abstract
The combination of machine learning (ML) models with chemistry-related tasks requires the description of molecular structures in a machine-readable way. The nature of these so-called molecular descriptors has a direct and major impact on the performance of ML models and remains an open problem in the field. Structural descriptors like SMILES strings or molecular graphs lack size-independence and can be memory intensive. Machine-learned descriptors can be of low dimensionality and constant size but lack physical significance and human interpretability. Sigma profiles, which are unnormalized histograms of the surface charge distributions of solvated molecules, combine physical significance with low dimensionality and size-independence, making them a suitable candidate for a universal molecular descriptor. However, their widespread adoption in ML applications requires open access to sigma profile generation, which is currently not available. This work details the development of OpenSPGen - an open-source tool for generating sigma profiles. Also presented are studies on the effect of different settings on the efficacy of the generated sigma profiles at predicting thermophysical material properties when used as inputs to a Gaussian process as a simple surrogate ML model. We find that a higher level of theory does not translate to more accurate results. We also provide further recommendations for sigma profile calculation and use in ML models.
keywords
MOLECULAR-ORBITAL METHODS; DEEP EUTECTIC SOLVENTS; GAUSSIAN-2 G2 THEORY; COSMO-RS; BASIS-SETS; ATOMS K; PREDICTION; SOLUBILITY; EXTENSION; ENERGY
subject category
Chemistry; Computer Science
authors
Salih, FYM; Abranches, DO; Maginn, EJ; Colón, YJ
our authors
Dinis O. Abranches
Assistant ResearcherProjects
CICECO - Aveiro Institute of Materials (UIDB/50011/2020)
CICECO - Aveiro Institute of Materials (UIDP/50011/2020)
Associated Laboratory CICECO-Aveiro Institute of Materials (LA/P/0006/2020)
acknowledgements
This work was supported by the Breakthrough Electrolytes for Energy Storage (BEES) Energy Frontier Research Center (award #DE-SC0019409), funded by the U.S. Department of Energy, Office of Science, Basic Energy Sciences program, and by the U.S. Department of Energy through subcontract 630340 from Los Alamos National Laboratory, Materials and Chemical Sciences Division. This research is also based upon work partially supported by the National Science Foundation under award number ERC-2330175 for the Engineering Research Center EARTH. Additionally, partial funding was provided by the CICECO Aveiro Institute of Materials (UIDB/50011/2020, UIDP/50011/2020, and LA/P/0006/2020), financed by Portuguese funds through FCT/MCTES (PIDDAC).