ORNL’s AI Recommender System to Enhance Experimentation Performance

Oak Ridge National Laboratory has developed a human-artificial intelligence collaboration recommender system to improve how machine learning algorithms use data during experiments.

The system uses a human-in-the-loop approach in identifying significant data sets for research and development. With the technology, researchers will be provided with data set suggestions based on their activities and the capability to order AI to show similar content or different data sets.

Tests showed that guidance from researchers enabled the system’s machine learning algorithms to find relevant data with minimal human input, ORNL said.

National Cheng Kung University in Taiwan and the University of Tennessee participated in the system development effort. The Center for Nanophase Materials Sciences and the MLExchange project, both funded by the Department of Energy, supported ORNL experiments and autonomous workflows and algorithm development, respectively.

Arpan Biswas, a postdoctoral research associate at ORNL, said the system aims to improve the quality of data for enhanced experimentation performance.