Scientific and Information visualization
While making museum exhibits in Randall Landsberg's section of Leo Kadanoff's SCOPE project, Mark SubbaRao (UC Astro/Adler Planetarium) introduced me to Stuart Levy's Partiview, free 3d visualization software that has traditionally been used in the astronomy domain. This turned out to be a surprisingly useful tool for testing machine learning algorithms.
"Visualizing high-dimensional datasets with Partiview", poster with Stuart Levy at the 2004 IEEE Information Visualization conference. We also have a demo at the 2004 NIPS Conference.
Machine Learning Projects
Matlab interface to Partiview. Comments welcome.
Here are some demos showing how different spectral clustering and kernel-based algorithms perform on different types of data in my research and other people's research.
I've also processed (aligned, documented, etc) several stereo photographs from several terrestial observatories (Kitt Peak, Apache Point, Mt Hopkins, Owens Valley) and Mars. The pictures were taken by Brian Humensky, Jurek Krzesinski, Chris Greer, and the NASA/JPL Spirit and Opportunity Rovers.
Here are some cute demos I've made for friends.
The original Partiview paper is Stuart Levy's "Interactive 3-D Visualization of Particle Systems with Partiview", in the 2001 Proceedings (Vol 208) of the International Astronomical Union Symposium on Astrophysical Supercomputing using Particles.
The two standard Partiview references are
Partiview Users Guide by Brian Abbott of Hayden Planetarium / AMNH.
Partiview Reference Manual by Peter Teuben and Stuart Levy.
And here's some stuff I've written: