The emergence of artificial intelligence is influencing all fields of science, from biology to materials science.
A new technique for X-ray crystallography was recently introduced, allowing researchers to make adjustments to their experiments in real time.
AI to speed up processes during X-ray work
When it comes to some types of X-ray experiments, new AI approaches have enabled researchers to get a more precise analysis of their samples, and to do so in a much shorter period of time. A group of researchers at the US Department of Energy’s (DOE) Argonne National Laboratory are harnessing AI to perform the challenging task of analyzing data from high-energy X-ray experiments. Using a new neural network-based method called BraggNN, the Argonne team can more accurately identify Bragg peaks — data points that indicate positions and orientations of small individual crystals — in a fraction of the time they used to.
“Conventional methods work like trying on a suit at a tailor. You have to test an estimated shape first, and then do the shape adjustment after it. With our technique, using a neural network, it is like generating the perfect suit instantly from a photograph of a person»said Hemant Sharma of the Argonne National Laboratory
Neural networks, denoted c0k0 NN in the BraggNN name, are algorithms that look for patterns in data and, over time, learn to predict outcomes, speeding up the analysis of that data.
“BraggNN gives us much greater efficiency and speed than conventional techniques”added Antonino Miceli of Argonne, the paper’s author and group leader at Advanced Photon Source (APS), a division of the DOE Office of Science in Argonne.
In recent years, a technique called high-energy diffraction microscopy (HEDM) has become one of the most popular ways scientists use to accurately characterize complicated materials with high resolution. Although HEDM has proven to be a great improvement over conventional techniques, it can also be costly and time consuming. It involves collecting huge data sets, analyzing millions of Bragg diffraction peaks, and reconstructing the sample using those peaks.
To address the technical challenges of this study, researchers working on the APS turned to AI to speed up and streamline Bragg peak analysis. The conventional method involves using a 2D or 3D model and fitting peak data to it, but the research team’s new model can directly determine peak positions from the data.
“Conventional methods work like trying on a suit at a tailor”said Argonne computer scientist Hemant Sharma, an author of the study. “You have to test an estimated shape first, and then do the shape adjustment after it. With our technique, using a neural network, it is like generating the perfect suit instantly from a photograph of a person».
After the model was trained on data containing diffraction spikes, the researchers were able to dramatically speed up the analysis and improve accuracy. “The real achievement is that we made peak determinations much faster and also delivered sub-pixel accuracy, the gold standard for drawing useful conclusions”said Argonne computer scientist Zhengchun Liu, the paper’s first author.
The advanced computational methods used by BraggNN are especially conducive to use on a graphics processing unit (GPU) chip, helping to further speed up its performance.
An article reporting the main findings of this study was published in the Journal of the International Union of Crystallography.