My main area of research is in two areas :
*) Generative models. My group is working with variational autoencoders, generative adversarial networks, auto-regressive models, diffusion, and flow models.
Example subtopics are the modeling of CAD models using auto-regressive models with transformer, modeling floorplans using auto-regressive models, editing images and videos using GANs, style transfer using GANs, domain transfer using GANs, and modeling roofs.

*) 3D Reconstruction. We are also working on single image depth estimation using CONV nets / transforer / ... and NERF-style reconstruction. There is also a connection of NERF reconstruction and 3D GANs we are exploring.

Almost all of my group is working on problems using deep learning.