Cross-modal Deep Variational Autoencoders

Eligible: Masters students

Mentor: Michal Muszynski

Description: The mail goal of this project is to develop new variational autoencoder techniques to learn a representation of physiological or behavioral signals in a cross-modal latent space. For example, this latent space can be directly used to estimate and synthesize human physiological and behavioral reactions.

Skills/Experience: Prior machine learning experience, intermediate skills in Matlab/Python/R programming, and willingness to learn new deep machine learning techniques and advanced statistical analysis required.

Contact: Interested students should send an email to Michal Muszynski with their CV and a short cover letter.