DiSCoVeR: a materials discovery screening tool for high performance, unique chemical compositions

Abstract

We present Descending from Stochastic Clustering Variance Regression (DiSCoVeR) (https://github.com/sparks-baird/mat discover), a Python tool for identifying and assessing high-performing, chemically unique compositions relative to existing compounds using a combination of...

Quantifying the performance of machine learning models in materials discovery

“…In recent studies, the number of experiments necessary to identify a high-performing material has been used as a metric for monitoring SL performance. 8,16,33 Modeling benchmark datasets and tools, such as Olympus, 34 MatBench, 35 and DiSCoVeR, 36 have started to standardize assessment of model and dataset performance. Notably, a recent study by Rohr et al 37 considers additional metrics that quantify SL performance relative to a benchmark case (typically random search).…”

Serendipity based recommender system for perovskites material discovery: balancing exploration and exploitation across multiple models

“…Recommending a diverse set of reactions increases the likelihood of identifying both phases, rather than being trapped in a local minimum of one or the other. While we have focused on the problem of optimizing reaction composition, the serendipity approach could be applied to compositional optimization problems by incorporating a suitable composition-space distance metric (45) during in the calculation.…”

Not as simple as we thought: a rigorous examination of data aggregation in materials informatics

“…DiSCoVeR 31 algorithm is a recently proposed ensemble of machine learning methods aimed at facilitating the identification of chemistries lying at the intersection between novelty and performance. In practice, DiSCoVeR can be used to provide novelty scores of a given pool of data with respect to another and it was recently employed to identify new chemically novel high-temperature superconductors.…”