When do we Clean, When do we Render, When do we Model? — Applying Novel Visualization and Data Science techniques to Biology and Medical Imaging Applications

This seminar took place through Zoom on October 22th, 2020 from 1:00 pm to 2:00 pm.
Presented by: Curtis Lisle, Ph.D., CEO, KnowledgeVis, LLC

Click Here to Access Seminar Recording

Follow these steps to gain access to the recording:

  1. Click on the link above and login to Zoom with SSO using UCF as your company domain
  2. Enter your NID username and password

When do we Clean, When do we Render, When do we Model? — Applying Novel Visualization and Data Science techniques to Biology and Medical Imaging Applications
Over the past few years, data science technology has matured and become more useful in many different scientific applications. This talk covers the use of data science techniques in several scientific domains and is suitable to attendees with varied backgrounds. The talk will begin by covering recent experience with an AutoML (automatic machine learning) system developed by DARPA’s Data-Driven Discovery of Models (D3M) program. We will discuss the architecture of the D3M system, show examples, and describe how this open-source system is becoming available for anyone to use. However, even with a leading AutoML system, when models are applied to analyze raw scientific data, much effort is often required to prepare the data before good modeling results can be achieved. We will illustrate the data preparation process using a large US Department of Energy TERRA-REF (Transportation Energy Resources from Renewable Agriculture Phenotyping Reference Platform) agricultural growth dataset. Finally, we will cover current work using deep learning techniques applied to medical imaging (both radiology and histology datasets using the PyTorch framework and discuss the newly-emerging MONAI framework (Medical Open Network for AI), supported by NVIDIA and internationally-respected data scientists. Attendees will learn data science and model development techniques through several examples and will be exposed to existing machine learning frameworks they may use later on their own projects.

Dr. Curtis Lisle is an experienced computer scientist working in the deep learning, data science, scientific visualization, and parallel computing communities. Dr. Lisle is the Principal of KnowledgeVis, LLC — a data science, imaging, and visualization consulting company in Altamonte Springs. He has consulted for the NIH/National Cancer Institute (NCI) Intramural Research program for multiple years. KnowledgeVis has performed leading ¬edge research, including scientific workflows for biology funded by NSF, imaging and visualization for NIH/NCI, and developing data science interfaces for DARPA. In the radiology imaging community, Dr. Lisle has also contributed to the National Alliance for Medical Image Computing project, developing radiology imaging components for the popular 3D Slicer imaging analysis system. Dr. Lisle completed his Masters and Ph.D. in Computer Science at UCF in 1990 and 1998, respectively. He held an appointment as a Research Faculty member at IST while pursuing his Doctoral degree. Dr. Lisle is a UCF Graduate Scholar and recently co¬-supervised Ph.D. students in Computer Science with Dr. Charles Hughes and Dr. Ulas Bagci. Dr. Lisle has also served as an Adjunct Faculty member for UCF’s Department of Industrial Engineering and Management Systems. In addition to UCF, Dr. Lisle’s career includes real-time visualization and parallel processing at Silicon Graphics Computer Systems and General Electric.