A5024828367- Scientific Computing and Data Management
- Distributed and Parallel Computing Systems
- Cancer Genomics and Diagnostics
- Cancer Cells and Metastasis
- Radiomics and Machine Learning in Medical Imaging
- 3D Printing in Biomedical Research
- Computational Drug Discovery Methods
- Advanced Data Storage Technologies
- Enzyme function and inhibition
- Synthesis and biological activity
- Biomedical Text Mining and Ontologies
- Data Visualization and Analytics
- Anomaly Detection Techniques and Applications
- Gene expression and cancer classification
- Protein Degradation and Inhibitors
- Software System Performance and Reliability
- Disaster Management and Resilience
- Data Management and Algorithms
- Explainable Artificial Intelligence (XAI)
- Artificial Intelligence in Healthcare and Education
- Bioinformatics and Genomic Networks
- Cell Image Analysis Techniques
- Public Relations and Crisis Communication
- African history and culture studies
- Optimization and Search Problems
Oak Ridge National Laboratory
2024
University of Southern California
2021-2023
Hunter College
2018-2019
City University of New York
2018-2019
Icahn School of Medicine at Mount Sinai
2018
Abstract High throughput drug screening is an established approach to investigate tumor biology and identify therapeutic leads. Traditional platforms use two-dimensional cultures which do not accurately reflect the of human tumors. More clinically relevant model systems such as three-dimensional organoids can be difficult scale screen. Manually seeded coupled destructive endpoint assays allow for characterization treatment response, but capture transitory changes intra-sample heterogeneity...
Many complex diseases such as cancer are associated with multiple pathological manifestations. Moreover, the therapeutics for their treatments often lead to serious side effects. Thus, it is needed develop multi-indication that can simultaneously target clinical indications of interest and mitigate However, conventional one-drug-one-gene drug discovery paradigm emerging polypharmacology approach rarely tackle challenge design. For first time, we propose a...
Abstract Biomedical data repositories such as the Gene Expression Omnibus (GEO) enable search and discovery of relevant biomedical digital objects. Similarly, resources OMICtools, index bioinformatics tools that can extract knowledge from these However, systematic access to pre-generated ‘canned’ analyses applied by objects is currently not available. Datasets2Tools a repository indexing 31,473 canned 6,431 datasets. The also contains 4,901 published software tools, all analyzed enables...
Abstract High-throughput drug screening is an established approach to investigate tumor biology and identify therapeutic leads. Traditional platforms for high-throughput use two-dimensional cultures of immortalized cell lines which do not accurately reflect the human tumors. More clinically relevant model systems, such as three-dimensional organoids, can be difficult screen scale. For example, manually seeded organoids coupled destructive endpoint assays allow characterization response...
Fundamental progress towards reliable modern science depends on accurate anomaly detection during application execution. In this paper, we suggest a novel approach to tackle problem by applying Convolutional Neural Network (CNN) classification methods high-resolution visualizations that capture the end-to-end workflow execution timeline. Subtle differences in timeline reveal information about performance of and infrastructure’s components. We collect 1000 traces scientific workflow’s...
Scientific workflows drive most modern large-scale science breakthroughs by allowing scientists to define their computations as a set of jobs executed in given order based on data dependencies. Workflow management systems (WMSs) have become key automating scientific workflows-executing computational and orchestrating transfers between those running complex high-performance computing (HPC) platforms. Traditionally, WMSs use files communicate jobs: job writes out that are read other jobs....
Reliable execution of scientific workflows is a fundamental concern in computational campaigns. Therefore, detecting and diagnosing anomalies are both important challenging for workflow executions that span complex, distributed computing infrastructures. In this paper we model the as directed acyclic graph apply neural networks (GNNs) to identify at individual job levels. addition, generalize our GNN take into account set together anomaly detection task rather than specific workflow. By...
Abstract The National Cancer Institute’s (NCI) Surveillance, Epidemiology, and End Results (SEER) registries maintain organize cancer incidence information allowing researchers to derive valuable insights into epidemiology. While significant attention has been devoted identifying cancers either from clinical text or through tabular data collected by SEER registries, there less emphasis on integrating these distinct modes of data. In our multimodal deep learning approach, we use longitudinal...
Geospatial data come from various sources, such as satellites, aircraft, and LiDAR. The variability of the source is not limited to types acquisition techniques, we have maps different time periods. To incorporate these for a coherent analysis, it essential first align "styles" geospatial its matching images that point same location on surface Earth. In this paper, approach image registration two-step process (1) extracting contents invariant visual (and any other non-content-related)...
Abstract No existing algorithm can reliably identify metastasis from pathology reports across multiple cancer types and the entire US population. In this study, we develop a deep learning model that automatically detects patients with metastatic by using many laboratories of types. We trained validated our on cohort 29,632 four Surveillance, Epidemiology, End Results (SEER) registries linked to 60,471 unstructured reports. Our architecture task-specific data outperforms general-purpose LLM,...
An increasing number of people use social media (SM) platforms like Twitter and Instagram to report critical emergencies or disaster events. Multimodal data shared on these often contain useful information about the scale event, victims, infrastructure damage. The can provide local authorities humanitarian organizations with a big-picture understanding emergency (situational awareness). Moreover, it be used effectively timely plan relief responses. In our project, we aim address challenge...
Scientific workflows are one of the well-established pillars modern large-scale computational science. More recently, scientists have started to leverage machine learning (ML) capabilities in their workflows, leading a new category scientific denoted as ML workflows. is not only about training and inference, also involve complex data processing steps before can start, which often accounted for most performance studies. In this work, we consider from pre-processing training, model evaluation....
Abstract Although remarkable progresses have been made in the cancer treatment, existing anti-cancer drugs are associated with increasing risk of heart failure, variable drug response, and acquired resistance. To address these challenges, for first time, we develop a novel genome-scale multi-target screening platform 3D-REMAP that integrates data from structural genomics chemical as well synthesize methods bioinformatics, biophysics, machine learning. enables us to discover marked...