Variational autoencoders (VAEs) are a family of deep, probabilistic models with emerging use-cases in the analysis of high-dimensional genomics data. In this talk I will cover two topics: I will start by providing a primer on the mathematical foundations behind VAEs and present them through two lenses: First, as probabilistic, latent variable models, and second, as a type of autoencoding neural network. In the second part, I will discuss how VAEs are being adapted and applied to the analysis of single-cell genomics data in order to find latent representations of cells.
Dr. Matthew Bernstein is Principal Scientist, Computational Biology at Immunitas Therapeutics. He develops computational and statistical methods to analyze large scale single-cell and spatial transcriptomics datasets to discover new immunotherapies for challenging cancers. Prior to joining Immunitas, Dr. Bernstein completed a Ph.D. in Computer Sciences from the University of Wisconsin - Madison and a postdoctoral fellowship at the Morgridge Institute for Research. Broadly, he is interested in applications of machine learning to the biomedical sciences.