CSAIL Event Calendar: Previous Series

Active Learning, Distilled Sensing, or how to close the loop between data analysis and acquisition

Speaker: Rui Castro, Assistant Professor , Columbia University, Dept of Electrical Engineering
Date: April 12 2010
Time: 4:00PM to 5:00PM
Location: Sem Rm G449 (Patil/ Kiva)
Host: Prof. Tomaso A. Poggio, McGovern Inst., BCS Dept. & CSAIL

Abstract: Many traditional approaches to statistical inference and machine learning are passive, in the sense that all data are collected prior to analysis in a non-adaptive fashion. However, in many practical scenarios it is possible to adjust the data collection process based on information gleaned from previous observations, in the spirit of the "twenty-questions" game. Learning in such settings is known as active learning or inference using sequential experimental designs. Despite the potential to dramatically improve inference performance, analysis of such procedures is difficult, due to the complicated data dependencies created by the closed-loop observation process. These difficulties are further exasperated by the presence of measurement uncertainty or noise.
In this talk I present a quantitative analysis of active learning in a variety of scenarios, including non-parametric settings. I also present a novel selective sensing procedure - Distilled Sensing - which is highly effective for detection and estimation of high-dimensional sparse signals in noise. Large-sample analysis shows that the proposed procedure provably outperforms the best possible detection methods based on non-adaptive sensing, allowing for the detection and estimation of extremely weak signals, imperceptible without adaptive sensing.

See other events that are part of Brains and Machines Seminar Series 2009/2010

See other events happening in April 2010