Machine learning of molecular properties: Locality and active learning

Gubaev, Konstantin; Podryabinkin, Evgeny V.; Shapeev, Alexander V.

doi:10.1063/1.5005095

Title: Machine learning of molecular properties: Locality and active learning

Abstract

In recent years, the machine learning techniques have shown great potent1ial in various problems from a multitude of disciplines, including materials design and drug discovery. The high computational speed on the one hand and the accuracy comparable to that of density functional theory on another hand make machine learning algorithms efficient for high-throughput screening through chemical and configurational space. However, the machine learning algorithms available in the literature require large training datasets to reach the chemical accuracy and also show large errors for the so-called outliers—the out-of-sample molecules, not well-represented in the training set. In the present paper, we propose a new machine learning algorithm for predicting molecular properties that addresses these two issues: it is based on a local model of interatomic interactions providing high accuracy when trained on relatively small training sets and an active learning algorithm of optimally choosing the training set that significantly reduces the errors for the outliers. We compare our model to the other state-of-the-art algorithms from the literature on the widely used benchmark tests.

Authors:

;

; Shapeev, Alexander V.

Publication Date:: Thu Apr 19 00:00:00 EDT 2018

Sponsoring Org.:: USDOE

OSTI Identifier:: 1433745

Grant/Contract Number:: 1150-06_2015

Resource Type:: Publisher's Accepted Manuscript

Journal Name:: Journal of Chemical Physics

Additional Journal Information:: Journal Name: Journal of Chemical Physics Journal Volume: 148 Journal Issue: 24; Journal ID: ISSN 0021-9606

Publisher:: American Institute of Physics

Country of Publication:: United States

Language:: English

Citation Formats


                    Gubaev, Konstantin, Podryabinkin, Evgeny V., and Shapeev, Alexander V. Machine learning of molecular properties: Locality and active learning.  United States: N. p., 2018. 
Web.  doi:10.1063/1.5005095.

Copy to clipboard


                    Gubaev, Konstantin, Podryabinkin, Evgeny V., & Shapeev, Alexander V. Machine learning of molecular properties: Locality and active learning.  United States.  https://doi.org/10.1063/1.5005095

Copy to clipboard


                    Gubaev, Konstantin, Podryabinkin, Evgeny V., and Shapeev, Alexander V. Thu .  
"Machine learning of molecular properties: Locality and active learning".  United States.  https://doi.org/10.1063/1.5005095.

Copy to clipboard


                    
@article{osti_1433745,

  title        = {Machine learning of molecular properties: Locality and active learning},

  author       = {Gubaev, Konstantin and Podryabinkin, Evgeny V. and Shapeev, Alexander V.},

  abstractNote = {In recent years, the machine learning techniques have shown great potent1ial in various problems from a multitude of disciplines, including materials design and drug discovery. The high computational speed on the one hand and the accuracy comparable to that of density functional theory on another hand make machine learning algorithms efficient for high-throughput screening through chemical and configurational space. However, the machine learning algorithms available in the literature require large training datasets to reach the chemical accuracy and also show large errors for the so-called outliers—the out-of-sample molecules, not well-represented in the training set. In the present paper, we propose a new machine learning algorithm for predicting molecular properties that addresses these two issues: it is based on a local model of interatomic interactions providing high accuracy when trained on relatively small training sets and an active learning algorithm of optimally choosing the training set that significantly reduces the errors for the outliers. We compare our model to the other state-of-the-art algorithms from the literature on the widely used benchmark tests.},

  doi          = {10.1063/1.5005095},

  journal      = {Journal of Chemical Physics},

  number       = 24,

  volume       = 148,

  place        = {United States},

  year         = {Thu Apr 19 00:00:00 EDT 2018},

  month        = {Thu Apr 19 00:00:00 EDT 2018}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (Publisher)

Publisher's Version of Record
https://doi.org/10.1063/1.5005095

Other availability

Search WorldCat to find libraries that may hold this journal

Citation Metrics:

Cited by: 101 works

Citation information provided by
Web of Science

Save / Share:

Export Metadata

Save to My Library

Works referenced in this record:

Machine learning of molecular electronic properties in chemical compound space
journal, September 2013

Montavon, Grégoire; Rupp, Matthias; Gobre, Vivekanand
New Journal of Physics, Vol. 15, Issue 9
DOI: 10.1088/1367-2630/15/9/095003

Quantum-chemical insights from deep tensor neural networks
journal, January 2017

Schütt, Kristof T.; Arbabzadah, Farhad; Chmiela, Stefan
Nature Communications, Vol. 8, Issue 1
DOI: 10.1038/ncomms13890

Genetic Optimization of Training Sets for Improved Machine Learning Models of Molecular Properties
journal, March 2017

Browning, Nicholas J.; Ramakrishnan, Raghunathan; von Lilienfeld, O. Anatole
The Journal of Physical Chemistry Letters, Vol. 8, Issue 7
DOI: 10.1021/acs.jpclett.7b00038

Big Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach
journal, April 2015

Ramakrishnan, Raghunathan; Dral, Pavlo O.; Rupp, Matthias
Journal of Chemical Theory and Computation, Vol. 11, Issue 5
DOI: 10.1021/acs.jctc.5b00099

Communication: Understanding molecular representations in machine learning: The role of uniqueness and target similarity
journal, October 2016

Huang, Bing; von Lilienfeld, O. Anatole
The Journal of Chemical Physics, Vol. 145, Issue 16
DOI: 10.1063/1.4964627

Quantum chemistry structures and properties of 134 kilo molecules
journal, August 2014

Ramakrishnan, Raghunathan; Dral, Pavlo O.; Rupp, Matthias
Scientific Data, Vol. 1, Issue 1
DOI: 10.1038/sdata.2014.22

Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space
journal, June 2015

Hansen, Katja; Biegler, Franziska; Ramakrishnan, Raghunathan
The Journal of Physical Chemistry Letters, Vol. 6, Issue 12
DOI: 10.1021/acs.jpclett.5b00831

Prediction Errors of Molecular Machine Learning Models Lower than Hybrid DFT Error
journal, October 2017

Faber, Felix A.; Hutchison, Luke; Huang, Bing
Journal of Chemical Theory and Computation, Vol. 13, Issue 11
DOI: 10.1021/acs.jctc.7b00577

Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17
journal, November 2012

Ruddigkeit, Lars; van Deursen, Ruud; Blum, Lorenz C.
Journal of Chemical Information and Modeling, Vol. 52, Issue 11
DOI: 10.1021/ci300415d

Hierarchical modeling of molecular energies using a deep neural network
journal, June 2018

Lubbers, Nicholas; Smith, Justin S.; Barros, Kipton
The Journal of Chemical Physics, Vol. 148, Issue 24
DOI: 10.1063/1.5011181

Active learning of linearly parametrized interatomic potentials
journal, December 2017

Podryabinkin, Evgeny V.; Shapeev, Alexander V.
Computational Materials Science, Vol. 140
DOI: 10.1016/j.commatsci.2017.08.031

A computational high-throughput search for new ternary superalloys
journal, January 2017

Nyshadham, Chandramouli; Oses, Corey; Hansen, Jacob E.
Acta Materialia, Vol. 122
DOI: 10.1016/j.actamat.2016.09.017

Machine Learning for Quantum Mechanical Properties of Atoms in Molecules
journal, July 2015

Rupp, Matthias; Ramakrishnan, Raghunathan; von Lilienfeld, O. Anatole
The Journal of Physical Chemistry Letters, Vol. 6, Issue 16
DOI: 10.1021/acs.jpclett.5b01456

Similar Records in DOE PAGES and OSTI.GOV collections:

Accounting for Training Data Error in Machine Learning Applied to Earth Observations

Journal Article Elmes, Arthur ; Alemohammad, Hamed ; Avery, Ryan ; ... - Remote Sensing

Remote sensing, or Earth Observation (EO), is increasingly used to understand Earth system dynamics and create continuous and categorical maps of biophysical properties and land cover, especially based on recent advances in machine learning (ML). ML models typically require large, spatially explicit training datasets to make accurate predictions. Training data (TD) are typically generated by digitizing polygons on high spatial-resolution imagery, by collecting in situ data, or by using pre-existing datasets. TD are often assumed to accurately represent the truth, but in practice almost always have error, stemming from (1) sample design, and (2) sample collection errors. The latter ismore »« less
Cited by 25
https://doi.org/10.3390/rs12061034

Full Text Available
QSO photometric redshifts using machine learning and neural networks

Journal Article Curran, S. J. ; Moss, J. P. ; Perrott, Y. C. - Monthly Notices of the Royal Astronomical Society

ABSTRACT The scientific value of the next generation of large continuum surveys would be greatly increased if the redshifts of the newly detected sources could be rapidly and reliably estimated. Given the observational expense of obtaining spectroscopic redshifts for the large number of new detections expected, there has been substantial recent work on using machine learning techniques to obtain photometric redshifts. Here, we compare the accuracy of the predicted photometric redshifts obtained from deep learning (DL) with the k-nearest neighbour (kNN) and the decision tree regression (DTR) algorithms. We find using a combination of near-infrared, visible, and ultraviolet magnitudes, trainedmore »« less
https://doi.org/10.1093/mnras/stab485
Machine Learning Classical Interatomic Potentials for Molecular Dynamics from First-Principles Training Data

Journal Article Chan, Henry ; Narayanan, Badri ; Cherukara, Mathew J. ; ... - Journal of Physical Chemistry. C

The ever-increasing power of modern supercomputers, along with the availability of highly scalable atomistic simulation codes, has begun to revolutionize predictive modeling of materials. In particular, molecular dynamics (MD) has led to breakthrough advances in diverse fields, including tribology, catalysis, sensing, and nanoparticle self-assembly. Additionally, recent integration of MD simulations with X-ray characterization has demonstrated promise in real-time 3-D characterization of materials on the atomic scale. The popularity of MD is driven by its applicability at disparate length/time scales, ranging from ab initio MD (hundreds of atoms and tens of picoseconds) to all-atom classical MD (millions of atoms over microseconds),more »« less
Cited by 50
https://doi.org/10.1021/acs.jpcc.8b09917

Full Text Available
Machine Learning Classical Interatomic Potentials for Molecular Dynamics from First-Principles Training Data

Journal Article Chan, Henry ; Narayanan, Badri ; Cherukara, Matthew J. ; ... - Journal of Physical Chemistry. C

The ever-increasing power of modern super-computers, along with the availability of highly scalable atomistic simulation codes, has begun to revolutionize predictive modeling of materials. In particular, molecular dynamics (MD) has led to breakthrough advances in diverse fields, including tribology, catalysis, sensing, and nanoparticle self-assembly. Furthermore, recent integration of MD simulations with X-ray characterization has demonstrated promise in real-time 3-D characterization of materials on the atomic scale. The popularity of MD is driven by its applicability at disparate length/time scales, ranging from ab initio MD (hundreds of atoms and tens of picoseconds) to all-atom classical MD (millions of atoms over microseconds),more »« less
https://doi.org/10.1021/acs.jpcc.8b09917
Predicting Metabolic Reaction Networks with Perturbation-Theory Machine Learning (PTML) Models

Journal Article Diéguez-Santana, Karel ; Casañola-Martin, Gerardo M. ; Green, James R. ; ... - Current Topics in Medicinal Chemistry

Background:Checking the connectivity (structure) of complex Metabolic Reaction Networks(MRNs) models proposed for new microorganisms with promising properties is an importantgoal for chemical biology. Objective:In principle, we can perform a hand-on checking (Manual Curation). However, this is achallenging task due to the high number of combinations of pairs of nodes (possible metabolic reactions). Results:The CPTML linear model obtained using the LDA algorithm is able to discriminate nodes(metabolites) with the correct assignation of reactions from incorrect nodes with values of accuracy,specificity, and sensitivity in the range of 85-100% in both training and external validation dataseries. Methods:In this work, we used Combinatorial Perturbationmore »« less
https://doi.org/10.2174/1568026621666210331161144

Similar Records

Title: Machine learning of molecular properties: Locality and active learning

Abstract

Citation Formats

Machine learning of molecular electronic properties in chemical compound space journal, September 2013

Quantum-chemical insights from deep tensor neural networks journal, January 2017

Genetic Optimization of Training Sets for Improved Machine Learning Models of Molecular Properties journal, March 2017

Big Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach journal, April 2015

Communication: Understanding molecular representations in machine learning: The role of uniqueness and target similarity journal, October 2016

Quantum chemistry structures and properties of 134 kilo molecules journal, August 2014

Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space journal, June 2015

Prediction Errors of Molecular Machine Learning Models Lower than Hybrid DFT Error journal, October 2017

Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17 journal, November 2012

Hierarchical modeling of molecular energies using a deep neural network journal, June 2018

Active learning of linearly parametrized interatomic potentials journal, December 2017

A computational high-throughput search for new ternary superalloys journal, January 2017

Machine Learning for Quantum Mechanical Properties of Atoms in Molecules journal, July 2015

Machine learning of molecular electronic properties in chemical compound space
journal, September 2013

Quantum-chemical insights from deep tensor neural networks
journal, January 2017

Genetic Optimization of Training Sets for Improved Machine Learning Models of Molecular Properties
journal, March 2017

Big Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach
journal, April 2015

Communication: Understanding molecular representations in machine learning: The role of uniqueness and target similarity
journal, October 2016

Quantum chemistry structures and properties of 134 kilo molecules
journal, August 2014

Machine Learning Predictions of Molecular Properties: Accurate Many-Body Potentials and Nonlocality in Chemical Space
journal, June 2015

Prediction Errors of Molecular Machine Learning Models Lower than Hybrid DFT Error
journal, October 2017

Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17
journal, November 2012

Hierarchical modeling of molecular energies using a deep neural network
journal, June 2018

Active learning of linearly parametrized interatomic potentials
journal, December 2017

A computational high-throughput search for new ternary superalloys
journal, January 2017

Machine Learning for Quantum Mechanical Properties of Atoms in Molecules
journal, July 2015