Home > Events > 2013 Seminars & Colloquia > JOSEPH RICHARDS - University of California at Berkeley

JOSEPH RICHARDS - University of California at Berkeley

Main Content

”Astronomical Discovery and Classification for the Synoptic Survey Era”
12 February 2013 from 4:00 PM to 5:00 PM
201 Thomas Bldg.
Add event to calendar

We have entered the Synoptic Survey Era of observational astronomy, where data collection rates will soon reach several terabytes per night.  A growing army of telescopes monitor, nightly, the luminosities of millions of astronomical objects such as stars, galaxies and quasars, searching both for distant stellar explosions and more nearby variable stars that exhibit time-varying brightness due to pulsation, chromospheric activity, or eclipse.  Real-time analysis of these data is critical to determine which objects and events require timely observations with expensive follow-up resources.  To maximize the scientific returns from these immense projects, sophisticated statistical tools must be used.  I will describe my ongoing research in several areas, including real-time discovery and classification of transient events for the Palomar Transient Factory, multi-band photometric supernova typing, and probabilistic classification of variable stars from irregularly sampled time series.  I will outline some of the methodology that I have developed for these problems, including the use of manifold learning for time-series feature extraction, active learning to overcome sample-selection biases, and semi-supervised learning to detect anomalies.  I will conclude by describing my vision for the future of automated discovery and classification in time-domain astronomy.

Filed under: ,