27. Variational Autoencoders
Generative machine learning models have the potential to allow us to move beyond screening to true materials discovery. Generative adversarial networks (GANs) are one powerful tool and variational autoencoders (VAEs) are another. This video descrbies autoencoders, latent space, reparameterization trick,
Check out the whole materials informatics series at • Materials Informatics with workbooks and course notes available at https://github.com/sp8rks/MaterialsIn...
0:00 what is an autoencoder and how does it work?
2:30 what is an autoencoder used for? neural image compression, denoising, neural inpainting
6:35 variational autoencoder architecture
8:18 reparameterization trick
11:30 disentangled VAEs
14:13 comparing GANs and VAEs