Researchers at the U.S. Department of Energy’s (DOE) Argonne National Laboratory have developed a self-driving microscopy technique that uses AI to selectively target points of interest for scanning. This can revolutionize data acquisition, allowing researchers to preserve the integrity of precious samples.
Modern scanning microscopes allow imaging materials with sub-atomic spatial and sub-picosecond time resolutions, but storing and analyzing the large volumes of data generated is difficult. The innovative approach identifies clusters of intriguing features, bypassing humdrum regions of monotonous uniformity, speeding up the experimental process, and enhancing data acquisition.
According to Charudatta Phatak, group leader and materials scientist at Argonne and co-author of the study, “many regions of a sample can be safely disregarded or at least not sampled heavily, but regions, where there are discontinuities and boundaries, can instead contribute the vast majority of information about the sample.”
The AI model first selects a set of random points on the sample. It then simultaneously gathers data from these points while predicting subsequent points of interest, accelerating data acquisition and eliminating the need for human intervention.
“Taking the human component out of the prediction process saves a great deal of time and really speeds up the experiment,” highlights Saugat Kandel, a postdoctoral researcher at Argonne and lead author of the paper.
“The ability to automate experiments with AI will significantly accelerate scientific progress in the coming years,” said Mathew Cherukara, group leader and computational scientist at Argonne and co-author of the study. “This is a demonstration of our ability to do autonomous research with a very complex instrument.”
Importantly, “the AI can be trained on a generic image, and it knows immediately how to recognize the areas of interest,” explains Zichao Di, a computational mathematician at Argonne and another study co-author.
Tao Zhou, nanoscientist at Argonne and co-author, emphasizes that the new approach can be broadly employed across various microscopies, including X-ray, electron, and atomic probe microscopy. It offers unparalleled speed and opens up new avenues for scientific exploration. By combining AI and microscopy, the researchers expect to enhance the efficacy of scanning microscopy experiments and, thus, the study of dynamic physical phenomena.
For more information: Nature Communications
Image: Artist’s representation of the autonomous scanning microscopy experiment at the APS. This experimental setup allows the AI-driven FAST system to autonomously control the beam position and the acquisition of data from the detector. (Image Credit: Argonne National Laboratory/Saugat Kandel).
