Alan Woessner1,2, Usman Anjum3, Hadi Salman4, Jacob Lear4, JT Turner5, Ross Campbell5, Laura Beaudry6, Justin Zhan3, Lawrence Cornett7,8, Susan Gauch4, and Kyle Quinn1,2
1Arkansas Integrative Metabolic Research Center, University of Arkansas, Fayetteville, AR, 2Department of Biomedical Engineering, University of Arkansas, Fayetteville, AR, 3Department of Computer Science, University of Cincinnati, Cincinnati, OH, 4Department of Computer Science and Computer Engineering, University of Arkansas, Fayetteville, AR, 5Deliotte Consulting LLP, New York, NY, 6Google, Mountain View, CA, 7IDeA Network of Biomedical Research Excellence, University of Arkansas for Medical Sciences, Little Rock, AR, 8Department of Physiology and Cell Biology, University of Arkansas for Medical Sciences, Little Rock, AR
Introduction/Background. Biomedical-related datasets are widely used in both research and clinical settings. However, as the size and breadth of these datasets increases, the ability for professionally trained clinicians and researchers to interpret datasets becomes difficult and prone to systematic or user bias. Artificial intelligence (AI), and specifically deep learning convolutional neural networks (CNNs), have recently become an important tool in novel biomedical research, but are hard to utilize due to their computational requirements and confusion regarding different neural network architectures.
Hypothesis/Goal of Study. The goal of this learning module is to provide a gentle introduction to the types of deep learning neural networks and practices that are commonly used in biomedical research via public and published datasets.
Methods and Results. This module is subdivided into four submodules that cover image classification CNNs, data augmentation, image segmentation CNNs, and regression CNNs. Each complementary submodule was written as a Jupyter notebook on the Google Cloud Platform (GCP) and contains detailed code and explanations, as well as quizzes and challenges to facilitate user training.
Discussion/Conclusions. Overall, the goal of this learning module is to enable users to identify and integrate the correct type of neural network with their data while highlighting the ease-of-use of cloud computing for implementing neural networks.
Citation/Acknowledgements. We would like to acknowledge the following funding grants for this work: 3P20GM103429-21S2, R01AG056560, R01EB031032, P20GM139768.