Functional Material Design
with the Combination of Artificial Intelligence and Crystal Structure
Prediction
Yindi Jin*
XtalPi
ABSTRACT:
Xtalpi has developed and used efficient global optimization
algorithms, quantum chemical calculations, molecular dynamics other simulation
methods, combined with machine learning algorithms and huge cloud computing
resources, to accurately predict the potential polymorphs of small molecule drugs
and their relative stability with temperature. And Xtalpi design cocrystals and
solvates through the combination of CSP and traditional algorithms.
Furthermore, Xtalpi combines CSP and artificial intelligence algorithms to
explore the prediction of organic molecular crystal properties, and realize the
rapid prediction of crystal solubility, crystal habit, mechanical properties,
sublimation enthalpy, formation enthalpy, electron mobility etc. Provide
powerful methodology support for the screening, design and customization of
related material properties.
Keywords: CSP; crystal properties; AI.
Dr. Yingdi Jin, Chief Research Scientist at XtalPi Inc. She conducted the doctoral and post-doctoral researches at the University of Science and Technology of China and the University of Hong Kong, respectively. Since 2017, XtalPi crystal research team lead by Dr. Jin is developing optimization algorithms for CSP of small-molecule drug candidates. A series of optimization algorithms have been developed by this team. Combined with the schedulable ultra-large-scale cloud computing resources, the algorithms achieved a 100% success rate of customer cases in 2019, including several extremely complex and challenging cases.