High-resolution imaging of materials produces complex, copious data. Researchers at Oak Ridge National Laboratory are developing a visual analytics system that could essentially "look over a scientist's shoulder," learning from human actions and improving its predictions of what the scientist thinks is interesting.
The approach combines computational data analytics with interactive data visualization--human intuition and knowledge. "Humans are very good at visual processing and intuition tasks that are hard to do with a computer," ORNL's Chad Steed said.
"If we could grasp that strength of the humans in some type of intelligent interface, we could go after problems that neither approach in isolation could solve."