Artur Dubrawski, director of the Auton Lab, Carnegie Mellon University, said that the lab researches new approaches to statistical data mining, specifically the underlying computer science, mathematics, statistics, and artificial intelligence of detection and exploitation of patterns in data.
Government agencies routinely collect different kinds of data reflecting various issues regarding food safety, Dubrawski said, and researchers at Auton Lab have been working with them for years. Among other things, they have looked at a database of food consumer complaints maintained by the U.S. Dept. of Agriculture’s Food Safety and Inspection Service (FSIS), developed analytic components, read data, and evaluated the hypothesis that complaints may be probabilistically related to complaints by other consumers and may have the same underlying cause. The work resulted in a consumer complaint monitoring system. The researchers have also looked at other databases used by government agencies. A few years ago, the Centers for Disease Control and Prevention (CDC) joined the effort, isolating bacteria and viruses and tracing them back to food. Now, he said, both agencies can identify pathways of transmission from foods to humans quickly and perhaps develop better processes and policies to mitigate future events. He said that comprehensively analyzing the data can bene t society because agencies can be agile and informed and act more swiftly and, in some cases, more efficiently.