Machine learning of accurate energy-conserving molecular force fields
By Stefan Chmiela, Alexandre Tkatchenko, Huziel E. Sauceda, Igor Poltavsky, Kristof T. Schütt, K. Müller. 2016
1,263 citations recorded by the source corpus.
The law of energy conservation is used to develop an efficient machine learning approach to construct accurate force fields. Using conservation of energy—a fundamental property of closed classical and quantum mechanical systems—we develop an efficient gradient-domain machine learning (GDML) approach to construct accurate molecular force fields using a restricted number of samples from ab initio molecular dynamics (AIMD) trajectories. The GDML implementation is able to reproduce global potential energy surfaces of intermediate-sized molecules with an accuracy of 0.3 kcal mol−1 for energies and 1 kcal mol−1 Å̊−1 for atomic forces using only 1000 conformational geometries for training. We demonstrate this accuracy for AIMD trajectories of molecules, including benzene, toluene, naphthalene, ethanol, uracil, and aspirin. The challenge of constructing conservative force fields is accomplished in our work by learning in a Hilbert space of vector-valued functions that obey the law of energy conservation. The GDML approach enables quantitative molecular dynamics simulations for molecules at a fraction of cost of explicit AIMD calculations, thereby allowing the construction of efficient force fields with the accuracy and transferability of high-level ab initio methods.
Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology, Computational Chemistry & Drug Discovery, AI for Physical Sciences
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computational Theory and Mathematics
- Tess E. Smidt
- Michael Gastegger
- Kristof T. Schütt
- K. Müller
- Oliver T. Unke
- Risi Kondor
- Stephan Günnemann
- Jian Tang
In Materials Chemistry
- Tess E. Smidt
- Risi Kondor
- K. Müller
- Michael Gastegger
- Kristof T. Schütt
- Stephan Günnemann
- Johannes Klicpera
- Oliver T. Unke
In Molecular Biology
- Tess E. Smidt
- Michael Gastegger
- K. Müller
- Oliver T. Unke
- Risi Kondor
- Kristof T. Schütt
- Stephan Günnemann
- Wojciech Samek
In Computational Chemistry & Drug Discovery
- Tess E. Smidt
- K. Müller
- Michael Gastegger
- Stephan Günnemann
- Johannes Klicpera
- Kristof T. Schütt
- Oliver T. Unke
- Risi Kondor
In AI for Physical Sciences
- Wojciech Samek
- K. Müller
- Risi Kondor
- Tess E. Smidt
- Grgoire Montavon
- Stephan Günnemann
- Johannes Klicpera
- Dongkun Zhang
Citations in the corpus
- Directional Message Passing for Molecular Graphs (2020, methodology): “, 2013) with Gaussian processes (Bartók et al., 2010; 2017; Chmiela et al., 2017) or neural networks (Behler & Parrinello, 2007).”
- Methods for interpreting and understanding deep neural networks (2017, methodology): “In the domain of atomistic simulations, powerful machine learning models have been produced to link molecular structure to electronic properties [48,23,58,18].”
- SchNet: A continuous-filter convolutional neural network for modeling quantum interactions (2017, methodology): “GDML captures the relationship between energy and interatomic forces (see Eq.”
- Cormorant: Covariant Molecular Neural Networks (2019, methodology): “MD-17 [Chmiela et al., 2016] is a dataset of eight small organic molecules (see .)”
- Tensor Field Networks: Rotation- and Translation-Equivariant Neural Networks for 3D Point Clouds (2018, methodology): “We have successfully trained tensor field networks on large standard datasets, including QM9 energy [26], MD17 molecular dynamics forces [27], and ModelNet40 [28] classification.”
- N-body Networks: a Covariant Hierarchical Neural Network Architecture for Learning Atomic Potentials (2018, positive): “…amount of activity in machine learned atomic potentials (MLAP), and molecular dynamics simulations based on this approach are starting to yield results that outperform other methods (Bart´ok et al., 2010; Behler, 2015; Shapeev, 2015; Chmiela et al., 2016; Zhang et al., 2017; Sch¨utt et al., 2017).”
- Utilizing graph machine learning within drug discovery and development (2021, methodology): “This is a popular task with well-established benchmarks such as QM9 [179] and MD17 [180].”
- Machine Learning and Materials Informatics: Recent Applications and Prospects (2017, background): “Several candidates, including those based on symmetry functions [76–78], bispectra of neighborhood atomic densities [79], Coulomb matrices (and its variants) [80, 81], smooth overlap of atomic positions (SOAP) [82–85], and others [86,87], have been proposed.”
- Recent advances and applications of deep learning methods in materials science (2021, neutral): “org/datasets [81] Tox21 760k https://tox21.”
- Emerging materials intelligence ecosystems propelled by machine learning (2020, methodology): “For instance, rotationally invariant fingerprints were designed to learn system energies, and a gradient-based ML approach was developed to construct energy-conserving force fields 226 .”