ML-driven discovery and classification of transients and variable stars in the era of synoptic surveys (PTF to LSST): real/bogus discrimination, probabilistic variable-star catalogs, and automating the discovery loop.
With Joseph Richards, on the first day of the five-day Berkeley workshop on machine learning for real-time and streaming data (May 7–11, 2012, Banatao Auditorium, Sutardja Dai Hall). The 51-slide deck is on SlideShare under the title “Machine Learning and Classification in the Synoptic Survey Era”; its title slide reads “Berkeley Streaming Workshop; 7 May 2012”.