Cameron Smith

I've been working with Vincent Sitzmann's Scene Representation Group on self-supervised 3D representations in between undergrad and PhD, and next I'll start my PhD with Yue Wang's group at USC!
Generally, I am interested in building AI with the right inductive biases to enable sample-efficient, self-supervised, and generalizable learning on tasks which involve self-simulation and planning. Currently, my emphasis is building self-supervised 3D representations as scaffolding to support more principled inductive biases.

[Email] [Google Scholar] [Github]

Publications