Predictions of new AB O 3 perovskite compounds by combining machine learning and density functional theory
Not Available
Balachandran, Prasanna V.↗
Engineering topics
Gubernatis, James E.
Publications and source records attributed to Gubernatis, James E..
Multi-fidelity machine learning models for accurate bandgap predictions of solids
Here, we present a multi-fidelity co-kriging statistical learning framework that combines variable-fidelity quantum mechanical calculations of bandgaps to generate a machine-learned model that enables low-cost accurate predictions of the bandgaps at the highest fidelity level. Additionally, the adopted Gaussian process regression formulation allows us to predict the underlying uncertainties as a measure of our confidence in the predictions. In using a set of 600 elpasolite compounds as an example dataset and using semi-local and hybrid exchange correlation functionals within density functional theory as two levels of fidelities, we demonstrate the excellent learning performance of the method against actual high fidelity quantum mechanical calculations of the bandgaps. The presented statistical learning method is not restricted to bandgaps or electronic structure methods and extends the utility of high throughput property predictions in a significant way.
36 MATERIALS SCIENCE↗
Multi-objective optimization techniques to design the Pareto front of organic dielectric polymers
Here, we present two Monte Carlo algorithms to find the Pareto front of the chemical space of a class of dielectric polymers that is most interesting with respect to optimizing both the bandgap and dielectric constant. Starting with a dataset generated from density functional theory calculations, we used machine learning to construct surrogate models for the bandgaps and dielectric constants of all physically meaningful 4-block polymers (that is, polymer systems with a 4-block repeat unit). We parameterized these machine learning models in such a way that the surrogates built for the 4-block polymers were readily extendable to polymers beyond a 4-block repeat unit. By using translational invariance, chemical intuition, and domain knowledge, we were able to enumerate all possible 4, 6, and 8 block polymers and benchmark our Monte Carlo sampling of the chemical space against the exact enumeration of the surrogate predictions. We obtained exact agreement for the fronts of 4-block polymers and at least a 90% agreement for those of 6 and 8-block polymers. We present fronts for 10-block polymer that are not possible to obtain by direct enumeration. We note that our Monte Carlo methods also return polymers close to the predicted front and a measure of the closeness. Both quantities are useful information for the design and discovery of new polymers.
36 MATERIALS SCIENCE↗
Triplet p + ip pairing correlations in the doped Kane-Mele-Hubbard model: A quantum Monte Carlo study
By using the constrained-phase quantum Monte Carlo method, we performed a systematic study of the pairing correlations in the ground state of the doped Kane-Mele-Hubbard model on a honeycomb lattice. We find that pairing correlations with d + id symmetry dominate close to half filling, but pairing correlations with p+ip symmetry dominate as hole doping moves the system below three-quarters filling. We correlate these behaviors of the pairing correlations with the topology of the Fermi surfaces of the non-interacting problem. We also find that the effective pairing correlation is enhanced greatly as the interaction increases, and these superconducting correlations are robust against varying the spin-orbit coupling strength. Finally, our numerical results suggest a possible way to realize spin triplet superconductivity in doped honeycomb-like materials or ultracold atoms in optical traps.
71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
Structure classification and melting temperature prediction in octet AB solids via machine learning
Machine learning methods are being increasingly used in condensed matter physics and materials science to classify crystals structures and predict material properties. However, the reliability of these methods for a given problem, especially when large data sets are unavailable, has not been well studied. By addressing the tasks of classifying crystal structure and predicting melting temperatures of the octet subset of AB solids, we performed such a study and found potential problems with using machine learning methods on relatively small data sets. At the same time, however, we can reaffirm the potential power of such methods for these tasks. In particular, we uncovered an important new material feature, the excess Born effective charge, that significantly increased the accuracy of the predictions for the classification problem we defined. This discovery leads us to propose a new scale for the degree of ionicity and covalency in these solids. More specifically, we partitioned the crystal structures of a set of 75 octet solids into those that are ionic and covalent bonded and thus performed a binary classification task. We found that using the standard indices (r σ ,r π ), suggested by St. John and Bloch several decades ago, enabled an average success in classification of 92%. Using just r σ and the excess Born effective charge ΔZ A of the A atom enabled an average success of 97%, but we also found relatively large variations about these averages that were dependent on how certain machine learning methods were used and for which a standard deviation was not a proper measure of the degree of confidence we can place in either average. Instead, we calculated and report with 95 % confidence that the traditional classification pair predicts an accuracy in the interval [ 89%, 95%] and the accuracy of the new pair lies in the interval [96 %, 99%]. For melting temperature predictions, the size of our data set was 46. We estimate the root-mean-squared error of our resulting model to be 11% of the mean melting temperature of the data, but we note that if the accuracy of this predicted error is itself measured, our estimated fitting error itself has a root-mean-square error of 50%. In short, what we illustrate is that classification and regression predictions can vary significantly, depending on the details of how machine learning methods are applied to small data sets. This variation makes it important, if not essential, to average the predictions and compute confidence intervals about these averages to report results meaningfully. However, when properly used, these statistical methods can advance our understanding and improve predictions of material properties even for small data sets.
36 MATERIALS SCIENCE↗
Materials Classification & Accelerated Property Predictions using Machine Learning [Slides]
Examine reliability of ML methods for “small” data sets and potenial for new insights – role of excess effective born charge as a feature.
36 MATERIALS SCIENCE↗