Machine learning prediction errors better than DFT accuracy
By Felix A. Faber, Luke Hutchison, Bing Huang, Justin Gilmer, Sam Schoenholz, George E. Dahl, Oriol Vinyals, Steven M. Kearnes, Patrick F. Riley, O. V. Lilienfeld. 2017
36 citations recorded by the source corpus.
We investigate the impact of choosing regressors and molecular representations for the construction of fast machine learning (ML) models of thirteen electronic ground-state properties of organic molecules. The performance of each regressor/representation/property combination is assessed using learning curves which report out-of-sample errors as a function of training set size with up to $\sim$117k distinct molecules. Molecular structures and properties at hybrid density functional theory (DFT) level of theory used for training and testing come from the QM9 database [Ramakrishnan et al, {\em Scientific Data} {\bf 1} 140022 (2014)] and include dipole moment, polarizability, HOMO/LUMO energies and gap, electronic spatial extent, zero point vibrational energy, enthalpies and free energies of atomization, heat capacity and the highest fundamental vibrational frequency. Various representations from the literature have been studied (Coulomb matrix, bag of bonds, BAML and ECFP4, molecular graphs (MG)), as well as newly developed distribution based variants including histograms of distances (HD), and angles (HDA/MARAD), and dihedrals (HDAD). Regressors include linear models (Bayesian ridge regression (BR) and linear regression with elastic net regularization (EN)), random forest (RF), kernel ridge regression (KRR) and two types of neural net works, graph convolutions (GC) and gated graph networks (GG). We present numerical evidence that ML model predictions deviate from DFT less than DFT deviates from experiment for all properties. Furthermore, our out-of-sample prediction errors with respect to hybrid DFT reference are on par with, or close to, chemical accuracy. Our findings suggest that ML models could be more accurate than hybrid DFT if explicitly electron correlated quantum (or experimental) data was available.
Subfields: AI for Physical Sciences
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In AI for Physical Sciences
Citations in the corpus
- Directional Message Passing for Molecular Graphs (2020, background): “3 kcal mol − 1 for energy (Faber et al., 2017)), so any remaining difference to real-world data is almost exclusively due to errors in the DFT simulation.”
- Message passing neural networks for molecular property prediction (2019, background): “Other approaches have explored the use of 3D atomic coordinates to augment the information provided to these models [46, 39, 25, 17, 18].”
- Nanophotonic particle simulation and inverse design using artificial neural networks (2017, methodology): “al.(9) used NNs to approximate Density Functional Theory.”