Thomas M. Sutter
YOU?
Author Swipe
View article: You Only Train Once: Differentiable Subset Selection for Omics Data
You Only Train Once: Differentiable Subset Selection for Omics Data Open
Selecting compact and informative gene subsets from single-cell transcriptomic data is essential for biomarker discovery, improving interpretability, and cost-effective profiling. However, most existing feature selection approaches either …
View article: Temporal Representation Learning for Real-Time Ultrasound Analysis
Temporal Representation Learning for Real-Time Ultrasound Analysis Open
Ultrasound (US) imaging is a critical tool in medical diagnostics, offering real-time visualization of physiological processes. One of its major advantages is its ability to capture temporal dynamics, which is essential for assessing motio…
View article: Predicting Pulmonary Hypertension in Newborns: A Multi-view VAE Approach
Predicting Pulmonary Hypertension in Newborns: A Multi-view VAE Approach Open
Pulmonary hypertension (PH) in newborns is a critical condition characterized by elevated pressure in the pulmonary arteries, leading to right ventricular strain and heart failure. While right heart catheterization (RHC) is the diagnostic …
View article: A Denoising VAE for Intracardiac Time Series in Ischemic Cardiomyopathy
A Denoising VAE for Intracardiac Time Series in Ischemic Cardiomyopathy Open
In the field of cardiac electrophysiology (EP), effectively reducing noise in intra-cardiac signals is crucial for the accurate diagnosis and treatment of arrhythmias and cardiomyopathies. However, traditional noise reduction techniques fa…
View article: Leveraging the Structure of Medical Data for Improved Representation Learning
Leveraging the Structure of Medical Data for Improved Representation Learning Open
Building generalizable medical AI systems requires pretraining strategies that are data-efficient and domain-aware. Unlike internet-scale corpora, clinical datasets such as MIMIC-CXR offer limited image counts and scarce annotations, but e…
View article: From Pixels to Components: Eigenvector Masking for Visual Representation Learning
From Pixels to Components: Eigenvector Masking for Visual Representation Learning Open
Predicting masked from visible parts of an image is a powerful self-supervised approach for visual representation learning. However, the common practice of masking random patches of pixels exhibits certain failure modes, which can prevent …
View article: RadVLM: A Multitask Conversational Vision-Language Model for Radiology
RadVLM: A Multitask Conversational Vision-Language Model for Radiology Open
The widespread use of chest X-rays (CXRs), coupled with a shortage of radiologists, has driven growing interest in automated CXR analysis and AI-assisted reporting. While existing vision-language models (VLMs) show promise in specific task…
View article: Weakly-Supervised Multimodal Learning on MIMIC-CXR
Weakly-Supervised Multimodal Learning on MIMIC-CXR Open
Multimodal data integration and label scarcity pose significant challenges for machine learning in medical settings. To address these issues, we conduct an in-depth evaluation of the newly proposed Multimodal Variational Mixture-of-Experts…
View article: Can Generative AI Learn Physiological Waveform Morphologies? A Study on Denoising Intracardiac Signals in Ischemic Cardiomyopathy
Can Generative AI Learn Physiological Waveform Morphologies? A Study on Denoising Intracardiac Signals in Ischemic Cardiomyopathy Open
Reducing electrophysiological (EP) signal noise is essential for diagnosis, mapping, and ablation, yet traditional approaches are suboptimal. This study tests the hypothesis that generative artificial intelligence (AI), specifically Variat…
View article: Two Is Better Than One: Aligned Representation Pairs for Anomaly Detection
Two Is Better Than One: Aligned Representation Pairs for Anomaly Detection Open
Anomaly detection focuses on identifying samples that deviate from the norm. Discovering informative representations of normal samples is crucial to detecting anomalies effectively. Recent self-supervised methods have successfully learned …
View article: Unity by Diversity: Improved Representation Learning in Multimodal VAEs
Unity by Diversity: Improved Representation Learning in Multimodal VAEs Open
Variational Autoencoders for multimodal data hold promise for many tasks in data analysis, such as representation learning, conditional generation, and imputation. Current architectures either share the encoder output, decoder input, or bo…
View article: M(otion)-mode Based Prediction of Ejection Fraction using Echocardiograms
M(otion)-mode Based Prediction of Ejection Fraction using Echocardiograms Open
Early detection of cardiac dysfunction through routine screening is vital for diagnosing cardiovascular diseases. An important metric of cardiac function is the left ventricular ejection fraction (EF), where lower EF is associated with car…
View article: Differentiable Random Partition Models
Differentiable Random Partition Models Open
Partitioning a set of elements into an unknown number of mutually exclusive subsets is essential in many machine learning problems. However, assigning elements, such as samples in a dataset or neurons in a network layer, to an unknown and …
View article: Hierarchical Organization of Single Crystal Polymorphs of Azobenzene Chromophore in Anisotropic Media Subjected to Local Thermal Gradients upon Cold Spraying
Hierarchical Organization of Single Crystal Polymorphs of Azobenzene Chromophore in Anisotropic Media Subjected to Local Thermal Gradients upon Cold Spraying Open
The present article entails the emergence of diverse crystal polymorphs following thermal quenching into various coexistence regions of binary azobenzene chromophore (ACh)/diacrylate (DA) solution and of azobenzene/nematic liquid crystal (…
View article: Learning Group Importance using the Differentiable Hypergeometric Distribution
Learning Group Importance using the Differentiable Hypergeometric Distribution Open
Partitioning a set of elements into subsets of a priori unknown sizes is essential in many applications. These subset sizes are rarely explicitly learned - be it the cluster sizes in clustering applications or the number of shared versus i…
View article: On the Limitations of Multimodal VAEs
On the Limitations of Multimodal VAEs Open
Multimodal variational autoencoders (VAEs) have shown promise as efficient generative models for weakly-supervised data. Yet, despite their advantage of weak supervision, they exhibit a gap in generative quality compared to unimodal VAEs, …
View article: Machine Learning Algorithms Evaluate Immune Response to Novel Mycobacterium tuberculosis Antigens for Diagnosis of Tuberculosis
Machine Learning Algorithms Evaluate Immune Response to Novel Mycobacterium tuberculosis Antigens for Diagnosis of Tuberculosis Open
Rationale Tuberculosis diagnosis in children remains challenging. Microbiological confirmation of tuberculosis disease is often lacking, and standard immunodiagnostic including the tuberculin skin test and interferon- γ release assay for t…
View article: Multimodal Generative Learning Utilizing Jensen-Shannon Divergence
Multimodal Generative Learning Utilizing Jensen-Shannon Divergence Open
Learning from different data types is a long-standing goal in machine learning research, as multiple information sources co-occur when describing natural phenomena. However, existing generative models that approximate a multimodal ELBO rel…
View article: A comparison of general and disease-specific machine learning models for the prediction of unplanned hospital readmissions
A comparison of general and disease-specific machine learning models for the prediction of unplanned hospital readmissions Open
Unplanned hospital readmissions are a burden to patients and increase healthcare costs. A wide variety of machine learning (ML) models have been suggested to predict unplanned hospital readmissions. These ML models were often specifically …
View article: Generation of Differentially Private Heterogeneous Electronic Health Records
Generation of Differentially Private Heterogeneous Electronic Health Records Open
Electronic Health Records (EHRs) are commonly used by the machine learning community for research on problems specifically related to health care and medicine. EHRs have the advantages that they can be easily distributed and contain many f…
View article: Multimodal Generative Learning Utilizing Jensen-Shannon Divergence
Multimodal Generative Learning Utilizing Jensen-Shannon Divergence Open
Learning from different data types is a long-standing goal in machine learning research, as multiple information sources co-occur when describing natural phenomena. However, existing generative models that approximate a multimodal ELBO rel…
View article: Feedback enhanced entanglement in a spin-1/2 XY dimer model permeated by a transverse magnetic field
Feedback enhanced entanglement in a spin-1/2 XY dimer model permeated by a transverse magnetic field Open
This paper studies the ability of feedback control to reduce the effect of decoherence and enhance entanglement in an interacting spin-1/2 XY dimer model in the presence of a transverse magnetic field. The system reaches an improved steady…