A5044596225- Topic Modeling
- Diabetes Management and Research
- Explainable Artificial Intelligence (XAI)
- Machine Learning in Healthcare
- Heart Rate Variability and Autonomic Control
- Big Data and Business Intelligence
- Corporate Social Responsibility Reporting
- Expert finding and Q&A systems
- Sentiment Analysis and Opinion Mining
- Non-Invasive Vital Sign Monitoring
- Computational and Text Analysis Methods
- Forecasting Techniques and Applications
- ECG Monitoring and Analysis
- Nonlinear Dynamics and Pattern Formation
- Sustainable Finance and Green Bonds
- Stock Market Forecasting Methods
- Natural Language Processing Techniques
- Climate Change Communication and Perception
- Artificial Intelligence in Healthcare
- Chaos control and synchronization
- Industrial Vision Systems and Defect Detection
- Sepsis Diagnosis and Treatment
- Machine Learning and Data Classification
- Advanced Text Analysis Techniques
- Energy, Environment, Economic Growth
University of Regensburg
2024-2025
Friedrich-Alexander-Universität Erlangen-Nürnberg
2021-2024
ETH Zurich
2018-2024
Institut für Höhere Studien - Institute for Advanced Studies (IHS)
2023
Schiller International University
2023
École Polytechnique Fédérale de Lausanne
2022
University Medical Center Utrecht
2022
Board of the Swiss Federal Institutes of Technology
2020
University of Freiburg
2017
University of Würzburg
1989-1999
Business analytics refers to methods and practices that create value through data for individuals, firms, organizations. This field is currently experiencing a radical shift due the advent of deep learning: neural networks promise improvements in prediction performance as compared models from traditional machine learning. However, our research into existing body literature reveals scarcity works utilizing learning discipline. Accordingly, objectives this overview article are follows: (1) we...
Disclosure of climate-related financial risks greatly helps investors assess companies’ preparedness for climate change. Voluntary disclosures such as those based on the recommendations Task Force Climate-related Financial Disclosures (TCFD) are being hailed an effective measure better risk management. We ask whether this expectation is justified. do so by training ClimateBERT, a deep neural language model fine-tuned BERT. In analyzing TCFD-supporting firms, ClimateBERT comes to sobering...
Over the recent years, large pretrained language models (LM) have revolutionized field of natural processing (NLP). However, while pretraining on general has been shown to work very well for common language, it observed that niche poses problems. In particular, climate-related texts include specific LMs can not represent accurately. We argue this shortcoming today's limits applicability modern NLP broad text texts. As a remedy, we propose ClimateBert, transformer-based model is further over...
Across many industries, visual quality assurance has transitioned from a manual, labor-intensive, and error-prone task to fully automated precise assessment of industrial quality. This transition been made possible due advances in machine learning general, supervised particular. However, the majority approaches only allow identify pre-defined categories, such as certain error types on manufactured objects. New, unseen are unlikely be detected by models. As remedy, this work studies...
Abstract Large Language Models have made remarkable progress in question-answering tasks, but challenges like hallucination and outdated information persist. These issues are especially critical domains climate change, where timely access to reliable is vital. One solution granting these models external, scientifically accurate sources enhance their knowledge reliability. Here, we GPT-4 by providing the Sixth Assessment Report of Intergovernmental Panel on Climate Change (IPCC AR6), most...
The ability to understand and explain the outcomes of data analysis methods, with regard aiding decision-making, has become a critical requirement for many applications. For example, in operational research domains, analytics have long been promoted as way enhance decision-making. This study proposes comprehensive, normative framework define explainable artificial intelligence (XAI) (XAIOR) reconciliation three subdimensions that constitute its requirements: performance, attributable,...
Navigating the complex landscape of corporate climate disclosures and their real impacts is crucial for managing climate-related financial risks. However, current oftentimes suffer from imprecision, inaccuracy, greenwashing. We introduce ClimateBert CTI, a deep learning algorithm, to identify cheap talk in MSCI World index firms' annual reports. find that only targeted engagement associated with less talk. Voluntary are more Moreover, correlates increased negative news coverage higher...
Predicting the remaining useful life of machinery, infrastructure, or other equipment can facilitate preemptive maintenance decisions, whereby a failure is prevented through timely repair replacement. This allows for better decision support by considering anticipated time-to-failure and thus promises to reduce costs. Here common baseline may be derived fitting probability density function past lifetimes then utilizing (conditional) expected as prognostic. approach finds widespread use in...
To develop a noninvasive hypoglycemia detection approach using smartwatch data.We prospectively collected data from two wrist-worn wearables (Garmin vivoactive 4S, Empatica E4) and continuous glucose monitoring values in adults with diabetes on insulin treatment. Using these data, we developed machine learning (ML) to detect (<3.9 mmol/L) noninvasively unseen individuals solely based wearable data.Twenty-two were included the final analysis (age 54.5 ± 15.2 years, HbA1c 6.9 0.6%, 16 males)....
Environmental, social, and governance (ESG) criteria take a central role in fostering sustainable development economies. This paper introduces class of novel Natural Language Processing (NLP) models to assess corporate disclosures the ESG subdomains. Using over 13.8 million texts from reports news, specific E, S, G were pretrained. Additionally, three 2k datasets developed classify ESG-related texts. The effectively explain variations ratings, showcasing robust method for enhancing...
Large language models (LLMs) have significantly transformed the landscape of artificial intelligence by demonstrating their ability to generate human-like text across diverse topics. However, despite impressive capabilities, LLMs lack recent information and often employ imprecise language, which can be detrimental in domains where accuracy is crucial, such as climate change. In this study, we make use ideas harness potential viewing them agents that access multiple sources, including...
We propose Interpretable Generalized Additive Neural Networks (IGANN), a novel machine learning model that uses gradient boosting and tailored neural networks to obtain high predictive performance while being interpretable humans. derive an efficient training algorithm based on the theory of extreme machines, allows reducing process solving sequence regularized linear regressions. analyze theoretically, provide insights into rate change so-called shape functions, show computational...
Micro- and macrovascular complications are a major burden for individuals with diabetes can already arise in prediabetic state. To allocate effective treatments to possibly prevent these complications, identification of those at risk is essential.This study aimed build machine learning (ML) models that predict the developing micro- or complication prediabetes diabetes.In this study, we used electronic health records from Israel contain information about demographics, biomarkers, medications,...
Jingwei Ni, Julia Bingler, Chiara Colesanti-Senni, Mathias Kraus, Glen Gostlow, Tobias Schimanski, Dominik Stammbach, Saeid Ashraf Vaghefi, Qian Wang, Nicolas Webersinke, Wekhof, Tingyu Yu, Markus Leippold. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: System Demonstrations. 2023.
Rigorous blood glucose management is vital for individuals with diabetes to prevent states of too low (hypoglycemia). While there are continuous monitors available, they expensive and not available many patients. Related work suggests a correlation between the level physiological measures, such as heart rate variability. We therefore propose machine learning model detect hypoglycemia on basis data from smartwatch sensors gathered in proof-of-concept study. In further work, we want integrate...
The climate impact of AI, and NLP research in particular, has become a serious issue given the enormous amount energy that is increasingly being used for training running computational models. Consequently, increasing focus placed on efficient NLP. However, this important initiative lacks simple guidelines would allow systematic reporting research. We argue deficiency one reasons why very few publications report key figures more thorough examination environmental impact, present quantitative...
Abstract Proactive analysis of patient pathways helps healthcare providers anticipate treatment-related risks, identify outcomes, and allocate resources. Machine learning (ML) can leverage a patient’s complete health history to make informed decisions about future events. However, previous work has mostly relied on so-called black-box models, which are unintelligible humans, making it difficult for clinicians apply such models. Our introduces PatWay-Net, an ML framework designed...
To develop and evaluate the concept of a non-invasive machine learning (ML) approach for detecting hypoglycaemia based exclusively on combined driving (CAN) eye tracking (ET) data.We first developed tested our ML in pronounced hypoglycaemia, then we applied it to mild its early warning potential. For this, conducted two consecutive, interventional studies individuals with type 1 diabetes. In study (n = 18), collected CAN ET data simulator during euglycaemia (blood glucose [BG] 2.0-2.5 mmol...
Advances towards more faithful and traceable answers of Large Language Models (LLMs) are crucial for various research practical endeavors. One avenue in reaching this goal is basing the on reliable sources. However, Evidence-Based QA has proven to work insufficiently with LLMs terms citing correct sources (source quality) truthfully representing information within (answer attributability). In work, we systematically investigate how robustly fine-tune better source quality answer...