Cornell Lab of Ornithology
Information Science
Home What We Do research Staff Partners Sponsors

Avian Knowledge Network / ITR

NSF- ITR- (ASE+EVS)- (dmc+sim): Tracking Environmental Change through the Data Resources of the Bird-Monitoring Community


The goal of this project is to advance research in the computer sciences by solving a fundamental problem in population biology. By exploring new models of machine learning and new methodologies for the mining of large data sets we will estimate the true abundance of wild bird populations across North America. We are taking an interdisciplinary approach to bring computational specialists together with population biologists and statisticians to develop broadly applicable technologies and expand knowledge in both the computer and biological sciences. For example, by addressing issues of variation in monitoring data, such as observer or protocol biases in the detection of birds, we will advance new techniques in ensemble learning, statistical smoothing, and multi-task machine learning. By taking a novel approach to estimating change within time series data to quantify variation in spatio-temporal abundance, we will produce accurate maps showing the distribution and variation in bird population abundances, and relate this to environmental conditions. Our goal is to allow rapid browsing of bird-monitoring data over the Internet, and to accomplish this we will advance new strategies for fast approximate data-exploration queries.

Avian Knowledge Network Report

Avian Knowledge
Network Report

(236 KB)