A5043039867- Advanced Neuroimaging Techniques and Applications
- Galaxies: Formation, Evolution, Phenomena
- Gamma-ray bursts and supernovae
- MRI in cancer diagnosis
- Glioma Diagnosis and Treatment
- Astronomy and Astrophysical Research
- Statistical and numerical algorithms
- Astrophysical Phenomena and Observations
- Fetal and Pediatric Neurological Disorders
- Dermatologic Treatments and Research
- Anomaly Detection Techniques and Applications
- Stellar, planetary, and galactic studies
- Bone and Joint Diseases
- Particle Detector Development and Performance
- Laser Applications in Dentistry and Medicine
- Body Contouring and Surgery
- Algebraic structures and combinatorial models
- Astronomical Observations and Instrumentation
- Radiomics and Machine Learning in Medical Imaging
- Multiple Sclerosis Research Studies
- Algebraic Geometry and Number Theory
- Image and Object Detection Techniques
- CCD and CMOS Imaging Sensors
- Hepatocellular Carcinoma Treatment and Prognosis
- Neural Networks and Reservoir Computing
University of Southern California
2019-2024
Western Sydney University
2024
University of Illinois Urbana-Champaign
2020-2023
LAC+USC Medical Center
2022
National Center for Supercomputing Applications
2022
Simons Foundation
2022
Keck Hospital of USC
2022
Cypress College
2022
Washington University in St. Louis
2019-2021
Abstract Objective Multiple sclerosis (MS) lesions are heterogeneous with regard to inflammation, demyelination, axonal injury, and neuronal loss. We previously developed a diffusion basis spectrum imaging (DBSI) technique better address MS lesion heterogeneity. hypothesized that the profiles of multiple DBSI metrics can identify lesion‐defining patterns. Here we test this hypothesis by combining deep learning algorithm using neural network (DNN) other methods. Methods Thirty‐eight patients...
We investigate the use of approximate Bayesian neural networks (BNNs) in modeling hundreds time-delay gravitational lenses for Hubble constant ($H_0$) determination. Our BNN was trained on synthetic HST-quality images strongly lensed active galactic nuclei (AGN) with lens galaxy light included. The can accurately characterize posterior PDFs model parameters governing elliptical power-law mass profile an external shear field. then propagate BNN-inferred into ensemble $H_0$ inference, using...
There is a shortage of multi-wavelength and spectroscopic followup capabilities given the number transient variable astrophysical events discovered through wide-field, optical surveys such as upcoming Vera C. Rubin Observatory. From haystack potential science targets, astronomers must allocate scarce resources to study selection needles in real time. Here we present variational recurrent autoencoder neural network encode simulated Observatory extragalactic using 1% PLAsTiCC dataset train...
ABSTRACT Protein therapeutic design and property prediction are frequently hampered by data scarcity. Here we propose a new model, DyAb, that addresses these issues leveraging pair-wise representation to predict differences in protein properties, rather than absolute values. DyAb is built on top of pre-trained language model achieves Spearman rank correlation up 0.85 binding affinity across molecules targeting three different antigens (EGFR, IL-6, an internal target), given as few 100...
Abstract Purpose: Glioblastoma (GBM) is one of the deadliest cancers with no cure. While conventional MRI has been widely adopted to examine GBM clinically, accurate neuroimaging assessment tumor histopathology for improved diagnosis, surgical planning, and treatment evaluation remains an unmet need in clinical management GBMs. Experimental Design: We employ a novel diffusion histology imaging (DHI) approach, combining basis spectrum (DBSI) machine learning, detect, differentiate, quantify...
Abstract High-grade pediatric brain tumors exhibit the highest cancer mortality rates in children. While conventional MRI has been widely adopted for examining high-grade clinically, accurate neuroimaging detection and differentiation of tumor histopathology improved diagnosis, surgical planning, treatment evaluation, remains an unmet need their clinical management. We employed a novel Diffusion Histology Imaging (DHI) approach employing diffusion basis spectrum imaging (DBSI) derived...
Automated searches for strong gravitational lensing in optical imaging survey datasets often employ machine learning and deep approaches. These techniques require more example systems to train the algorithms than have presently been discovered, which creates a need simulated images as training dataset supplements. This work introduces summarizes deeplenstronomy, an open-source Python package that enables efficient, large-scale, reproducible simulation of astronomical systems. A full suite...
Quantifying the morphology of galaxies has been an important task in astrophysics to understand formation and evolution galaxies. In recent years, data size dramatically increasing due several on-going upcoming surveys. Labeling identifying interesting objects for further investigations explored by citizen science through Galaxy Zoo Project machine learning particular with convolutional neural networks (CNNs). this work, we explore usage Vision Transformer (ViT) galaxy classification first...
Quantifying the parameters and corresponding uncertainties of hundreds strongly lensed quasar systems holds key to resolving one most important scientific questions: Hubble constant ($H_{0}$) tension. The commonly used Markov chain Monte Carlo (MCMC) method has been too time-consuming achieve this goal, yet recent work shown that convolution neural networks (CNNs) can be an alternative with seven orders magnitude improvement in speed. With 31,200 simulated images, we explore usage Vision...
Dark matter substructures are interesting since they can reveal the properties of dark matter. Collisionless N-body simulations cold show more compared with population dwarf galaxy satellites observed in our local group. Therefore, understanding and property subhalos at cosmological scale would be an test for In recent years, it has become possible to detect individual near images strongly lensed extended background galaxies. this work, we discuss possibility using deep neural networks...
Abstract Background: High-grade pediatric brain tumors exhibit the highest cancer mortality rates in children. While conventional MRI has been widely adopted for examining high-grade clinically, accurate neuroimaging detection and differentiation of tumor histopathology improved diagnosis, surgical planning, treatment evaluation, remains an unmet need their clinical management. Methods: We employed a novel Diffusion Histology Imaging (DHI) approach employing diffusion basis spectrum imaging...
The Event Horizon Telescope (EHT) recently released the first horizon-scale images of black hole in M87. Combined with other astronomical data, these constrain mass and spin as well accretion rate magnetic flux trapped on hole. An important question for EHT is how key parameters such can be extracted from present future data alone. Here we explore parameter extraction using a neural network trained high resolution synthetic drawn state-of-the-art simulations. We find that able to recover...
Introduction: Treatment of vestibular schwannoma (VS) has been extensively studied, but a gap in knowledge exists demonstrating how racial and socioeconomic status influence VS presentation. Our institution unique setting with public safety net hospital (PSNH) tertiary academic medical center (TAMC) the same zip code, which we study to evaluate initial presentation disparities patient populations presenting these settings. Methods: Retrospective chart review was performed all adult patients...
Abstract High-grade pediatric brain tumors constitute the highest mortality of cancer-death in children. While conventional MRI has been widely adopted for examining high-grade tumor clinically, accurate neuroimaging detection and differentiation histopathology improved diagnosis, surgical planning, treatment evaluation, remains an unmet need clinical management tumor. We employed a novel Diffusion Histology Imaging (DHI) approach that incorporates diffusion basis spectrum imaging (DBSI)...
Supernovae mark the explosive deaths of stars and enrich cosmos with heavy elements. Future telescopes will discover thousands new supernovae nightly, creating a need to flag astrophysically interesting events rapidly for followup study. Ideally, such an anomaly detection pipeline would be independent our current knowledge sensitive unexpected phenomena. Here we present unsupervised method search anomalous time series in real transient, multivariate, aperiodic signals. We use RNN-based...
We investigate Siamese networks for learning related embeddings augmented samples of molecular conformers. find that a non-contrastive (positive-pair only) auxiliary task aids in supervised training Euclidean neural (E3NNs) and increases manifold smoothness (MS) around point-cloud geometries. demonstrate this property multiple drug-activity prediction tasks while maintaining relevant performance metrics, propose an extension MS to probabilistic regression settings. provide analysis...
ABSTRACT Supermassive black holes (SMBHs) are commonly found at the centres of most massive galaxies. Measuring SMBH mass is crucial for understanding origin and evolution SMBHs. Traditional approaches, on other hand, necessitate collection spectroscopic data, which costly. We present an algorithm that weighs SMBHs using quasar light time series information, including colours, multiband magnitudes, variability curves, circumventing need expensive spectra. train, validate, test neural...
Abstract Purpose Glioblastoma (GBM) is one of the deadliest cancers with no cure. While conventional MRI has been widely adopted for examining GBM clinically, accurate neuroimaging assessment tumor histopathology improved diagnosis, surgical planning, and treatment evaluation, remains an unmet need in clinical management GBMs. Experimental Design We employ a novel Diffusion Histology Imaging (DHI) approach, combining diffusion basis spectrum imaging (DBSI) machine learning, to detect,...