Fraunhofer-Institut für Angewandte Informationstechnik FIT

Filters

Reset filters

Settings
Now showing 1 - 10 of 4697
No Thumbnail Available
Publication

Towards Access Control for Machine Learning Embeddings

2024-06 , Matzutt, Roman

In this work, we explore the potential to make embeddings, which are becoming an integral part of machine-learning pipelines, shareable with the general public while providing self-contained access control. To this end, we apply attribute-based encryption and discuss a potential application for supply chain management.

View
No Thumbnail Available
Publication

Managing artificial intelligence applications in healthcare: Promoting information processing among stakeholders

2024-04 , Hofmann, Peter , Lämmermann, Luis , Urbach, Nils

AI applications hold great potential for improving healthcare. However, successfully operating AI is a complex endeavor requiring organizations to establish adequate management approaches. Managing AI applications requires functioning information exchange between a diverse set of stakeholders. Lacking information processing among stakeholders increases task uncertainty, hampering the operation of AI applications. Existing research lacks an understanding of holistic AI management approaches. To shed light on AI management in healthcare, we conducted a multi-perspective literature analysis followed by an interview study. Based on the organizational information processing theory, this paper investigates AI management in healthcare from an organizational perspective. As a result, we develop the AI application management model (AIAMA) that illustrates the managerial factors of AI management in healthcare and its interrelations. Furthermore, we provide managerial practices that improve information processing among stakeholders. We contribute to the academic discourse by providing a conceptual framework that increases the theoretical understanding of AI's management factors and understanding of management interrelations. Moreover, we contribute to practice by providing management practices that promote information processing and decrease task uncertainty when managing AI applications in healthcare.

View
No Thumbnail Available
Publication

Vergleich zweier Ansätze zur Bekämpfung der kalten Progression: Tarifverschiebung vs. einkommensteuerpflichtige Kopfpauschale

2024 , Broer, Michael , Stöwhase, Sven

Die Bundesregierung hat im Jahr 2022 das Inflationsausgleichsgesetz verabschiedet, das zum Ausgleich der kalten Progression in der Einkommensteuer sowohl den Grundfreibetrag erhöht als auch die übrigen Tarifeckpunkte anpasst. Bei dem in Gesetzentwürfen verpflichtenden Hinweis auf Alternativen steht "keine". Doch es gibt sehr wohl eine Alternative zur Tarifverschiebung, schreiben Michael Broer und Sven Stöwhase. Sie besteht darin, allein den Grundfreibetrag zu erhöhen, was gemäß höchstrichterlicher Rechtsprechung zwingend notwendig ist. Die fiskalischen Einnahmen, die durch den Verzicht auf eine Anpassung der übrigen Tarifeckpunkte zustande kommen, werden dann über eine steuerpflichtige Kopfpauschale an die Steuerpflichtigen ausgeschüttet, ähnlich wie im Fall der Energiepauschale im Jahr 2022. Es zeigt sich beim Vergleich, dass gerade einkommensschwache Steuerzahler, die von der Inflation aktuell besonders belastet werden, von einer Kopfpauschalenlösung profitieren würden.

View
No Thumbnail Available
Publication

Der Einfluss von digitalen Technologien auf Wissensarbeit: Kompetenzen im Wandel

2024 , Lanzl, Julia , Schnaak, Feline , Schöttl, Frederik , Gimpel, Henner

Die voranschreitende Digitalisierung und der technologische Fortschritt haben in der Arbeitswelt Auswirkungen auf die Ausbildung von Fachkräften und insbesondere auf die geforderten Kompetenzen für einen langfristigen Unternehmenserfolg. Neben klassischen Basiskompetenzen, wie beispielsweise kritischem Denken, treten Digitalkompetenzen wie beispielswiese die Verarbeitung von Informationen zunehmend in den Vordergrund. Deshalb bestehen die Fragen: Was sind durch die fortschreitende Digitalisierung relevante Digitalkompetenzen in der Wissensarbeit? Wie präsentieren und vermitteln Unternehmen und Hochschulen diese Digitalkompetenzen und welche Handlungsempfehlungen lassen sich diesbezüglich ableiten? Durch eine Literaturrecherche in praxisnaher und wissenschaftlicher Literatur wurden Digitalkompetenzen gesammelt, definiert und klassifiziert. Zwölf wichtige Digitalkompetenzen werden herausgestellt und es wird beobachtet, ob und wie sie in der Praxis bereits in betriebswirtschaftlichen Studiengängen, unternehmensinternen sowie -externen Off-the-job-Weiterbildungen und im Recruiting Beachtung finden. Dies ermöglicht das Ableiten von Handlungsempfehlungen für Hochschulen und Unternehmen, sowie deren Zusammenarbeit. Zusätzlich verhelfen die Ergebnisse Weiterbildungen entsprechend an die neuen Gegebenheiten anzupassen und die Wichtigkeit der Digitalkompetenzen für den Berufseinstieg und die berufliche Laufbahn zu verdeutlichen.

View
No Thumbnail Available
Publication

Towards Improving Accountability in Sensitive-Disclosure Scenarios

2024-06 , Matzutt, Roman , Wagner, Eric

Public transparency has become increasingly important to uphold trust in government agencies and private companies alike, e.g., by establishing police accountability and proving abiding to ethical supply chain practices. Oftentimes, however, this public interest conflicts with the need for confidentiality of ongoing processes. In this paper, we investigate these sensitive-disclosure scenarios and the requirements for technical solutions to support the data dissemination in these scenarios. We identify translucent blockchains as a promising building block to provide transparency in sensitive-disclosure scenarios with fine-granular access control.

View
No Thumbnail Available
Publication

Identification of Medical Ecosystems in the Field of Mental Health and Cardiovascular Diseases at the Cologne Site

2024-03-15 , Dannenberg, Cara , Heimann, Johannes , Koumpis, Adamantios , Beyan, Oya Deniz

As part of the Europe-wide smart health innovation hub implemented in the context of the Horizon Europe SHIFT-HUB project, our work concerns the identification of specific medical research ecosystems in the two fields, namely cardiovascular diseases and mental illness, with Cologne as the central location. To achieve this aim, the websites of involved organizations were used for data research purposes, and the members of each respective ecosystem or network were identified by acquiring information about their cooperation partners. A variety of selection criteria have been applied to filter out whether these partners were suitable to be considered as a further starting point for the research. The results indicate the existence of ecosystems in the two fields, with Cologne as the central location, in which various stakeholders, including healthcare institutions, healthcare providers, foundations, NGOs, and the business community, work closely together. Larger institutions are usually networked at an international level, while smaller institutions increasingly depend on and foster regional partnerships. This promotes cooperation and the exchange of knowledge at the regional level and facilitates direct contact with the people affected, i.e., patients' groups. Research institutions in both fields often receive financial support from commercial organizations, which highlights the importance of the business community's involvement in exploiting research results and promoting the quality of healthcare. The article highlights the complexity and interdisciplinarity of the particular ecosystems, with all the different categories of institutions comprising an indispensable position. The interaction amongst stakeholders at international, regional, and local levels can significantly help to deploy resources more effectively and improve the quality of life of people suffering from any of the two conditions.

View
No Thumbnail Available
Publication

Skeleton Ground Truth Extraction: Methodology, Annotation Tool and Benchmarks

2024 , Yang, Cong , Indurkhya, Bipin , See, John , Gao, Bo , Ke, Yan , Boukhers, Zeyd , Yang, Zhenyu , Grzegorzek, Marcin

Skeleton Ground Truth (GT) is critical to the success of supervised skeleton extraction methods, especially with the popularity of deep learning techniques. Furthermore, we see skeleton GTs used not only for training skeleton detectors with Convolutional Neural Networks (CNN), but also for evaluating skeleton-related pruning and matching algorithms. However, most existing shape and image datasets suffer from the lack of skeleton GT and inconsistency of GT standards. As a result, it is difficult to evaluate and reproduce CNN-based skeleton detectors and algorithms on a fair basis. In this paper, we present a heuristic strategy for object skeleton GT extraction in binary shapes and natural images. Our strategy is built on an extended theory of diagnosticity hypothesis, which enables encoding human-in-the-loop GT extraction based on clues from the target's context, simplicity, and completeness. Using this strategy, we developed a tool, SkeView, to generate skeleton GT of 17 existing shape and image datasets. The GTs are then structurally evaluated with representative methods to build viable baselines for fair comparisons. Experiments demonstrate that GTs generated by our strategy yield promising quality with respect to standard consistency, and also provide a balance between simplicity and completeness.

View
No Thumbnail Available
Publication

Towards an ELSA Curriculum for Data Scientists

2024-04-11 , Christoforaki, Maria , Beyan, Oya Deniz

The use of artificial intelligence (AI) applications in a growing number of domains in recent years has put into focus the ethical, legal, and societal aspects (ELSA) of these technologies and the relevant challenges they pose. In this paper, we propose an ELSA curriculum for data scientists aiming to raise awareness about ELSA challenges in their work, provide them with a common language with the relevant domain experts in order to cooperate to find appropriate solutions, and finally, incorporate ELSA in the data science workflow. ELSA should not be seen as an impediment or a superfluous artefact but rather as an integral part of the Data Science Project Lifecycle. The proposed curriculum uses the CRISP-DM (CRoss-Industry Standard Process for Data Mining) model as a backbone to define a vertical partition expressed in modules corresponding to the CRISP-DM phases. The horizontal partition includes knowledge units belonging to three strands that run through the phases, namely ethical and societal, legal and technical rendering knowledge units (KUs). In addition to the detailed description of the aforementioned KUs, we also discuss their implementation, issues such as duration, form, and evaluation of participants, as well as the variance of the knowledge level and needs of the target audience.

View
No Thumbnail Available
Publication

Consumer-centric electricity markets: Six design principles

2024-03 , Michaelis, Anne , Hanny, Lisa-Maria , Körner, Marc-Fabian , Strüker, Jens , Weibelzahl, Martin

Key to the necessary decarbonization of energy systems is the large-scale expansion of renewable energy sources and their integration into the electricity system. This integration is challenging because the feed-in from renewable energy sources is highly intermittent and largely dependent on uncontrollable factors such as weather patterns. To maintain grid stability, which refers to the required balance between demand and supply in the electricity system, flexibility is key. Large flexibility potentials can be found on the electricity demand side. However, current electricity market design in Europe, while providing major flexibility incentives, often neglects small-scale electricity consumers and distributed energy resources. We contribute to shape future electricity markets with consumers at the heart by developing six design principles for a consumer-centric electricity market design. We proceed by conducting a systematic literature review and evaluate the findings by expert interviews. Based on the developed design principles, we define a consumer-centric electricity market design as a set of market rules that align with the rules of other relevant energy markets and allow for the efficient matching of electricity demand and supply, with consumers having nondiscriminatory market access, being exposed to fine-grained price signals, being able to express their preferences, and having sufficient possibilities to protect themselves against unexpected price spikes. By actively incorporating consumers into electricity markets, we contribute to the overarching goal of integrating renewable energy sources while promoting energy justice, i.e., supporting a balanced mix of economic, political, environmental, and social interests.

View
No Thumbnail Available
Publication

Negative electricity prices as a signal for lacking flexibility? On the effects of demand flexibility on electricity prices

2024 , Halbrügge, Stephanie , Heeß, Paula , Schott, Paul , Weibelzahl, Martin

Purpose: The purpose of this paper is to examine how active consumers, i.e. consumers that can inter-temporally shift their load, can influence electricity prices. As demonstrated in this paper, inter-temporal load shifting can induce negative electricity prices, a recurring phenomenon on power exchanges. Design/methodology/approach: The paper presents a novel electricity-market model assuming a nodal-pricing, energy-only spot market with active consumers. This study formulates an economic equilibrium problem as a linear program and uses an established six-node case study to compare equilibrium prices of a model with inflexible demand to a model with flexible demand of active consumers. Findings: This study illustrates that temporal coupling of hourly market clearing through load shifting of active consumers can cause negative electricity prices that are not observed in a model with ceteris paribus inflexible demand. In such situations, where compared to the case of inflexible demand more flexibility is available in the system, negative electricity prices signal lower total system costs. These negative prices result from the use of demand flexibility, which, however, cannot be fully exploited due to limited transmission capacities, respectively, loop-flow restrictions. Originality/value: Literature indicates that negative electricity prices result from lacking flexibility. The results illustrate that active consumers and their additional flexibility can lead to negative electricity prices in temporally coupled markets, which in general contributes to increased system efficiency as well as increased use of renewable energy sources. These findings extend existing research in both the area of energy flexibility and causes for negative electricity prices. Therefore, policymakers should be aware of such (temporal coupling) effects and, e.g. continue to allow negative electricity prices in the future that can serve as investment signals for active consumers.

View