity of synthesized data in relation to their utility, which often remains hidden in current discourse on synthetic data.
- 30 May 2026
- Published by Springer Nature in Minds and Machines
- Vol. 36 (2)
Open AccessOn the Conceptualization of Synthetic Data: Why Not All Synthetic Data are Useful and Some are
Synthetic data are increasingly used across the sciences and industry and are widely regarded as a solution to data problems such as privacy, bias mitigation, and data insufficiency. At the same time, the “promises” of synthetic data have faced substantial scholarly critique. We argue that much of the disagreement surrounding the usefulness of synthetic data stems from a lack of clarity about what synthetic data actually are and what they can do. In this paper, we examine several conceptualizations of synthetic data and propose a refined concept of synthesized data that more accurately captures the phenomenon commonly referred to as synthetic data. We argue that synthetic data cannot be conceptualized in contrast with other types of data. Moreover, there is no meaningful difference between synthetic data and data. We further prove this argument through an Iceberg model of the process of data synthetization and utilization, which addresses the “submerged” data risks such as bias, privacy, and validity, thus ensuring the usefulness of synthesized data. Our approach is grounded in a relational view of data, which links their intended purpose with their method of generation. We further elaborate this connection through concrete examples. In conclusion, we demonstrate the complexity of synthesized data in relation to their utility, which often remains hidden in current discourse on synthetic data.
- Open Access
- 25 October 2023
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 18 (10) , e0291689
Student and teacher performance during COVID-19 lockdown: An investigation of associated features and complex interactions using multiple data sources
Due to the COVID-19 pandemic, testing what is required to support teachers and students while subject to forced online teaching and learning is relevant in terms of similar situations in the future. To understand the complex relationships of numerous factors with teaching during the lockdown, we used administrative data and survey data from a large Danish university. The analysis employed scores from student evaluations of teaching and the students’ final grades during the first wave of the COVID-19 lockdown in the spring of 2020 as dependent targets in a linear regression model and a random forest model. This led to the identification of linear and non-linear relationships, as well as feature importance and interactions for the two targets. In particular, we found that many factors, such as the age of teachers and their time use, were associated with the scores in student evaluations of teaching and student grades, and that other features, including peer interaction among teachers and student gender, also exerted influence, especially on grades. Finally, we found that for non-linear features, in terms of the age of teachers and students, the average values led to the highest response values for scores in student evaluations of teaching and grades.
- 25 October 2023
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 18 (10) , e0291689
Open AccessStudent and teacher performance during COVID-19 lockdown: An investigation of associated features and complex interactions using multiple data sources
Due to the COVID-19 pandemic, testing what is required to support teachers and students while subject to forced online teaching and learning is relevant in terms of similar situations in the future. To understand the complex relationships of numerous factors with teaching during the lockdown, we used administrative data and survey data from a large Danish university. The analysis employed scores from student evaluations of teaching and the students’ final grades during the first wave of the COVID-19 lockdown in the spring of 2020 as dependent targets in a linear regression model and a random forest model. This led to the identification of linear and non-linear relationships, as well as feature importance and interactions for the two targets. In particular, we found that many factors, such as the age of teachers and their time use, were associated with the scores in student evaluations of teaching and student grades, and that other features, including peer interaction among teachers and student gender, also exerted influence, especially on grades. Finally, we found that for non-linear features, in terms of the age of teachers and students, the average values led to the highest response values for scores in student evaluations of teaching and grades.
- Open Access
- 29 September 2023
- Published by Elsevier in International Journal of Human-Computer Studies
- Vol. 181
Innovation tactics for implementing an ML application in healthcare: A long and winding road
- 29 September 2023
- Published by Elsevier in International Journal of Human-Computer Studies
- Vol. 181
Open AccessInnovation tactics for implementing an ML application in healthcare: A long and winding road
- Open Access
- 20 January 2023
- Published by Springer Nature in Synthese
- Vol. 201 (2) , 1-21
The importance of expert knowledge in big data and machine learning
According to popular belief, big data and machine learning provide a wholly novel approach to science that has the potential to revolutionise scientific progress and will ultimately lead to the ‘end of theory’. Proponents of this view argue that advanced algorithms are able to mine vast amounts of data relating to a given problem without any prior knowledge and that we do not need to concern ourselves with causality, as correlation is sufficient for handling complex issues. Consequently, the human contribution to scientific progress is deemed to be non-essential and replaceable. We, however, following the position most commonly represented in the philosophy of science, argue that the need for human expertise remains. Based on an analysis of big data and machine learning methods in two case studies—skin cancer detection and protein folding—we show that expert knowledge is essential and inherent in the application of these methods. Drawing on this analysis, we establish a classification of the different kinds of expert knowledge that are involved in the application of big data and machine learning in scientific contexts. We address the ramifications of a human-driven expert knowledge approach to big data and machine learning for scientific practice and the discussion about the role of theory. Finally, we show that the ways in which big data and machine learning both influence and are influenced by scientific methodology involve continuous conceptual shifts rather than a rigid paradigm change.
- 20 January 2023
- Published by Springer Nature in Synthese
- Vol. 201 (2) , 1-21
Open AccessThe importance of expert knowledge in big data and machine learning
According to popular belief, big data and machine learning provide a wholly novel approach to science that has the potential to revolutionise scientific progress and will ultimately lead to the ‘end of theory’. Proponents of this view argue that advanced algorithms are able to mine vast amounts of data relating to a given problem without any prior knowledge and that we do not need to concern ourselves with causality, as correlation is sufficient for handling complex issues. Consequently, the human contribution to scientific progress is deemed to be non-essential and replaceable. We, however, following the position most commonly represented in the philosophy of science, argue that the need for human expertise remains. Based on an analysis of big data and machine learning methods in two case studies—skin cancer detection and protein folding—we show that expert knowledge is essential and inherent in the application of these methods. Drawing on this analysis, we establish a classification of the different kinds of expert knowledge that are involved in the application of big data and machine learning in scientific contexts. We address the ramifications of a human-driven expert knowledge approach to big data and machine learning for scientific practice and the discussion about the role of theory. Finally, we show that the ways in which big data and machine learning both influence and are influenced by scientific methodology involve continuous conceptual shifts rather than a rigid paradigm change.
- Open Access
- 17 September 2018
- Published by Springer Nature in Synthese
- Vol. 197 (4) , 1511-1541
Truth tracking performance of social networks: how connectivity and clustering can make groups less competent
- 17 September 2018
- Published by Springer Nature in Synthese
- Vol. 197 (4) , 1511-1541
Open AccessTruth tracking performance of social networks: how connectivity and clustering can make groups less competent
- 7 October 2016
- Published by Springer Nature in Journal of Logic, Language and Information
- Vol. 25 (3-4) , 299-333
Reflecting on Social Influence in Networks
- 7 October 2016
- Published by Springer Nature in Journal of Logic, Language and Information
- Vol. 25 (3-4) , 299-333
Reflecting on Social Influence in Networks
- 20 December 2015
- Published by Oxford University Press (OUP) in Journal of Logic and Computation
- Vol. 29 (7) , 1121-1137
Reasoning about opinion dynamics in social networks
This article introduces a logic to reason about a well-known model of opinion dynamics in social networks initially developed by Morris DeGroot as well as Keith Lehrer and Carl Wagner. The proposed logic is an extension of Łukasiewicz' fuzzy logic with additional equational expressivity, modal operators, machinery from hybrid logic and dynamic modalities. The model of opinion dynamics in social networks is simple enough to be easily grasped, but still complex enough to have interesting mathematical properties and applications. Thus, developing a logic to reason about this particular model serves as a paradigmatic example of how logic can be useful in social network analysis.
- 20 December 2015
- Published by Oxford University Press (OUP) in Journal of Logic and Computation
- Vol. 29 (7) , 1121-1137
Reasoning about opinion dynamics in social networks
This article introduces a logic to reason about a well-known model of opinion dynamics in social networks initially developed by Morris DeGroot as well as Keith Lehrer and Carl Wagner. The proposed logic is an extension of Łukasiewicz' fuzzy logic with additional equational expressivity, modal operators, machinery from hybrid logic and dynamic modalities. The model of opinion dynamics in social networks is simple enough to be easily grasped, but still complex enough to have interesting mathematical properties and applications. Thus, developing a logic to reason about this particular model serves as a paradigmatic example of how logic can be useful in social network analysis.
- 15 June 2015
- Published by Oxford University Press (OUP) in Journal of Logic and Computation
- Vol. 28 (5) , 883-908
Many-valued hybrid logic
In this article we define a family of many-valued semantics for hybrid logic, where each semantics is based on a finite Heyting algebra of truth-values. We provide sound and complete tableau systems for these semantics. Moreover, we show how the tableau systems can be made terminating and thereby give rise to decision procedures for the logics in question. Our many-valued hybrid logics turn out to be ‘intermediate’ logics between intuitionistic hybrid logic and classical hybrid logic in a specific sense explained in the article. Our results show that many-valued hybrid logic is indeed a natural enterprise.
- 15 June 2015
- Published by Oxford University Press (OUP) in Journal of Logic and Computation
- Vol. 28 (5) , 883-908
Many-valued hybrid logic
In this article we define a family of many-valued semantics for hybrid logic, where each semantics is based on a finite Heyting algebra of truth-values. We provide sound and complete tableau systems for these semantics. Moreover, we show how the tableau systems can be made terminating and thereby give rise to decision procedures for the logics in question. Our many-valued hybrid logics turn out to be ‘intermediate’ logics between intuitionistic hybrid logic and classical hybrid logic in a specific sense explained in the article. Our results show that many-valued hybrid logic is indeed a natural enterprise.
- 1 March 2015
- Published by Elsevier in Journal of Applied Logic
- Vol. 13 (1) , 48-77
A logic for diffusion in social networks
- 1 March 2015
- Published by Elsevier in Journal of Applied Logic
- Vol. 13 (1) , 48-77
A logic for diffusion in social networks
- 25 March 2014
- Published by Springer Nature in Synthese
- Vol. 191 (11) , 2445-2470
On the rationality of pluralistic ignorance
- 25 March 2014
- Published by Springer Nature in Synthese
- Vol. 191 (11) , 2445-2470
On the rationality of pluralistic ignorance