I am an Assistant Professor of Computer Science at Brown University. In my research, I’m broadly interested in the intersection of computer graphics with artificial intelligence and machine learning. I primarily focus on learning, inference, and synthesis with generative graphics models and programs. Questions I care about in this domain include:

  • How can we learn structured generative models of visual content from (small or large) data?
  • How can we discover re-usable abstractions in visual content and exploit them in generative models?
  • How can we make it easier to create procedural representations of visual content?
  • How can we infer procedural representations from unstructured, noisy, perceptual input, such as images?
  • How can generative models efficiently incorporate functional constraints, such as physical stability?

I received my PhD from Stanford University, where I worked with Pat Hanrahan in the Graphics Lab and with Noah Goodman in the Computation and Cognition Lab. I received my undergraduate degree from the University of California, Berkeley.

19 Feb 2018
Angela Dai's work on large-scale 3D scan completion accepted to CVPR 2018.
15 Dec 2017
New paper on learning procedural models from examples accepted to Eurographics 2018.
17 Nov 2017
Maxime Voisin's work on improved training for neural autoregressive data completion models accepted to the NIPS 2017 Bayesian Deep Learning Workshop.
12 Oct 2017