Effective Molecular Descriptors for Chemical Accuracy at DFT Cost: Fragmentation, Error-Cancellation, and Machine Learning

Collins, Eric M.; Raghavachari, Krishnan

doi:10.1021/acs.jctc.0c00236

Cited by 23 publications

(26 citation statements)

References 65 publications

(132 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The former relies heavily on chemical intuition by requiring the user to choose which atomic or molecular attributes are important for the problem at hand. Recent developments from our group proposed a class of fragmentation-based representations termed ML(CBH) or simply MLCBH, in which a system is broken apart into smaller fragments based on the generalized isodesmic schemes of the Connectivity-Based Hierarchy. , CBH reactions are characterized by deconstructing the molecule into smaller n diameter fragments, corresponding to the n th rung on CBH, as well as their overlaps, to satisfy the inclusion–exclusion principle. Once the full reaction scheme is constructed, the coefficients of the fragments along with their overlaps are multi-hot encoded into a vector of all possible fragments (Figure b).…”

Section: Methodsmentioning

confidence: 99%

“…Nevertheless, many of the most successful representations thus far have been designed to replace QM through standard ML techniques by describing the composition and the 3D structure of a chemical system as either an encoded vector through popular fingerprinting algorithms, such as Morgan FP or ECFP, or as a higher dimensional tensor through machine learning interatomic potentials. , Such models are typically benchmarked against large datasets of DFT-calculated properties. Although these ML models can achieve mean absolute errors (MAEs) below the threshold of “chemical accuracy” (∼1 kcal/mol), the reference values being reproduced (typically DFT) are still significantly inaccurate compared to experiments or more sophisticated CCSD(T)-based cWFTs such as G4 or G4(MP2). − …”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Fragmentation-Based Graph Embedding Framework for QM/ML

Collins

Raghavachari

2021

J. Phys. Chem. A

Self Cite

View full text Add to dashboard Cite

We introduce a new fragmentation-based molecular representation framework "FragGraph" for QM/ML methods involving embedding fragment-wise fingerprints onto molecular graphs. Our model is specifically designed for delta machine learning (Δ-ML) with the central goal of correcting the deficiencies of approximate methods such as DFT to achieve high accuracy. Our framework is based on a judicious combination of ideas from fragmentation, error cancellation, and a state-of-the-art deep learning architecture. Broadly, we develop a general graph-network framework for molecular machine learning by incorporating the inherent advantages prebuilt into error cancellation methods such as the generalized Connectivity-Based Hierarchy. More specifically, we develop a QM/ML representation through a fragmentationbased attributed graph representation encoded with fragment-wise molecular fingerprints. The utility of our representation is demonstrated through a graph network fingerprint encoder in which a global fingerprint is generated through message passing of local neighborhoods of fragment-wise fingerprints, effectively augmenting standard fingerprints to also include the inbuilt molecular graph structure. On the 130k-GDB9 dataset, our method predicts an out-of-sample mean absolute error significantly lower than 1 kJ/mol compared to target G4(MP2) calculated energies, rivaling current deep learning methods with reduced computational scaling.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Fragmentation-Based Graph Embedding Framework for QM/ML

Collins

Raghavachari

2021

J. Phys. Chem. A

Self Cite

View full text Add to dashboard Cite

show abstract

“…The rise of powerful machine learning (ML) methods and hardware has provided tools for chemists to approach the problem of accurate enthalpy calculations from a completely different, data-driven, angle exploiting the fact that ML calculations practically come at no cost compared to high-level QM methods. − A series of special-purpose ML approaches were suggested either to specifically target prediction of enthalpies of formation directly ,− or by correcting predictions of a baseline lower-level, DFT, − Hartree–Fock, or SQM method to experimental data ,− ,− or higher-level QM methods such as G4 and G4MP2 . Impressively, many of these special-purpose ML approaches were reported to come close to reaching chemical accuracy, but all of them are based on correcting DFT-level predictions, − which strongly limits their computational efficiency.…”

Section: Uncertainty Quantificationmentioning

confidence: 99%

Toward Chemical Accuracy in Predicting Enthalpies of Formation with General-Purpose Data-Driven Methods

Zheng

Yang

et al. 2022

J. Phys. Chem. Lett.

View full text Add to dashboard Cite

Enthalpies of formation and reaction are important thermodynamic properties that have a crucial impact on the outcome of chemical transformations. Here we implement the calculation of enthalpies of formation with a general-purpose ANI-1ccx neural network atomistic potential. We demonstrate on a wide range of benchmark sets that both ANI-1ccx and our other general-purpose data-driven method AIQM1 approach the coveted chemical accuracy of 1 kcal/mol with the speed of semiempirical quantum mechanical methods (AIQM1) or faster (ANI-1ccx). It is remarkably achieved without specifically training the machine learning parts of ANI-1ccx or AIQM1 on formation enthalpies. Importantly, we show that these data-driven methods provide statistical means for uncertainty quantification of their predictions, which we use to detect and eliminate outliers and revise reference experimental data. Uncertainty quantification may also help in the systematic improvement of such datadriven methods.

show abstract

“…Indeed, the development of new descriptors and techniques to represent chemical information for specialized use cases is subject of ongoing research [497], [498], [499], [500], [501], [502], [503], [504], [505], [506]. For example, Ref.…”

Section: E Technical Aspectsmentioning

confidence: 99%

A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry

Fiedler¹,

Shah²,

Bußmann³

et al. 2021

Preprint

View full text Add to dashboard Cite

With the growth of computational resources, the scope of electronic structure simulations has increased greatly. Artificial intelligence and robust data analysis hold the promise to accelerate large-scale simulations and their analysis to hitherto unattainable scales. Machine learning is a rapidly growing field for the processing of such complex datasets. It has recently gained traction in the domain of electronic structure simulations, where density functional theory takes the prominent role of the most widely used electronic structure method. Thus, DFT calculations represent one of the largest loads on academic high-performance computing systems across the world. Accelerating these with machine learning can reduce the resources required and enables simulations of larger systems. Hence, the combination of density functional theory and machine learning has the potential to rapidly advance electronic structure applications such as in-silico materials discovery and the search for new chemical reaction pathways. We provide the theoretical background of both density functional theory and machine learning on a generally accessible level. This serves as the basis of our comprehensive review including research articles up to December 2020 in chemistry and materials science that employ machine-learning techniques. In our analysis, we categorize the body of research into main threads and extract impactful results. We conclude our review with an outlook on exciting research directions in terms of a citation analysis.

show abstract

Effective Molecular Descriptors for Chemical Accuracy at DFT Cost: Fragmentation, Error-Cancellation, and Machine Learning

Cited by 23 publications

References 65 publications

A Fragmentation-Based Graph Embedding Framework for QM/ML

A Fragmentation-Based Graph Embedding Framework for QM/ML

Toward Chemical Accuracy in Predicting Enthalpies of Formation with General-Purpose Data-Driven Methods

A Deep Dive into Machine Learning Density Functional Theory for Materials Science and Chemistry

Contact Info

Product

Resources

About