A5087735114- Acute Myeloid Leukemia Research
- Cancer Genomics and Diagnostics
- Digital Imaging for Blood Diseases
- Single-cell and spatial transcriptomics
- Imbalanced Data Classification Techniques
- Natural Language Processing Techniques
- Topic Modeling
- Clusterin in disease pathology
- Total Knee Arthroplasty Outcomes
- Radiomics and Machine Learning in Medical Imaging
- Patient-Provider Communication in Healthcare
- Artificial Intelligence in Healthcare and Education
- AI in cancer detection
- Anomaly Detection Techniques and Applications
- Metallurgical and Alloy Processes
- Delphi Technique in Research
- Magnetic and transport properties of perovskites and related materials
- SARS-CoV-2 detection and testing
- Neutropenia and Cancer Infections
- Frailty in Older Adults
- COVID-19 diagnosis using AI
- Magnetic Properties of Alloys
- Machine Learning and Data Classification
- SARS-CoV-2 and COVID-19 Research
- Generative Adversarial Networks and Image Synthesis
University Hospital Carl Gustav Carus
2021-2025
German Research Centre for Artificial Intelligence
2023-2024
TU Dresden
1999-2024
Center for Scalable Data Analytics and Artificial Intelligence
2023-2024
Zimmer Biomet (Germany)
2023-2024
Zimmer Biomet (Netherlands)
2023
Center for Systems Biology Dresden
2020
Clinical research relies on high-quality patient data, however, obtaining big data sets is costly and access to existing often hindered by privacy regulatory concerns. Synthetic generation holds the promise of effectively bypassing these boundaries allowing for simplified accessibility prospect synthetic control cohorts. We employed two different methodologies generative artificial intelligence - CTAB-GAN+ normalizing flows (NFlow) synthesize derived from 1606 patients with acute myeloid...
The use of artificial intelligence (AI) in healthcare is transforming a number medical fields, including nephrology. integration various AI techniques nephrology facilitates the prediction early detection, diagnosis, prognosis, and treatment kidney disease. Nevertheless, recent reports have demonstrated that majority published clinical studies lack uniform reporting standards, which poses significant challenges interpreting, replicating, translating into routine use. In response to these...
Oversampling is commonly used to improve classifier performance for small tabular imbalanced datasets. State-of-the-art linear interpolation approaches can be generate synthetic samples from the convex space of minority class. Generative networks are common deep learning sample generation. However, their scope on data generation in context classification not adequately explored. In this article, we show that existing generative models perform poorly compared interpolation-based problems To...
Abstract Background Total knee replacement (TKR) is one of the most commonly performed routine procedures in world. Prognostic studies indicate that number TKR will further increase constituting growing burden on healthcare systems. There also substantial regional heterogeneity rates within and between countries. Despite known therapeutic effects, a subset patients undergoing does not benefit from procedure as intended. To improve appropriateness indication, EKIT initiative (“evidence...
Current challenges of rare diseases need to involve patients, physicians, and the research community generate new insights on comprehensive patient cohorts. Interestingly, integration context has been insufficiently considered, but might tremendously improve accuracy predictive models for individual patients. Here, we conceptualized an extension European Platform Rare Disease Registration data model with contextual factors. This extended can serve as enhanced baseline is well-suited analyses...
Oversampling is commonly used to improve classifier performance for small tabular imbalanced datasets. State-of-the-art linear interpolation approaches can be generate synthetic samples from the convex space of minority class. Generative networks are common deep learning sample generation. However, their scope on data generation in context classification not adequately explored. In this article, we show that existing generative models perform poorly compared interpolation-based problems To...
We studied whether an individualized digital decision aid can improve decision-making quality for or against knee arthroplasty.
Abstract Clinical research relies on high-quality patient data, however, obtaining big data sets is costly and access to existing often hindered by privacy regulatory concerns. Synthetic generation holds the promise of effectively bypassing these boundaries allowing for simplified accessibility prospect synthetic control cohorts. We employed two different methodologies generative artificial intelligence – CTAB-GAN+ normalizing flows (NFlow) synthesize derived from 1606 patients with acute...
This study advances the utility of synthetic data in hematology, particularly for Acute Myeloid Leukemia (AML), by facilitating its integration into healthcare systems and research platforms through standardization Observational Medical Outcomes Partnership (OMOP) Fast Healthcare Interoperability Resources (FHIR) formats. In our previous work, we addressed need high-quality patient used CTAB-GAN+ Normalizing Flow (NFlow) to synthesize from 1606 patients across four multicenter AML clinical...
Abstract In the ongoing SARS-CoV-2 pandemic, there is a need for new strategies surveillance and identification of arising infection waves. Reported cases infections based on individual testing are soon deemed inaccurate due to ever changing regulations limited capacity. Wastewater epidemiology one promising solution that can be broadly applied with low efforts in comparison current large-scale individuals. Here, we combining local wastewater data from city Dresden (Germany) along reported...
Transfer Learning approaches are a promising means to analyze low-resource domain specific texts. The German SmartData corpus is the first corpus, annotated with entities from different domains, and thus allows investigate transfer learning for Named Entity Recognition (NER) on domains. In order prepare such investigations, this work includes thorough analysis of revision w.r.t. annotations split into training test data, considering distribution document entity types. Based that baseline...
Automated text analysis as named entity recognition (NER) heavily relies on large amounts of high-quality training data. Transfer learning approaches aim to overcome the problem lacking domain-specific In this paper, we investigate different transfer recognize unknown entities, including influence varying data size. The experiments are based revised German SmartData Corpus, and a baseline model, trained corpus.
Data is commonly stored in tabular format. Several fields of research are prone to small imbalanced data. Supervised Machine Learning on such data often difficult due class imbalance. Synthetic generation, i.e., oversampling, a common remedy used improve classifier performance. State-of-the-art linear interpolation approaches, as LoRAS and ProWRAS can be generate synthetic samples from the convex space minority performance cases. Deep generative networks deep learning approaches for sample...