The Data Science Core specializes in artificial intelligence-based approaches to elucidate relationships between large imaging, bioenergetics, genomic, and proteomic data sets.

Data acquired from the AIMRC Imaging and Spectroscopy Core and Bioenergetics Core can be directly uploaded to a cluster storage sever for use by data analysis tools and computational resources (including a dedicated AIMRC GPU cluster).

Key Equipment Contained in the Data Science Core:

Image of server racks at the Arkansas high-performance computing center.
  • Eighteen single A100 GPU dual CPU servers for DART.
  • Nineteen single V100 GPU dual CPU servers and about 400 non-GPU dual CPU servers.
  • Three Petabytes of DNN Lustre parallel storage, four PB of nearline Lustre and object storage, and InfiniBand and Ethernet networks to act as a single system.
  • 450 application packages installed from source or commercial sources, about 150 packages installed as Python or R modules, and hundreds of packages installed as RPMs (system packages) from CentOS, EPEL, and OpenHPC.
  • An OpenOnDemand web portal, Jupter Notebooks, RStudio, and other graphical interfaces in addition to command line batch computing.
Image of a dell xe8545 dual cpu and quad gpu server, nvidia a100 gpu, and an amd 7543 cpu.
  • 450 application packages installed from source or commercial sources, about 150 packages installed as Python or R modules, and hundreds of packages installed as RPMs (system packages) from CentOS, EPEL, and OpenHPC.
  • An OpenOnDemand web portal, Jupyter Notebooks, RStudio, and other graphical interfaces in addition to command line batch computing.
Image of a hyperplane 8 server.
  • Operating system: Ubuntu 20.04 + Lambda Stack
  • Software: TensorFlow, PyTorch, Caffe, Keras, CUDA, cuDNN

Data Science Core Services:

  • Provide foundational training for those getting started with high-performance computing and Arkansas Research Platform (ARP). Topics would cover OpenOnDemand portal, running jobs on terminal, modules, file systems, moving data, bioimaging software support on ARP, and demos.
  • Provide training for Python programming, basic data mining, and machine learning.
  • Provide training of open-source deep learning based biomedical imaging resources, including DeepImageJ, CSBDeep, ZeroCostDL4Mic, and Bioimage Model Zoo.
  • Provide customized deep learning solutions for biomedical imaging analysis tasks such as image restoration, image segmentation, and image quantification and tracking.
  • Provide customized solutions for multi-omics data integration and analysis.
  • Data Science Core staff are available to assist center members with their research by developing novel algorithms and quantitative analysis pipelines for large datasets.

Please contact our Data Science Core Director or Core Manager to schedule access to the equipment or discuss your research needs.

Core Director


Image of Dr. Xintao Wu.

Dr. Xintao Wu

Charles D. Morgan/Acxiom Endowed Graduate Research Chair
Electrical Engineering & Computer Science
College of Engineering
University of Arkansas
xintaowu@uark.edu
Profile

Core Manager


Image of Dr. Prateek Verma.

Dr. Prateek Verma

Post-Doctoral Fellow
Electrical Engineering and Computer Science
College of Engineering
University of Arkansas
prateek@uark.edu
Profile