A5057629937- AI in cancer detection
- COVID-19 diagnosis using AI
- Radiomics and Machine Learning in Medical Imaging
- Epigenetics and DNA Methylation
- Genetic Associations and Epidemiology
- Bioinformatics and Genomic Networks
- Gene expression and cancer classification
- Retinal Imaging and Analysis
- RNA and protein synthesis mechanisms
- Probiotics and Fermented Foods
- Genomics and Rare Diseases
- Traffic Prediction and Management Techniques
- Genetic and phenotypic traits in livestock
- Wireless Body Area Networks
- Advanced Authentication Protocols Security
- Advanced Chemical Sensor Technologies
- Traffic control and management
- Healthcare Policy and Management
- COVID-19 and healthcare impacts
- COVID-19 Clinical Research Studies
- Transportation Planning and Optimization
- SARS-CoV-2 detection and testing
- Gut microbiota and health
- Retinal Diseases and Treatments
- Medical Image Segmentation Techniques
University of California, Los Angeles
2020-2024
AndroScience (United States)
2022
New York University
2020
Abstract Biobanks that collect deep phenotypic and genomic data across many individuals have emerged as a key resource in human genetics. However, phenotypes biobanks are often missing individuals, limiting their utility. We propose AutoComplete, learning-based imputation method to impute or ‘fill-in’ population-scale biobank datasets. When applied collections of measured ~300,000 from the UK Biobank, AutoComplete substantially improved accuracy over existing methods. On three traits with...
Abstract Biobanks often contain several phenotypes relevant to diseases such as major depressive disorder (MDD), with partly distinct genetic architectures. Researchers face complex tradeoffs between shallow (large sample size, low specificity/sensitivity) and deep (small high phenotypes, the optimal choices are unclear. Here we propose integrate these combine benefits of each. We use phenotype imputation information across hundreds MDD-relevant which significantly increases genome-wide...
Worldwide, testing capacity for SARS-CoV-2 is limited and bottlenecks in the scale up of polymerase chain reaction (PCR-based exist. Our aim was to develop evaluate a machine learning algorithm diagnose COVID-19 inpatient setting. The based on basic demographic laboratory features serve as screening tool at hospitals where scarce or unavailable. We used retrospectively collected data from UCLA Health System Los Angeles, California. included all emergency room cases receiving PCR who also had...
Abstract Biobanks that collect deep phenotypic and genomic data across large numbers of individuals have emerged as a key resource for human genetic research. However, phenotypes acquired part are often missing many individuals, limiting the utility these datasets. The ability to accurately impute or “fill-in” is critical harness power population-scale Biobank We propose AutoComplete, learning-based imputation method which can in When applied collections measured ≈ 300K from UK Biobank,...
Microbial source tracking analysis has emerged as a widespread technique for characterizing the properties of complex microbial communities. However, this is currently limited to environments sampled in specific study. In order expand scope beyond one single study and allow exploration using large databases repositories, such Earth Microbiome Project, selection procedure required. Such will differentiating between contributing nuisance ones when number potential sources considered high....
Abstract Biobanks often contain several phenotypes relevant to a given disorder, and researchers face complex tradeoffs between shallow (high sample size, low specificity sensitivity) deep (low high sensitivity). Here, we study an extreme case: Major Depressive Disorder (MDD) in UK Biobank. Previous studies found that MDD have qualitatively distinct genetic architectures, but it remains unclear which are optimal for scientific or clinical prediction. We propose new framework get the best of...
The ability to forecast traffic congestion ahead of time given road conditions has remained a prominent problem in analysis. In this work, we leverage mobility traces public transport vehicles tracked by the New York City MTA and formulate Message-Passing Recurrent Neural Nets (MPRNN) produce long-term forecasting on data that is sparse but wide coverage. We model interactions among segments spread over entirety Manhattan, period 3 months, such can be propagated > 90% examined from just few...
<title>Abstract</title> We present SLIViT, a deep-learning framework that accurately measures disease-related risk factors in volumetric biomedical imaging, such as magnetic resonance imaging (MRI) scans, optical coherence tomography (OCT) and ultrasound videos. To evaluate we applied it to five different datasets of these three data modalities tackling seven learning tasks (including both classification regression) found consistently significantly outperforms domain-specific...
Breast cancer screening using Mammography serves as the earliest defense against breast cancer, revealing anomalous tissue years before it can be detected through physical screening. Despite use of high resolution radiography, presence densely overlapping patterns challenges consistency human-driven diagnosis and drives interest in leveraging state-of-art localization ability deep convolutional neural networks (DCNN). The growing availability digitized clinical archives enables training...
<title>Abstract</title> We present SLIViT, a deep-learning framework that accurately measures disease-related risk factors in volumetric biomedical imaging, such as magnetic resonance imaging (MRI) scans, optical coherence tomography (OCT) and ultrasound videos. To evaluate we applied it to five different datasets of these three data modalities tackling seven learning tasks (including both classification regression) found consistently significantly outperforms domain-specific...
Applying deep learning methods to mammography assessment has remained a challenging topic. Dense noise with sparse expressions, mega-pixel raw data resolution, lack of diverse examples have all been factors affecting performance. The pixel-level ground truths especially limited segmentation in pushing beyond approximately bounding regions. We propose classification approach grounded high performance tissue as an alternative all-in-one localization and models that is also capable pinpointing...
Abstract During the initial wave of COVID-19 pandemic in United States, hospitals took drastic action to ensure sufficient capacity, including canceling or postponing elective procedures, expanding number available intensive care beds and ventilators, creating regional overflow hospital capacity. However, most locations actual patients did not reach projected surge leaving available, unused As a result, may have delayed needed lost substantial revenue. These recommendations were made based...
Abstract Symptom screening is a widely deployed strategy to mitigate the COVID-19 pandemic and many public health authorities are mandating its use by employers for all employees in workplace. While symptom has benefit of reducing number infected individuals workplace, it raises some inherently difficult privacy issues as traditional approach requires employer collect data from each employee which essentially medical information. In this paper, we describe system implement Cryptographic...