I am a PhD student at the Media Research Laboratory of New York University.

My thesis work is in robot navigation of outdoor environments using computer vision and convolutional neural networks. Previous projects include performing manifold learning on human motion data, simulating biological models of limb control, and texture synthesis. My advisor is Yann LeCun. Most of my collaborators are from his group, the Computational and Biological Learning Lab.

Academics

• Sc.B, Engineering-Physics, Brown University
• MS, Media Arts and Science, MIT
• PhD in progress, Computer Science, New York University

Research

My dissertation is in enabling an autonomous off-road robot to use video input to both build a concise map of its surroundings, and cheaply localize itself within this map as it goes. Using vision for this purpose is essential for complex outdoor environments, which can cause other sensors such as GPS, odometry, and laser scanners to drift severely or become unusably noisy. As a more accurate alternative, we use a convolutional neural network for landmark recognition, and a novel, fast form of visual odometry.

In the past I have worked in learning human motion manifolds, quick prototyping of character animation, and biological models of limb control.

• Manifold learning for separating style and content in dance phrases.
• Prototyping character animations.
• Simulating control primitives for generating limb motion (MIT master's thesis).

More details of the above projects coming soon...

Personal

Please see my personal page for bits about scuba diving, home-made spring shoes, foldable welded angel wings, and other such parts of my life that one need only be exposed to by choice.