A5018488407- Medical Image Segmentation Techniques
- Gaussian Processes and Bayesian Inference
- Generative Adversarial Networks and Image Synthesis
- Adversarial Robustness in Machine Learning
- Domain Adaptation and Few-Shot Learning
- Reinforcement Learning in Robotics
- Anomaly Detection Techniques and Applications
- Brain Tumor Detection and Classification
- Bayesian Modeling and Causal Inference
- Advanced Neural Network Applications
- Information and Cyber Security
- Radiomics and Machine Learning in Medical Imaging
- Machine Learning and Data Classification
- Robot Manipulation and Learning
- Machine Learning and Algorithms
- Model Reduction and Neural Networks
- Digital Media Forensic Detection
- Time Series Analysis and Forecasting
- Machine Learning in Healthcare
- Explainable Artificial Intelligence (XAI)
- Cell Image Analysis Techniques
- Glioma Diagnosis and Treatment
- Machine Learning in Materials Science
- COVID-19 diagnosis using AI
- AI in cancer detection
Technical University of Munich
2024-2025
Helmholtz Zentrum München
2024-2025
Munich Center for Machine Learning
2025
Max Planck Institute for Intelligent Systems
2018-2023
Iomedico (Germany)
2023
KTH Royal Institute of Technology
2021-2022
Vienna University of Economics and Business
2012-2022
Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)
2019-2022
GlaxoSmithKline (Switzerland)
2020-2022
Canadian Institute for Advanced Research
2020-2022
The two fields of machine learning and graphical causality arose are developed separately. However, there is, now, cross-pollination increasing interest in both to benefit from the advances other. In this article, we review fundamental concepts causal inference relate them crucial open problems learning, including transfer generalization, thereby assaying how can contribute modern research. This also applies opposite direction: note that most work starts premise variables given. A central...
The key idea behind the unsupervised learning of disentangled representations is that real-world data generated by a few explanatory factors variation which can be recovered algorithms. In this paper, we provide sober look at recent progress in field and challenge some common assumptions. We first theoretically show fundamentally impossible without inductive biases on both models data. Then, train more than 12000 covering most prominent methods evaluation metrics reproducible large-scale...
High-entropy alloys are solid solutions of multiple principal elements that capable reaching composition and property regimes inaccessible for dilute materials. Discovering those with valuable properties, however, too often relies on serendipity, because thermodynamic alloy design rules alone fail in high-dimensional spaces. We propose an active learning strategy to accelerate the high-entropy Invar a practically infinite compositional space based very sparse data. Our approach works as...
The classification of sleep stages is the first and an important step in quantitative analysis polysomnographic recordings. Sleep stage scoring relies heavily on visual pattern recognition by a human expert time consuming subjective. Thus, there need for automatic classification. In this work we developed machine learning algorithms classification: random forest based features artificial neural networks working both with raw data. We tested our methods healthy subjects patients. Most yielded...
Background and Purpose Reproducible segmentation of brain tumors on magnetic resonance images is an important clinical need. This study was designed to evaluate the reliability a novel fully automated tool for tumor image analysis in comparison manually defined segmentations. Methods We prospectively evaluated preoperative MR Images from 25 glioblastoma patients. Two independent expert raters performed manual Automatic segmentations were using Brain Tumor Image Analysis software (BraTumIA)....
In organizations, users' compliance with information security policies (ISP) is crucial for minimizing (IS) incidents. To improve compliance, IS managers have implemented awareness (ISA) programs, which are systematically planned interventions to continuously transport a target audience. The underlying research analyzes managers' efforts design effective ISA programs by comparing current recommendations suggested scientific literature actual practices of in three banks. Moreover, this study...
Information about the size of a tumor and its temporal evolution is needed for diagnosis as well treatment brain patients. The aim study was to investigate potential fully-automatic segmentation method, called BraTumIA, longitudinal volumetry by comparing automatically estimated volumes with ground truth data acquired via manual segmentation. Longitudinal Magnetic Resonance (MR) Imaging 14 patients newly diagnosed glioblastoma encompassing 64 MR acquisitions, ranging from preoperative up 12...
COVID-19 is a rapidly emerging respiratory disease caused by SARS-CoV-2. Due to the rapid human-to-human transmission of SARS-CoV-2, many health care systems are at risk exceeding their capacities, in particular terms SARS-CoV-2 tests, hospital and intensive unit (ICU) beds, mechanical ventilators. Predictive algorithms could potentially ease strain on identifying those who most likely receive positive test, be hospitalized, or admitted ICU.The aim this study develop, study, evaluate...
Several clinical trials have recently proven the efficacy of mechanical thrombectomy for treating ischemic stroke, within a six-hour window therapy. To move beyond treatment windows and toward personalized risk assessment, it is essential to accurately identify extent tissue-at-risk ("penumbra"). We introduce fully automated method estimate penumbra volume using multimodal MRI (diffusion-weighted imaging, T2w- T1w contrast-enhanced sequence, dynamic susceptibility contrast perfusion MRI)....
Estimating what would be an individual's potential response to varying levels of exposure a treatment is high practical relevance for several important fields, such as healthcare, economics and public policy. However, existing methods learning estimate counterfactual outcomes from observational data are either focused on estimating average dose-response curves, or limited settings with only two treatments that do not have associated dosage parameter. Here, we present novel machine-learning...
The two fields of machine learning and graphical causality arose developed separately. However, there is now cross-pollination increasing interest in both to benefit from the advances other. In present paper, we review fundamental concepts causal inference relate them crucial open problems learning, including transfer generalization, thereby assaying how can contribute modern research. This also applies opposite direction: note that most work starts premise variables are given. A central...
Coronavirus Disease 2019 (COVID-19) is an emerging respiratory disease caused by the severe acute syndrome coronavirus 2 (SARS-CoV-2) with rapid human-to-human transmission and a high case fatality rate particularly in older patients. Due to exponential growth of infections, many healthcare systems across world are under pressure care for increasing amounts at-risk Given number infected patients, identifying patients highest mortality risk early critical enable effective intervention optimal...
Chemical short-range order (CSRO) refers to atoms of specific elements self-organising within a disordered crystalline matrix form particular atomic neighbourhoods. CSRO is typically characterized indirectly, using volume-averaged or through projection microscopy techniques that fail capture the three-dimensional atomistic architectures. Here, we present machine-learning enhanced approach break inherent resolution limits atom probe tomography enabling imaging multiple CSROs. We showcase our...
Image-based modeling of tumor growth combines methods from cancer simulation and medical imaging. In this context, we present a novel approach to adapt healthy brain atlas MR images patients. order establish correspondence between pathologic patient image, in combination with registration algorithms is employed. first step, the grown based on new multiscale, multiphysics model including cellular level up biomechanical level, accounting for cell proliferation tissue deformations. Large-scale...
Abstract Reproducible definition and quantification of imaging biomarkers is essential. We evaluated a fully automatic MR-based segmentation method by comparing it to manually defined sub-volumes experienced radiologists in the TCGA-GBM dataset, terms sub-volume prognosis association with VASARI features. MRI sets 109 GBM patients were downloaded from Cancer Imaging archive. sub-compartments automatically using Brain Tumor Image Analysis (BraTumIA). Spearman’s correlation was used evaluate...
Despite the importance of information security, far too many organizations, in particular banks, are facing behavioral security incidents. In context given by headquarters a large European banking organization, this single case study investigates whether individual compliance with policy is influenced accumulated and awareness embedded within theory reasoned action an extended norms approach. We collected empirical data through three-staged process which we conducted semi-structured...
Understanding sleep and its perturbation by environment, mutation, or medication remains a central problem in biomedical research. Its examination animal models rests on brain state analysis via classification of electroencephalographic (EEG) signatures. Traditionally, these states are classified trained human experts visual inspection raw EEG recordings, which is laborious task prone to inter-individual variability. Recently, machine learning approaches have been developed automate this...
Contrast enhancement is an important preprocessing technique for improving the performance of downstream tasks in image processing and computer vision. Among existing approaches based on nonlinear histogram transformations, contrast limited adaptive equalization (CLAHE) a popular choice dealing with 2D images obtained natural scientific settings. The recent hardware upgrade data acquisition systems results significant increase complexity, including their sizes dimensions. Measurements...