I'm interested in the motion of the visual world. Motion is critical for intelligent systems. Motion informs of changes in the world, its structure, physical properties, etc.  As Gibson nicely put it "We see in order to move and we move in order to see". In our lab we study what motion tells us about the world and how to estimate it from video. These concepts often fall into problem buckets like: Optical Flow, Depth and Structure from Motion, Instance Segmentation in Video, Object Tracking, Temporal Representations, Video Representations, Action Recognition, Vision and Simulated Environments, Active Vision. Solving these problems requires a fundamental ability to learn. In video, labels are expensive, and thus learning challenges broadly fall in the following Machine Learning areas: Deep Learning, in particular CNNs, Unsupervised or Weakly Supervised Learning, Reinforcement Learning.