A new AI algorithm is designed, created, and successfully tested by a team of
scientist the U.S. Department of Energy's Brookhaven National Laboratory and
Lawrence Berkeley National Laboratory to make smarter scientific measurement
decisions. A form of artificial intelligence (AI) algorithm can make autonomous
decisions to define and perform the next step of an experiment. The scientific
experiments performed to understand the world around us from Galileo and Newton
to the recent discovery of gravitational waves, has been the driving force of
our technological advancement for hundreds of years. The researchers can have
tremendous impact on how quickly those experiments yield applicable results for
new technologies improving the way they do their experiments.
Researchers have sped up their experiments
over the last decades through automation and an ever-growing assortment of fast
measurement tools. However, a very demanding and time-consuming experiments still
requires for some the most interesting and important scientific challenges --
such as creating improved battery materials for energy storage or new quantum
materials for new types of computers. The new decision-making algorithm offers
the possibility to study these challenges in a more efficient fashion team from
two of Brookhaven's DOE Office of Science user facilities -- the Center for
Functional Nanomaterials (CFN) and the National Synchrotron Light Source II
(NSLS-II) -- and Berkeley Lab's Center for Advanced Mathematics for Energy
Research Applications (CAMERA).
Experiments are done to gain knowledge
about the material that is studied by scientists in a well-tested way using take
a sample of the material and measuring how it reacts to change in its
environment. A standard approach for scientists is to manually scan through the
measurements from a given experiment at user facilities like NSLS-II and CFN
but access to these high-end materials-characterization tools is limited. A
research team need to make the most out of each measurement to measure their
materials, as they might have a few days. A minimum number of measurements and
maximum quality of the resulting model can be achieved where uncertainties are
large. The goal is not only to take data faster but also to improve the quality
of the collected data. The new approach is an example of applied artificial
intelligence. The decision-making algorithm can scan through the data and make
smart decisions about how the experiment should proceed replacing the intuition
of the human experimenter.
The team had to tackle three important
pieces to make autonomous experiments a reality - the automation of the data
collection, real-time analysis, and the decision-making algorithm. The
decision-making algorithm was first implemented at the Complex Materials Scattering
(CMS) beamline at NSLS-II. The team first fully autonomous performed was to map
the perimeter of a droplet where nano particles segregate using a technique
called small-angle x-ray scattering at the CMS beamline. The scientists
compared the standard approach of measuring the sample with measurements taken
when the new decision-making algorithm was calling the shots in the first
experiment, resulting The algorithm was able to identify the area of the
droplet and focused on its inner parts and edges instead of the background.
0 Comments