Optimality guarantees for crystal structure prediction

Computational approach inspired advancements of solid-state electrolytes for lithium secondary batteries: from first-principles to machine learning

“…ML algorithms can be trained on known compositions and properties to predict the performance of new compositions for inverse design, providing valuable insights into material behavior and aiding in the design and optimization process. 330 Hamad et al used a random forest-based model to predict the ionic conductivity of SSEs for LIBs and sodium-ion batteries, with experimental validation (Fig. 9(a)).…”

Crystal growth, first-principles calculations, and enhanced 1.5–1.7 μm fluorescence emission from the Er, Ce: CaGdAlO4 laser crystal

“…Based on the obtained stable structure, the local geometrical information, and electronic and optical properties of rare earth ions in the matrix materials can be obtained by combining the luminescence principle with the first-principle calculations. 37–50…”

Integration of generative machine learning with the heuristic crystal structure prediction code FUSE

“…Within the constraints of how the problem is formulated, this then provides a guarantee that the global minimum structure has been located. [12] The types of heuristic methods for crystal structure prediction mentioned above are all dependent on an algorithm to generate initial structure(s) with an element of randomness that then evolve in some specified way. The initial structures can be produced using simple rules, for example randomly selecting a unit cell then populating it with atoms with minimum inter-atomic distance constraints, or with more complex rule sets based upon knowledge of inorganic chemistry, as used, e.g., in FUSE having structures assembled from randomly generated blocks using rules based on how such blocks connect in known compounds.…”