Refine
Has Fulltext
- yes (6)
Is part of the Bibliography
- yes (6)
Year of publication
- 2023 (6) (remove)
Document Type
- Journal article (3)
- Doctoral Thesis (3)
Language
- English (6) (remove)
Keywords
Institute
- Betriebswirtschaftliches Institut (6) (remove)
Companies are expected to act as international players and to use their capabilities to provide customized products and services quickly and efficiently. Today, consumers expect their requirements to be met within a short time and at a favorable price. Order-to-delivery lead time has steadily gained in importance for consumers. Furthermore, governments can use various emissions policies to force companies and customers to reduce their greenhouse gas emissions. This thesis investigates the influence of order-to-delivery lead time and different emission policies on the design of a supply chain. Within this work different supply chain design models are developed to examine these different influences. The first model incorporates lead times and total costs, and various emission policies are implemented to illustrate the trade-off between the different measures. The second model reflects the influence of order-to-delivery lead time sensitive consumers, and different emission policies are implemented to study their impacts. The analysis shows that the share of order-to-delivery lead time sensitive consumers has a significant impact on the design of a supply chain. Demand uncertainty and uncertainty in the design of different emission policies are investigated by developing an appropriate robust mathematical optimization model. Results show that especially uncertainties on the design of an emission policy can significantly impact the total cost of a supply chain. The effects of differently designed emission policies in various countries are investigated in the fourth model. The analyses highlight that both lead times and emission policies can strongly influence companies' offshoring and nearshoring strategies.
Recent computing advances are driving the integration of artificial intelligence (AI)-based systems into nearly every facet of our daily lives. To this end, AI is becoming a frontier for enabling algorithmic decision-making by mimicking or even surpassing human intelligence. Thereupon, these AI-based systems can function as decision support systems (DSSs) that assist experts in high-stakes use cases where human lives are at risk. All that glitters is not gold, due to the accompanying complexity of the underlying machine learning (ML) models, which apply mathematical and statistical algorithms to autonomously derive nonlinear decision knowledge. One particular subclass of ML models, called deep learning models, accomplishes unsurpassed performance, with the drawback that these models are no longer explainable to humans. This divergence may result in an end-user’s unwillingness to utilize this type of AI-based DSS, thus diminishing the end-user’s system acceptance.
Hence, the explainable AI (XAI) research stream has gained momentum, as it develops techniques to unravel this black-box while maintaining system performance. Non-surprisingly, these XAI techniques become necessary for justifying, evaluating, improving, or managing the utilization of AI-based DSSs. This yields a plethora of explanation techniques, creating an XAI jungle from which end-users must choose. In turn, these techniques are preliminarily engineered by developers for developers without ensuring an actual end-user fit. Thus, it renders unknown how an end-user’s mental model behaves when encountering such explanation techniques.
For this purpose, this cumulative thesis seeks to address this research deficiency by investigating end-user perceptions when encountering intrinsic ML and post-hoc XAI explanations. Drawing on this, the findings are synthesized into design knowledge to enable the deployment of XAI-based DSSs in practice. To this end, this thesis comprises six research contributions that follow the iterative and alternating interplay between behavioral science and design science research employed in information systems (IS) research and thus contribute to the overall research objectives as follows: First, an in-depth study of the impact of transparency and (initial) trust on end-user acceptance is conducted by extending and validating the unified theory of acceptance and use of technology model. This study indicates both factors’ strong but indirect effects on system acceptance, validating further research incentives. In particular, this thesis focuses on the overarching concept of transparency. Herein, a systematization in the form of a taxonomy and pattern analysis of existing user-centered XAI studies is derived to structure and guide future research endeavors, which enables the empirical investigation of the theoretical trade-off between performance and explainability in intrinsic ML algorithms, yielding a less gradual trade-off, fragmented into three explainability groups. This includes an empirical investigation on end-users’ perceived explainability of post-hoc explanation types, with local explanation types performing best. Furthermore, an empirical investigation emphasizes the correlation between comprehensibility and explainability, indicating almost significant (with outliers) results for the assumed correlation. The final empirical investigation aims at researching XAI explanation types on end-user cognitive load and the effect of cognitive load on end-user task performance and task time, which also positions local explanation types as best and demonstrates the correlations between cognitive load and task performance and, moreover, between cognitive load and task time. Finally, the last research paper utilizes i.a. the obtained knowledge and derives a nascent design theory for XAI-based DSSs. This design theory encompasses (meta-) design requirements, design principles, and design features in a domain-independent and interdisciplinary fashion, including end-users and developers as potential user groups. This design theory is ultimately tested through a real-world instantiation in a high-stakes maintenance scenario.
From an IS research perspective, this cumulative thesis addresses the lack of research on perception and design knowledge for an ensured utilization of XAI-based DSS. This lays the foundation for future research to obtain a holistic understanding of end-users’ heuristic behaviors during decision-making to facilitate the acceptance of XAI-based DSSs in operational practice.
The collection at hand is concerned with learning curve effects in hospitals as highly specialized expert organizations and comprises four papers, each focusing on a different aspect of the topic. Three papers are concerned with surgeons, and one is concerned with the staff of the emergency room in a conservative treatment.
The preface compactly addresses the steadily increasing health care costs and economic pressure, the hospital landscape in Germany as well as its development. Furthermore, the DRG lump-sum compensation and the characteristics of the health sector, which is strongly regulated by the state and in which ethical aspects must be omnipresent, are outlined. Besides, the benefit of knowing about learning curve effects in order to cut costs and to keep quality stable or even improve it, is addressed.
The first paper of the collection investigates the learning effects in a hospital which has specialized on endoprosthetics (total hip and knee replacement). Doing so, the specialized as well as the non-specialized interventions are studied. Costs are not investigated directly, but cost indicators. The indicator of costs in the short term are operating room times. The one of medium- to long-term costs is quality. It is operationalized by complications in the post-anesthesia care unit. The study estimates regression models (OLS and logit). The results indicate that the specialization comes along with advantages due to learning effects in terms of shorter operating room times and lower complication rates in endoprosthetic interventions. For the non-specialized interventions, the results are the same. There are no possibly negative effects of specialization on non-specialized surgeries, but advantageous spillover effects. Altogether, the specialization can be regarded as reasonable, as it cuts costs of all surgeries in the short, medium, and long term. The authors are Carsten Bauer, Nele Möbs, Oliver Unger, Andrea Szczesny, and Christian Ernst.
In the second paper surgeons’ learning curves effects in a teamwork vs. an individual work setting are in the focus of interest. Thus, the study combines learning curve effects with teamwork in health care, an issue increasingly discussed in recent literature. The investigated interventions are tonsillectomies (surgical excision of the palatine tonsils), a standard intervention. The indicator of costs in the short and medium to long term are again operating room times and complications as a proxy for quality respectively. Complications are secondary bleedings, which usually occur a few days after surgery. The study estimates regression models (OLS and logit). The results show that operating room times decrease with increasing surgeon’s experience. Surgeons who also operate in teams learn faster than the ones always operating on their own. Thus, operating room times are shorter for surgeons who also take part in team interventions. As a special feature, the data set contains the costs per case. This enables assuring that the assumed cost indicators are valid. The findings recommend team surgeries especially for resident physicians. The authors are Carsten Bauer, Oliver Unger, and Martin Holderried.
The third paper is dedicated to stapes surgery, a therapy for conductive hearing loss caused by otosclerosis (overflow bone growth). It is conceptually simple, but technically difficult. Therefore, it is regarded as the optimum to study learning curve effects in surgery. The paper seeks a comprehensive investigation. Thus, operating room times are employed as short-term cost indicator and quality as the medium to long term one. To measure quality, the postoperative difference between air and bone conduction threshold as well as a combination of this difference and the absence of complications. This paper also estimates different regression models (OLS and logit). Besides investigating the effects on department level, the study also considers the individual level, this means operating room times and quality are investigated for individual surgeons. This improves the comparison of learning curves, as the surgeons worked under widely identical conditions. It becomes apparent that the operating room times initially decrease with increasing experience. The marginal effect of additional experience gets smaller until the direction of the effect changes and the operating room times increase with increasing experience, probably caused by the allocation of difficult cases to the most experienced surgeons. Regarding quality, no learning curve effects are observed. The authors are Carsten Bauer, Johannes Taeger, and Kristen Rak.
The fourth paper is a systematic literature review on learning effects in the treatment of ischemic strokes. In case of stroke, every minute counts. Therefore, there is the inherent need to reduce the time from symptom onset to treatment. The article is concerned with the reduction of the time from arrival at the hospital to thrombolysis treatment, the so-called “door-to-needle time”. In the literature, there are studies on learning in a broader sense caused by a quality improvement program as well as learning in a narrower sense, in which learning curve effects are evaluated. Besides, studies on the time differences between low-volume and high-volume hospitals are considered, as the differences are probably the result of learning and economies of scale. Virtually all the 165 evaluated articles report improvements regarding the time to treatment. Furthermore, the clinical results substantiate the common association of shorter times from arrival to treatment with improved clinical outcomes. The review additionally discusses the economic implications of the results. The author is Carsten Bauer.
The preface brings forward that after the measurement of learning curve effects, further efforts are necessary for using them in order to increase efficiency, as the issue does not admit of easy, standardized solutions. Furthermore, the postface emphasizes the importance of multiperspectivity in research for the patient outcome, the health care system, and society.
Robotic process automation is a disruptive technology to automate already digital yet manual tasks and subprocesses as well as whole business processes rapidly. In contrast to other process automation technologies, robotic process automation is lightweight and only accesses the presentation layer of IT systems to mimic human behavior. Due to the novelty of robotic process automation and the varying approaches when implementing the technology, there are reports that up to 50% of robotic process automation projects fail. To tackle this issue, we use a design science research approach to develop a framework for the implementation of robotic process automation projects. We analyzed 35 reports on real-life projects to derive a preliminary sequential model. Then, we performed multiple expert interviews and workshops to validate and refine our model. The result is a framework with variable stages that offers guidelines with enough flexibility to be applicable in complex and heterogeneous corporate environments as well as for small and medium-sized companies. It is structured by the three phases of initialization, implementation, and scaling. They comprise eleven stages relevant during a project and as a continuous cycle spanning individual projects. Together they structure how to manage knowledge and support processes for the execution of robotic process automation implementation projects.
Ever-growing data availability combined with rapid progress in analytics has laid the foundation for the emergence of business process analytics. Organizations strive to leverage predictive process analytics to obtain insights. However, current implementations are designed to deal with homogeneous data. Consequently, there is limited practical use in an organization with heterogeneous data sources. The paper proposes a method for predictive end-to-end enterprise process network monitoring leveraging multi-headed deep neural networks to overcome this limitation. A case study performed with a medium-sized German manufacturing company highlights the method’s utility for organizations.
Bioimages frequently exhibit low signal-to-noise ratios due to experimental conditions, specimen characteristics, and imaging trade-offs. Reliable segmentation of such ambiguous images is difficult and laborious. Here we introduce deepflash2, a deep learning-enabled segmentation tool for bioimage analysis. The tool addresses typical challenges that may arise during the training, evaluation, and application of deep learning models on ambiguous data. The tool’s training and evaluation pipeline uses multiple expert annotations and deep model ensembles to achieve accurate results. The application pipeline supports various use-cases for expert annotations and includes a quality assurance mechanism in the form of uncertainty measures. Benchmarked against other tools, deepflash2 offers both high predictive accuracy and efficient computational resource usage. The tool is built upon established deep learning libraries and enables sharing of trained model ensembles with the research community. deepflash2 aims to simplify the integration of deep learning into bioimage analysis projects while improving accuracy and reliability.