& UDelta;(2) machine learning for reaction property prediction

Article Multidisciplinary Sciences

Learning local equivariant representations for large-scale atomistic dynamics

Albert Musaelian et al.

Summary: This study introduces Allegro, a local equivariant deep neural network interatomic potential architecture that achieves excellent accuracy and scalability in quantum chemistry and molecular simulations.

NATURE COMMUNICATIONS (2023)

添加到收藏夹

Article Multidisciplinary Sciences

Comprehensive exploration of graphically defined reaction spaces

Qiyuan Zhao et al.

Summary: The concept of a graphically-defined model reaction has been used to address the data gap in existing reaction transition state (TS) databases. The resulting dataset, called the Reaction Graph Depth 1 (RGD1) dataset, is composed of 176,992 organic reactions with validated TS and various other properties. RGD1 is the largest and most chemically diverse TS dataset published to date, and can be used to develop novel machine learning models for predicting reaction properties.

SCIENTIFIC DATA (2023)

添加到收藏夹

Article Chemistry, Medicinal

Machine Learning of Reaction Properties via Learned Representations of the Condensed Graph of Reaction

Esther Heid et al.

Summary: This study combined the condensed graph of reaction (CGR) with a graph convolutional neural network (GCNN) to develop a versatile, robust, and compact deep learning model for chemical reactions. The model outperforms current state-of-the-art models in accuracy, applicability, and predictive capabilities for diverse target properties. A large curated dataset of atom-mapped reactions with their target properties is provided as benchmark data for future work.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2022)

添加到收藏夹

Article Chemistry, Physical

Quantum chemistry-augmented neural networks for reactivity prediction: Performance, generalizability, and explainability

Thijs Stuyver et al.

Summary: This study investigates the performance, generalizability, and explainability of structure-based and descriptor-based molecular representations for predictive chemistry tasks. The quantum mechanics-augmented graph neural network (ml-QM-GNN) architecture is introduced, which combines structure-based representations with quantum mechanics descriptors to improve predictive accuracy and generalization to unseen compounds.

JOURNAL OF CHEMICAL PHYSICS (2022)

添加到收藏夹

Article Chemistry, Physical

Conformational Sampling for Transition State Searches on a Computational Budget

Qiyuan Zhao et al.

Summary: In this study, a classifier was trained to learn the features of reaction conformers that lead to successful transition state searches, enabling the down-selection of optimal conformers before incurring the cost of a high-level transition state search. The results showed that neglecting conformer contributions can lead to qualitatively incorrect activation energy estimations, while machine learning classifiers reliably down-select low-barrier reaction conformers.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2022)

添加到收藏夹

Article Chemistry, Physical

Successes and challenges in using machine-learned activation energies in kinetic simulations

I. Ismail et al.

Summary: Machine-learning methods, such as artificial neural networks, are increasingly used to predict the thermodynamic and kinetic properties of chemical reactions. This study investigates the impact of uncertainty associated with machine-learning predictions of activation energies on observable properties of chemical reaction networks. Guidelines for the reliability of machine-learning-based activation energies in kinetics modeling are suggested based on microkinetics simulations.

JOURNAL OF CHEMICAL PHYSICS (2022)

添加到收藏夹

Article Chemistry, Physical

Fast Predictions of Reaction Barrier Heights: Toward Coupled- Cluster Accuracy

Kevin A. Spiekermann et al.

Summary: Quantitative estimates of reaction barriers are crucial for development and prediction of reaction outcomes. However, lack of experimental data and limitations of accurate quantum calculations often impede the obtainment of reliable kinetic values. In this study, a directed message passing neural network is trained on diverse gas-phase reactions to accurately predict barrier heights based on reactant and product information. The model outperforms good density functional theory calculations and suggests the use of transfer learning for future modeling efforts. This model is expected to enhance and expedite kinetic predictions for small molecule chemistry.

JOURNAL OF PHYSICAL CHEMISTRY A (2022)

添加到收藏夹

Article Chemistry, Physical

Graph-Driven Reaction Discovery: Progress, Challenges, and Future Opportunities

Idil Ismail et al.

Summary: Graph-based descriptors provide a simple and compact way to categorize molecular structures and have been widely used for generating chemical reaction network models. Recent advancements have incorporated graph-based methods with semiempirical and ab initio electronic structure calculations to explore mechanistic questions. However, challenges such as chemical accuracy, calculation speeds, and dealing with vast accessible reaction spaces still remain.

JOURNAL OF PHYSICAL CHEMISTRY A (2022)

添加到收藏夹

Article Multidisciplinary Sciences

Deep reaction network exploration at a heterogeneous catalytic interface

Qiyuan Zhao et al.

Summary: This study demonstrates the predictive characterization of reaction networks on heterogeneous catalytic surfaces using automated exploration algorithms. The results show that these algorithms have great potential for exploratory catalytic applications.

NATURE COMMUNICATIONS (2022)

添加到收藏夹

Article Chemistry, Multidisciplinary

Active learning accelerates ab initio molecular dynamics on reactive energy surfaces

Shi Jun Ang et al.

Summary: Modeling dynamical effects in chemical reactions often requires ab initio molecular dynamics simulations, but traditional methods have limitations in accuracy and cost. A data-driven approach combining high-throughput calculations and neural network potentials allows for cost-effective simulation of post-transition state dynamics in various pericyclic reactions. This method can accurately predict reaction outcomes in complex systems.

CHEM (2021)

添加到收藏夹

Article Chemistry, Medicinal

Transferable Ring Corrections for Predicting Enthalpy of Formation of Cyclic Compounds

Qiyuan Zhao et al.

Summary: Computational predictions of the thermodynamic properties of molecules and materials are crucial for contemporary reaction prediction and kinetic modeling. Approximate methods and machine learning are currently the only approaches capable of providing the necessary chemical coverage and throughput. Ring-containing molecules present a challenge for transferability due to nonlocal interactions, impacting thermodynamic properties.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2021)

添加到收藏夹

Article Chemistry, Physical

Toward the design of chemical reactions: Machine learning barriers of competing mechanisms in reactant space

Stefan Heinen et al.

Summary: The interplay of kinetics and thermodynamics governs reactive processes, with a newly introduced R2B machine learning model capable of rapidly and accurately inferring activation energies and transition state geometries, enabling experimental reaction design and providing new possibilities for synthesis efforts.

JOURNAL OF CHEMICAL PHYSICS (2021)

添加到收藏夹

Article Chemistry, Physical

Generative adversarial networks for transition state geometry prediction

Malgorzata Z. Makos et al.

Summary: This study introduces a novel application of generative adversarial networks (GANs) for predicting starting geometries in transition state searches of chemical reactions. The TS-GAN efficiently maps the potential energy space between reactants and products to generate reliable TS guess structures, showing high accuracy and efficiency compared to classical approaches. The current TS-GAN can be extended to any dataset with sufficient chemical reaction data for training, and the software is freely available for training, experimentation, and prediction.

JOURNAL OF CHEMICAL PHYSICS (2021)

添加到收藏夹

Article Multidisciplinary Sciences

Teaching a neural network to attach and detach electrons from molecules

Roman Zubatyuk et al.

Summary: Machine learning algorithms such as Deep-Neural Networks have achieved high accuracy comparable to quantum mechanical methods, allowing for massive simulations. The proposed AIMNet-NSE architecture can predict molecular energies and spin-charges with errors close to reference QM simulations, bypassing the need for QM calculations. This model shows potential in modeling chemical reactivity through learned atomic representations and descriptors.

NATURE COMMUNICATIONS (2021)

添加到收藏夹

Article Chemistry, Multidisciplinary

A transferable active-learning strategy for reactive molecular force fields

Tom A. Young et al.

Summary: This study presents a method of using machine learning to construct fast, accurate, and reactive interatomic potentials for predictive molecular simulations. By leveraging hierarchical and active learning, accurate Gaussian Approximation Potential (GAP) models can be developed autonomously for diverse chemical systems.

CHEMICAL SCIENCE (2021)

添加到收藏夹

Review Chemistry, Multidisciplinary

Autonomous Discovery in the Chemical Sciences Part I: Progress

Connor W. Coley et al.

ANGEWANDTE CHEMIE-INTERNATIONAL EDITION (2020)

添加到收藏夹

Review Chemistry, Physical

The Exploration of Chemical Reaction Networks

Jan P. Unsleber et al.

ANNUAL REVIEW OF PHYSICAL CHEMISTRY, VOL 71 (2020)

添加到收藏夹

Article Chemistry, Physical

Deep Learning of Activation Energies

Colin A. Grambow et al.

JOURNAL OF PHYSICAL CHEMISTRY LETTERS (2020)

添加到收藏夹

Article Multidisciplinary Sciences

Reactants, products, and transition states of elementary chemical reactions based on quantum chemistry

Colin A. Grambow et al.

SCIENTIFIC DATA (2020)

添加到收藏夹

Article Chemistry, Physical

Generating transition states of isomerization reactions with deep learning

Lagnajit Pattanaik et al.

PHYSICAL CHEMISTRY CHEMICAL PHYSICS (2020)

添加到收藏夹

Article Chemistry, Physical

GFN2-xTB-An Accurate and Broadly Parametrized Self-Consistent Tight-Binding Quantum Chemical Method with Multipole Electrostatics and Density-Dependent Dispersion Contributions

Christoph Bannwarth et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2019)

添加到收藏夹

Article Chemistry, Physical

Predicting Molecular Energy Using Force-Field Optimized Geometries and Atomic Vector Representations Learned from an Improved Deep Tensor Neural Network

Jianing Lu et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2019)

添加到收藏夹

Article Chemistry, Medicinal

Analyzing Learned Molecular Representations for Property Prediction

Kevin Yang et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2019)

添加到收藏夹

Article Multidisciplinary Sciences

Accurate and transferable multitask prediction of chemical properties with an atoms-in-molecules neural network

Roman Zubatyuk et al.

SCIENCE ADVANCES (2019)

添加到收藏夹

Article Chemistry, Physical

Combined Automated Reaction Pathway Searches and Sparse Modeling Analysis for Catalytic Properties of Lowest Energy Twins of Cu13

Takeshi Iwasa et al.

JOURNAL OF PHYSICAL CHEMISTRY A (2019)

添加到收藏夹

Article Chemistry, Multidisciplinary

Glucose to 5-Hydroxymethylfurfural: Origin of Site-Selectivity Resolved by Machine Learning Based Reaction Sampling

Pei-Lin Kang et al.

JOURNAL OF THE AMERICAN CHEMICAL SOCIETY (2019)

添加到收藏夹

Article Chemistry, Multidisciplinary

The TensorMol-0.1 model chemistry: a neural network augmented with long-range physics

Kun Yao et al.

CHEMICAL SCIENCE (2018)

添加到收藏夹

Article Chemistry, Physical

Stochastic surface walking reaction sampling for resolving heterogeneous catalytic reaction network: A revisit to the mechanism of water-gas shift reaction on Cu

Xiao-Jie Zhang et al.

JOURNAL OF CHEMICAL PHYSICS (2017)

添加到收藏夹

Article Chemistry, Physical

Automatic Generation of Microkinetic Mechanisms for Heterogeneous Catalysis

C. Franklin Goldsmith et al.

JOURNAL OF PHYSICAL CHEMISTRY C (2017)

添加到收藏夹

Article Chemistry, Multidisciplinary

ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost

J. S. Smith et al.

CHEMICAL SCIENCE (2017)

添加到收藏夹

Article Chemistry, Physical

Automated Discovery of Elementary Chemical Reaction Steps Using Freezing String and Berny Optimization Methods

Yury V. Suleimanov et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2015)

添加到收藏夹

Article Chemistry, Physical

Big Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach

Raghunathan Ramakrishnan et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2015)

添加到收藏夹

Article Chemistry, Multidisciplinary

Exploring Transition State Structures for Intramolecular Pathways by the Artificial Force Induced Reaction Method

Satoshi Maeda et al.

JOURNAL OF COMPUTATIONAL CHEMISTRY (2014)

添加到收藏夹

Article Chemistry, Physical

Growing string method with interpolation and optimization in internal coordinates: Method and examples

Paul M. Zimmerman

JOURNAL OF CHEMICAL PHYSICS (2013)

添加到收藏夹

Article Chemistry, Physical

Stochastic Surface Walking Method for Structure Prediction and Pathway Searching

Cheng Shang et al.

JOURNAL OF CHEMICAL THEORY AND COMPUTATION (2013)

添加到收藏夹

Article Thermodynamics

Comprehensive chemical kinetic modeling of the oxidation of 2-methylalkanes from C7 to C20

S. M. Sarathy et al.

COMBUSTION AND FLAME (2011)

添加到收藏夹

Article Chemistry, Physical

Efficient exploration of reaction paths via a freezing string method

Andrew Behn et al.

JOURNAL OF CHEMICAL PHYSICS (2011)

添加到收藏夹

Article Chemistry, Medicinal

Extended-Connectivity Fingerprints

David Rogers et al.

JOURNAL OF CHEMICAL INFORMATION AND MODELING (2010)

添加到收藏夹

Article Chemistry, Physical

A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu

Stefan Grimme et al.

JOURNAL OF CHEMICAL PHYSICS (2010)

添加到收藏夹

Article Chemistry, Physical

Gaussian-4 theory

Larry A. Curtiss et al.

JOURNAL OF CHEMICAL PHYSICS (2007)

添加到收藏夹

Article Thermodynamics

Computational combustion

CK Westbrook et al.

PROCEEDINGS OF THE COMBUSTION INSTITUTE (2005)

添加到收藏夹

Article Chemistry, Physical

A growing string method for determining transition states: Comparison to the nudged elastic band and string methods

B Peters et al.

JOURNAL OF CHEMICAL PHYSICS (2004)

添加到收藏夹

Article Chemistry, Physical

A climbing image nudged elastic band method for finding saddle points and minimum energy paths

G Henkelman et al.

JOURNAL OF CHEMICAL PHYSICS (2000)

添加到收藏夹

& UDelta;(2) machine learning for reaction property prediction

相关参考文献

导出引文

分享论文