11 Interfakultäre Einrichtungen
Permanent URI for this collectionhttps://elib.uni-stuttgart.de/handle/11682/12
Browse
Search Results
Item Open Access To bucket or not to bucket? : analyzing the performance and interpretability of hybrid hydrological models with dynamic parameterization(2024) Acuña Espinoza, Eduardo; Loritz, Ralf; Álvarez Chaves, Manuel; Bäuerle, Nicole; Ehret, UweHydrological hybrid models have been proposed as an option to combine the enhanced performance of deep learning methods with the interpretability of process-based models. Among the various hybrid methods available, the dynamic parameterization of conceptual models using long short-term memory (LSTM) networks has shown high potential. We explored this method further to evaluate specifically if the flexibility given by the dynamic parameterization overwrites the physical interpretability of the process-based part. We conducted our study using a subset of the CAMELS-GB dataset. First, we show that the hybrid model can reach state-of-the-art performance, comparable with LSTM, and surpassing the performance of conceptual models in the same area. We then modified the conceptual model structure to assess if the dynamic parameterization can compensate for structural deficiencies of the model. Our results demonstrated that the deep learning method can effectively compensate for these deficiencies. A model selection technique based purely on the performance to predict streamflow, for this type of hybrid model, is hence not advisable. In a second experiment, we demonstrated that if a well-tested model architecture is combined with an LSTM, the deep learning model can learn to operate the process-based model in a consistent manner, and untrained variables can be recovered. In conclusion, for our case study, we show that hybrid models cannot surpass the performance of data-driven methods, and the remaining advantage of such models is the access to untrained variables.Item Open Access Mapping the ethics of generative AI : a comprehensive scoping review(2024) Hagendorff, ThiloThe advent of generative artificial intelligence and the widespread adoption of it in society engendered intensive debates about its ethical implications and risks. These risks often differ from those associated with traditional discriminative machine learning. To synthesize the recent discourse and map its normative concepts, we conducted a scoping review on the ethics of generative artificial intelligence, including especially large language models and text-to-image models. Our analysis provides a taxonomy of 378 normative issues in 19 topic areas and ranks them according to their prevalence in the literature. The study offers a comprehensive overview for scholars, practitioners, or policymakers, condensing the ethical debates surrounding fairness, safety, harmful content, hallucinations, privacy, interaction risks, security, alignment, societal impacts, and others. We discuss the results, evaluate imbalances in the literature, and explore unsubstantiated risk scenarios.Item Open Access Metadata management in virtual product development to enable cross-organizational data analytics(2024) Ziegler, Julian; Mitschang, Bernhard (Prof. Dr.-Ing. habil.)Due to the advancing digitalization, companies are increasingly adopting computer-aided technologies. Especially in product development, computer-aided technologies enable a gradual shift from physical to virtual prototypes. This shift towards virtual product development includes design, simulation, testing, and optimization of products, and reduces costs and time needed for these tasks. Companies with strong activities in the field of virtual product development generate large amounts of heterogeneous data and wish to mine these data for knowledge. In this context, metadata is a key enabler for data discovery, data exploration, and data analyses but often neglected. The diversity in the structure and formats of virtual product development data makes it difficult for domain experts to analyze them. Domain experts struggle with this task because such engineering data are not sufficiently described with metadata. Moreover, data in companies are often isolated in data silos and difficult to explore by domain experts. This calls for an adequate data and metadata management that is able to cope with the significant data heterogeneity in virtual product development, and that empowers domain experts to discover and access data for further analyses. This thesis identifies previously unsolved challenges for a data and metadata management that is tailored to virtual product development and makes three contributions. First, a metadata model that provides a connected view on all data, metadata, and work activities of virtual product development projects. A prototypical implementation of this metadata model is already being applied to a real-world use case of an industry partner. Based on this foundation, the second contribution uses this metadata model to enable feature engineering with domain experts as part of data analyses projects. Going further, data analyses can directly use the metadata structure to provide added value without having to access the large amounts of product data. To this end, the third contribution utilizes the metadata structure itself to enable a novel approach to process discovery for product development projects. Thus, process structures in development projects can be analyzed with little effort, e.g., to identify good or inefficient processes in development projects.Item Open Access Fairness hacking : the malicious practice of shrouding unfairness in algorithms(2024) Meding, Kristof; Hagendorff, ThiloFairness in machine learning (ML) is an ever-growing field of research due to the manifold potential for harm from algorithmic discrimination. To prevent such harm, a large body of literature develops new approaches to quantify fairness. Here, we investigate how one can divert the quantification of fairness by describing a practice we call “fairness hacking” for the purpose of shrouding unfairness in algorithms. This impacts end-users who rely on learning algorithms, as well as the broader community interested in fair AI practices. We introduce two different categories of fairness hacking in reference to the established concept of p-hacking. The first category, intra-metric fairness hacking, describes the misuse of a particular metric by adding or removing sensitive attributes from the analysis. In this context, countermeasures that have been developed to prevent or reduce p-hacking can be applied to similarly prevent or reduce fairness hacking. The second category of fairness hacking is inter-metric fairness hacking. Inter-metric fairness hacking is the search for a specific fair metric with given attributes. We argue that countermeasures to prevent or reduce inter-metric fairness hacking are still in their infancy. Finally, we demonstrate both types of fairness hacking using real datasets. Our paper intends to serve as a guidance for discussions within the fair ML community to prevent or reduce the misuse of fairness metrics, and thus reduce overall harm from ML applications.