Multilayer atomic cluster expansion for semilocal interactions

Bochkarev, Anton S.; Lysogorskiy, Yury; Ortner, Christoph; Csänyi, Gábor; Drautz, Ralf

doi:10.1103/physrevresearch.4.l042019

Cited by 16 publications

(3 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, for bigger gains, we believe that more significant changes to the proposed architecture of the models are needed. For example, a promising pathway is to extend our BIPs model to the recently proposed framework for equivariant higher-order message passing [37,38] that has proven highly successful for modeling inter-atomic interactions [39,40], typically outperforming other approaches despite employing much shallower architectures. This work suggests that the BIP model could also be extended to a geometric deep-learning framework which would naturally lead to the automated discovery of the embedding Q n , and generally open up further model tuning possibilities that will likely significantly improve the already excellent accuracy we obtain with BIP models.…”

Section: Discussionmentioning

confidence: 99%

Boost invariant polynomials for efficient jet tagging

Munoz

Batatia

Ortner

2022

Mach. Learn.: Sci. Technol.

Self Cite

View full text Add to dashboard Cite

Given the vast amounts of data generated by modern particle detectors, computational efficiency is essential for many data-analysis jobs in high-energy physics. We develop a new class of physically interpretable boost invariant polynomial features (BIPs) for jet tagging that achieves such efficiency. We show that, for both supervised and unsupervised tasks, integrating BIPs with conventional techniques leads to models achieving high accuracy on jet tagging benchmarks while being orders of magnitudes faster to train and evaluate than contemporary deep learning systems.

show abstract

Section: Discussionmentioning

confidence: 99%

Boost invariant polynomials for efficient jet tagging

Munoz

Batatia

Ortner

2022

Mach. Learn.: Sci. Technol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…They have also been demonstrated to exhibit improved data efficiency and generalization capabilities compared to their invariant counterparts on predictions of scalar properties, albeit at a higher computational cost. Nevertheless, given an expressive enough architecture (i.e., using higher-order messages ,,,− and/or enough convolutional layers ,, ), invariant models are sufficient for many property prediction tasks…”

Section: Introductionmentioning

confidence: 99%

3DReact: Geometric Deep Learning for Chemical Reactions

van Gerwen,

Briling,

Bunne

et al. 2024

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Geometric deep learning models, which incorporate the relevant molecular symmetries within the neural network architecture, have considerably improved the accuracy and data efficiency of predictions of molecular properties. Building on this success, we introduce 3DREACT, a geometric deep learning model to predict reaction properties from three-dimensional structures of reactants and products. We demonstrate that the invariant version of the model is sufficient for existing reaction data sets. We illustrate its competitive performance on the prediction of activation barriers on the GDB7-22-TS, Cyclo-23-TS, and Proparg-21-TS data sets in different atom-mapping regimes. We show that, compared to existing models for reaction property prediction, 3DREACT offers a flexible framework that exploits atommapping information, if available, as well as geometries of reactants and products (in an invariant or equivariant fashion). Accordingly, it performs systematically well across different data sets, atom-mapping regimes, as well as both interpolation and extrapolation tasks.

show abstract

“…More sophisticated long-range schemes than that used in this work to name but a few include refs − , and a comparison of their performance would be highly insightful. Moreover, recent equivarient graph-based architectures including MACE, NEQUIP, and GRACE are by construction longer range and are again highly promising to address the question of long-range interactions in an efficient manner. Beyond the bulk, confinement of electrolytes leads to interesting physics and unexpected phenomena that are highly relevant to a range of applications, including blue energy harvesting and desalination. , In particular, extension of our current models to explore the intriguing phenomenon of confinement-induced ion pairing is very attractive.…”

mentioning

confidence: 99%

To Pair or not to Pair? Machine-Learned Explicitly-Correlated Electronic Structure for NaCl in Water

O’Neill,

Shi,

Fong

et al. 2024

J. Phys. Chem. Lett.

View full text Add to dashboard Cite

The extent of ion pairing in solution is an important phenomenon to rationalize transport and thermodynamic properties of electrolytes. A fundamental measure of this pairing is the potential of mean force (PMF) between solvated ions. The relative stabilities of the paired and solvent shared states in the PMF and the barrier between them are highly sensitive to the underlying potential energy surface. However, direct application of accurate electronic structure methods is challenging, since long simulations are required. We develop wave function based machine learning potentials with the random phase approximation (RPA) and second order Møller–Plesset (MP2) perturbation theory for the prototypical system of Na and Cl ions in water. We show both methods in agreement, predicting the paired and solvent shared states to have similar energies (within 0.2 kcal/mol). We also provide the same benchmarks for different DFT functionals as well as insight into the PMF based on simple analyses of the interactions in the system.

show abstract

Multilayer atomic cluster expansion for semilocal interactions

Cited by 16 publications

References 33 publications

Boost invariant polynomials for efficient jet tagging

Boost invariant polynomials for efficient jet tagging

3DReact: Geometric Deep Learning for Chemical Reactions

To Pair or not to Pair? Machine-Learned Explicitly-Correlated Electronic Structure for NaCl in Water

Contact Info

Product

Resources

About