AI agents help Penn State scientists find a new pattern in food physics
Data on how foods behave during processing is spread across decades of papers, often tucked into tables, figures and equations. A Penn State team led by food scientist Xu Zhou built AgentXtract, a system in which AI agents with different skills extract, check and organize those numbers into a searchable database, with human experts guiding data quality. Penn State reports that the agents collected more than 4,000 experimental records in about 30 minutes, a job Zhou says once took a trained researcher six months or more. The pooled data revealed a new, consistent relationship between two dielectric properties, which shape microwave heating, across more than 100 foods. The work is online in the Journal of Food Engineering.