Living Lab Lecture No. 26: Variational Autoencoder - How Asterix pushes data from high- to low dimensional representation

In the era of high-dimensional data in biomedical research and other application areas, representation learning and dimension reduction by deep learning networks have become a popular choice and vivid field of research. Variational Autoencoders (VAE) are probabilistic deep learning networks and their derivatives are promising solutions for explainable representation learning or cross-modal data integration. In our lecture we will touch the basics, show the application to molecular data in biomedical research and our latest framework development for the application of VAE to any data and research.