Search Menu
DOE PAGES title logo U.S. Department of Energy
Office of Scientific and Technical Information
Search Scholarly Publications

Title: Analytical ab initio hessian from a deep learning potential for transition state optimization

Journal Article · · Nature Communications
ORCiD logo [1];  [2];  [1]; ORCiD logo [3]; ORCiD logo [4]; ORCiD logo [5]; ORCiD logo [1]; ORCiD logo [2]
  1. University of California, Berkeley, CA (United States); Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
  2. Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
  3. Quantum-Si, Branford, CT (United States)
  4. Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States); University of California, Berkeley, CA (United States)
  5. Sandia National Laboratories (SNL-CA), Livermore, CA (United States)
+ Show Author Affiliations

Identifying transition states—saddle points on the potential energy surface connecting reactant and product minima—is central to predicting kinetic barriers and understanding chemical reaction mechanisms. In this work, we train a fully differentiable equivariant neural network potential, NewtonNet, on thousands of organic reactions and derive the analytical Hessians. By reducing the computational cost by several orders of magnitude relative to the density functional theory (DFT) ab initio source, we can afford to use the learned Hessians at every step for the saddle point optimizations. We show that the full machine learned (ML) Hessian robustly finds the transition states of 240 unseen organic reactions, even when the quality of the initial guess structures are degraded, while reducing the number of optimization steps to convergence by 2–3× compared to the quasi-Newton DFT and ML methods. All data generation, NewtonNet model, and ML transition state finding methods are available in an automated workflow.

Research Organization:
Lawrence Berkeley National Laboratory (LBNL), Berkeley, CA (United States)
Sponsoring Organization:
USDOE; USDOE Laboratory Directed Research and Development (LDRD) Program; USDOE Office of Science (SC), Basic Energy Sciences (BES). Chemical Sciences, Geosciences & Biosciences Division (CSGB); USDOE Office of Science (SC), Basic Energy Sciences (BES). Scientific User Facilities (SUF)
Grant/Contract Number:
AC02-05CH11231
OSTI ID:
2466264
Journal Information:
Nature Communications, Journal Name: Nature Communications Journal Issue: 1 Vol. 15; ISSN 2041-1723
Publisher:
Nature Publishing GroupCopyright Statement
Country of Publication:
United States
Language:
English

References (54)

A reduced-restricted-quasi-Newton-Raphson method for locating and optimizing energy crossing points between two potential energy surfaces journal June 1997
How good is a Broyden-Fletcher-Goldfarb-Shanno-like update Hessian formula to locate transition structures? Specific reformulation of Broyden-Fletcher-Goldfarb-Shanno for optimizing saddle points journal February 1998
Practical Methods of Optimization book January 2000
Optimization of equilibrium geometries and transition structures journal July 1982
An algorithm for the location of transition states journal August 1986
Remarks on the updated Hessian matrix methods journal January 2003
Location of saddle points and minimum energy paths by a constrained simplex optimization procedure journal January 1979
Estimating the hessian for gradient-type geometry optimizations journal January 1984
N�herungsformeln f�r die Zylinderfunktionen f�r gro�e Werte des Arguments und unbeschr�nkt ver�nderliche Werte des Index journal December 1909
On the automatic restricted-step rational-function-optimization method
  • Besalú, Emili; Bofill, Josep Maria
  • Theoretical Chemistry Accounts: Theory, Computation, and Modeling (Theoretica Chimica Acta), Vol. 100, Issue 5-6 https://doi.org/10.1007/s002140050387
journal December 1998
On the use of a Hessian model function in molecular geometry optimizations journal July 1995
The iterative calculation of a few of the lowest eigenvalues and corresponding eigenvectors of large real-symmetric matrices journal January 1975
Implementation of analytic derivative methods in quantum chemistry journal September 1989
Geometry optimization by direct inversion in the iterative subspace journal January 1984
A generalized synchronous transit method for transition state location journal October 2003
KinBot: Automated stationary point search on potential energy surfaces journal March 2020
DP-GEN: A concurrent learning platform for the generation of reliable deep learning based potential energy models journal February 2020
Sigmoid-weighted linear units for neural network function approximation in reinforcement learning journal November 2018
Hessian Matrix Update Scheme for Transition State Search Based on Gaussian Process Regression journal July 2020
Sella, an Open-Source Automation-Friendly Molecular Saddle Point Optimizer journal October 2022
Gaussian Process Regression for Transition State Search journal October 2018
Accelerated Saddle Point Refinement through Full Exploitation of Partial Hessian Diagonalization journal October 2019
Recent Advances toward Efficient Calculation of Higher Nuclear Derivatives in Quantum Chemistry journal October 2022
Using Hessian Updating To Increase the Efficiency of a Hessian Based Predictor-Corrector Reaction Path Following Method journal October 2004
General methods for geometry and wave function optimization journal November 1992
Search for stationary points on surfaces journal January 1985
Chemical Reaction Networks Explain Gas Evolution Mechanisms in Mg-Ion Batteries journal May 2023
Towards exact molecular dynamics simulations with machine-learned force fields journal September 2018
Diffusion-based generative AI for exploring transition states from 2D molecular graphs journal January 2024
Transition1x - a dataset for building generalizable reactive machine learning potentials journal December 2022
Using machine learning to go beyond potential energy surface benchmarking for chemical reactivity journal November 2023
Accurate transition state generation with an object-aware equivariant elementary reaction diffusion model journal December 2023
ANI-1: an extensible neural network potential with DFT accuracy at force field computational cost journal January 2017
Variational transition state theory: theoretical framework and recent developments journal January 2017
Predictive stochastic analysis of massive filter-based electrochemical reaction networks journal January 2023
A climbing image nudged elastic band method for finding saddle points and minimum energy paths journal December 2000
Using force fields methods for locating transition structures journal November 2003
Self‐Consistent Molecular‐Orbital Methods. IX. An Extended Gaussian‐Type Basis for Molecular‐Orbital Studies of Organic Molecules journal January 1971
A growing string method for determining transition states: Comparison to the nudged elastic band and string methods journal May 2004
Systematic optimization of long-range corrected hybrid density functionals journal February 2008
A finite difference Davidson procedure to sidestep full ab initio hessian calculation: Application to characterization of stationary points and transition state searches journal April 2014
Less is more: Sampling chemical space with active learning journal June 2018
Software for the frontiers of quantum chemistry: An overview of developments in the Q-Chem 5 package journal August 2021
Geometry optimization speedup through a geodesic approach to internal coordinates journal September 2021
Assessing and rationalizing the performance of Hessian update schemes for reaction path Hamiltonian rate calculations journal November 2021
A Simplex Method for Function Minimization journal January 1965
Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces journal April 2007
A solution for the best rotation to relate two sets of vectors journal September 1976
On implementing 2D rectangular assignment algorithms journal August 2016
Numerical Methods for Unconstrained Optimization and Nonlinear Equations book January 1996
Nudged elastic band method for finding minimum energy paths of transitions
  • JÓNsson, Hannes; Mills, Greg; Jacobsen, Karsten W.
  • Proceedings of the International School of Physics, Classical and Quantum Dynamics in Condensed Phase Simulations https://doi.org/10.1142/9789812839664_0016
conference November 2011
Open Babel: An open chemical toolbox journal October 2011
Methods of conjugate gradients for solving linear systems journal December 1952
Data for Deep Learning of ab initio Hessians for Transition State Optimization dataset January 2024

Similar Records

Related Subjects

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CHEMISTRY