COVID+DS: Simple introduction to deep learning

A key aspect of analysis of data involves classification and regression; these play a key role in the analysis of many types of data connected to COVID-19. To perform such analyses, one typically must extract features from the data, with which classification/regression is performed. Rather than relying on human-generated ("handcrafted") features, deep learning is a data-driven approach, in which the algorithm learns the best class of features from the data itself. Such an approach mitigates the human biases inherent to using handcrafted features. In this session we will discuss the fundamental principles of deep learning, with an emphasis on intuition, and with as little math as possible. The session is meant to be accessible to a wide audience interested in analyzing data, not just for quantitative specialists.

This session is part of the Duke+Data Science (+DS) program virtual series on COVID-19 + Data Science. Please join us for a 8-week series on data science methods with direct applications to the COVID-19 pandemic. Learn from Duke experts about the state-of-the-art in these 1-hour virtual sessions. For more information, please visit https://plus.datascience.duke.edu