Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Predicting Reaction Outcomes

Predicting Reaction Outcomes

Kinetics describes the rate and path by which a reaction occurs. In contrast, thermodynamics deals with state functions and describes the properties, behavior, and components of a system. It is not concerned with the path taken by the process and cannot address the rate at which a reaction occurs. Although it does provide information about what can happen during a reaction process, it does not describe the detailed steps of what appears on an atomic or a molecular level. On the other hand,...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Scalable emulation of protein equilibrium ensembles with generative deep learning.

Science (New York, N.Y.)·2025

Same author

Excited-state nonadiabatic dynamics in explicit solvent using machine learned interatomic potentials.

Digital discovery·2025

Same author

Correction to "Solvatochromic Effects on the Absorption Spectrum of 2-Thiocytosine".

The journal of physical chemistry. B·2024

Same author

SpaiNN: equivariant message passing for excited-state nonadiabatic molecular dynamics.

Chemical science·2024

Same author

Biomolecular dynamics with machine-learned quantum-mechanical force fields trained on diverse chemical fragments.

Science advances·2024

Same author

Transferability of atomic energies from alchemical decomposition.

The Journal of chemical physics·2024

Search research articles

Related Experiment Video

Updated: Mar 30, 2026

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

High-Dimensional Neural Network Potentials for Organic Reactions and an Improved Training Algorithm.

Michael Gastegger¹, Philipp Marquetand¹

¹Institute of Theoretical Chemistry, University of Vienna , Währinger Str. 17, 1090 Vienna, Austria.

Journal of Chemical Theory and Computation

|November 18, 2015

Summary

This summary is machine-generated.

High-dimensional neural networks (HDNNs) now predict molecular energies for organic reactions. A new training algorithm, the element-decoupled global extended Kalman filter (ED-GEKF), improves accuracy and speed for these complex simulations.

More Related Videos

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Related Experiment Videos

Last Updated: Mar 30, 2026

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Deep Neural Networks for Image-Based Dietary Assessment

Deep Neural Networks for Image-Based Dietary Assessment

Published on: March 13, 2021

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

Area of Science:

Computational Chemistry
Materials Science
Machine Learning

Background:

Artificial neural networks (NNs) offer accurate molecular potential energy prediction for large-scale simulations.
NNs provide computational efficiency similar to force fields but can model complex bonding and coordination.
High-dimensional neural networks (HDNNs) are adapted for molecular property prediction, typically in solid-state and surface interactions.

Purpose of the Study:

To apply the HDNN approach to an organic reaction, specifically the Claisen rearrangement.
To introduce and evaluate a novel training algorithm, the element-decoupled global extended Kalman filter (ED-GEKF), for HDNN potential construction.
To assess the performance of ED-GEKF against existing Kalman filter variants in HDNN training.

Main Methods:

Developed and implemented the element-decoupled global extended Kalman filter (ED-GEKF) training algorithm.
Utilized density functional theory (DFT) to generate reference data via metadynamics.
Trained HDNN potentials using the ED-GEKF algorithm on the Claisen rearrangement reaction.
Investigated the impact of including forces in the ED-GEKF training process.

Main Results:

The ED-GEKF algorithm demonstrated superior accuracy and training speed compared to other Kalman filter methods for HDNN training.
The first successful application of HDNNs to model an organic reaction (Claisen rearrangement) was achieved.
The study analyzed the influence of incorporating force data during ED-GEKF training on potential accuracy.

Conclusions:

The ED-GEKF is an effective and efficient algorithm for training HDNNs for molecular simulations, particularly for organic reactions.
HDNNs, coupled with advanced training algorithms, show significant promise for simulating complex chemical processes.
This work expands the application scope of HDNNs beyond solid-state systems to reactive organic chemistry.