Alec Owens

Bio: Alec Owens is an academic researcher from University College London. The author has contributed to research in topics: Ab initio & Rotational–vibrational spectroscopy. The author has an hindex of 15, co-authored 41 publications receiving 952 citations. Previous affiliations of Alec Owens include University of Hamburg & Max Planck Society.

The 2020 release of the ExoMol database: molecular line lists for exoplanet and other hot atmospheres

Abstract: The ExoMol database ( www.exomol.com ) provides molecular data for spectroscopic studies of hot atmospheres. While the data are intended for studies of exoplanets and other astronomical bodies, the dataset is widely applicable. The basic form of the database is extensive line lists; these are supplemented with partition functions, state lifetimes, cooling functions, Lande g-factors, temperature-dependent cross sections, opacities, pressure broadening parameters, k-coefficients and dipoles. This paper presents the latest release of the database which has been expanded to consider 80 molecules and 190 isotopologues totaling over 700 billion transitions. While the spectroscopic data are concentrated at infrared and visible wavelengths, ultraviolet transitions are being increasingly considered in response to requests from observers. The core of the database comes from the ExoMol project which primarily uses theoretical methods, albeit usually fine-tuned to reproduce laboratory spectra, to generate very extensive line lists for studies of hot bodies. The data have recently been supplemented by line lists derived from direct laboratory observations, albeit usually with the use of ab initiotransition intensities. A major push in the new release is towards accurate characterisation of transition frequencies for use in high resolution studies of exoplanets and other bodies.

Structure-based sampling and self-correcting machine learning for accurate calculations of potential energy surfaces and vibrational levels.

Abstract: We present an efficient approach for generating highly accurate molecular potential energy surfaces (PESs) using self-correcting, kernel ridge regression (KRR) based machine learning (ML). We introduce structure-based sampling to automatically assign nuclear configurations from a pre-defined grid to the training and prediction sets, respectively. Accurate high-level ab initio energies are required only for the points in the training set, while the energies for the remaining points are provided by the ML model with negligible computational cost. The proposed sampling procedure is shown to be superior to random sampling and also eliminates the need for training several ML models. Self-correcting machine learning has been implemented such that each additional layer corrects errors from the previous layer. The performance of our approach is demonstrated in a case study on a published high-level ab initio PES of methyl chloride with 44 819 points. The ML model is trained on sets of different sizes and then used to predict the energies for tens of thousands of nuclear configurations within seconds. The resulting datasets are utilized in variational calculations of the vibrational energy levels of CH3Cl. By using both structure-based sampling and self-correction, the size of the training set can be kept small (e.g., 10% of the points) without any significant loss of accuracy. In ab initio rovibrational spectroscopy, it is thus possible to reduce the number of computationally costly electronic structure calculations through structure-based sampling and self-correcting KRR-based machine learning by up to 90%.

Hierarchical machine learning of potential energy surfaces

Abstract: We present hierarchical machine learning (hML) of highly accurate potential energy surfaces (PESs). Our scheme is based on adding predictions of multiple Δ-machine learning models trained on energies and energy corrections calculated with a hierarchy of quantum chemical methods. Our (semi-)automatic procedure determines the optimal training set size and composition of each constituent machine learning model, simultaneously minimizing the computational effort necessary to achieve the required accuracy of the hML PES. Machine learning models are built using kernel ridge regression, and training points are selected with structure-based sampling. As an illustrative example, hML is applied to a high-level ab initio CH3Cl PES and is shown to significantly reduce the computational cost of generating the PES by a factor of 100 while retaining similar levels of accuracy (errors of ∼1 cm−1).

Accurate ab initio vibrational energies of methyl chloride

Abstract: Two new nine-dimensional potential energy surfaces (PESs) have been generated using high-level ab initio theory for the two main isotopologues of methyl chloride, CH335Cl and CH337Cl. The respective PESs, CBS-35 HL, and CBS-37 HL, are based on explicitly correlated coupled cluster calculations with extrapolation to the complete basis set (CBS) limit, and incorporate a range of higher-level (HL) additive energy corrections to account for core-valence electron correlation, higher-order coupled cluster terms, scalar relativistic effects, and diagonal Born-Oppenheimer corrections. Variational calculations of the vibrational energy levels were performed using the computer program TROVE, whose functionality has been extended to handle molecules of the form XY 3Z. Fully converged energies were obtained by means of a complete vibrational basis set extrapolation. The CBS-35 HL and CBS-37 HL PESs reproduce the fundamental term values with root-mean-square errors of 0.75 and 1.00 cm−1, respectively. An analysis of t...

Quantum-Chemical Insights from Deep Tensor Neural Networks

Abstract: Learning from data has led to paradigm shifts in a multitude of disciplines, including web, text and image search, speech recognition, as well as bioinformatics. Can machine learning enable similar breakthroughs in understanding quantum many-body systems? Here we develop an efficient deep learning approach that enables spatially and chemically resolved insights into quantum-mechanical observables of molecular systems. We unify concepts from many-body Hamiltonians with purpose-designed deep tensor neural networks, which leads to size-extensive and uniformly accurate (1 kcal mol−1) predictions in compositional and configurational chemical space for molecules of intermediate size. As an example of chemical relevance, the model reveals a classification of aromatic rings with respect to their stability. Further applications of our model for predicting atomic energies and local chemical potentials in molecules, reliable isomer energies, and molecules with peculiar electronic structure demonstrate the potential of machine learning for revealing insights into complex quantum-chemical systems.

Towards exact molecular dynamics simulations with machine-learned force fields.

Abstract: Molecular dynamics (MD) simulations employing classical force fields constitute the cornerstone of contemporary atomistic modeling in chemistry, biology, and materials science. However, the predictive power of these simulations is only as good as the underlying interatomic potential. Classical potentials often fail to faithfully capture key quantum effects in molecules and materials. Here we enable the direct construction of flexible molecular force fields from high-level ab initio calculations by incorporating spatial and temporal physical symmetries into a gradient-domain machine learning (sGDML) model in an automatic data-driven way. The developed sGDML approach faithfully reproduces global force fields at quantum-chemical CCSD(T) level of accuracy and allows converged molecular dynamics simulations with fully quantized electrons and nuclei. We present MD simulations, for flexible molecules with up to a few dozen atoms and provide insights into the dynamical behavior of these molecules. Our approach provides the key missing ingredient for achieving spectroscopic accuracy in molecular simulations.

The HITRAN2020 molecular spectroscopic database

Abstract: The HITRAN database is a compilation of molecular spectroscopic parameters. It was established in the early 1970s and is used by various computer codes to predict and simulate the transmission and emission of light in gaseous media (with an emphasis on terrestrial and planetary atmospheres). The HITRAN compilation is composed of five major components: the line-by-line spectroscopic parameters required for high-resolution radiative-transfer codes, experimental infrared absorption cross-sections (for molecules where it is not yet feasible for representation in a line-by-line form), collision-induced absorption data, aerosol indices of refraction, and general tables (including partition sums) that apply globally to the data. This paper describes the contents of the 2020 quadrennial edition of HITRAN. The HITRAN2020 edition takes advantage of recent experimental and theoretical data that were meticulously validated, in particular, against laboratory and atmospheric spectra. The new edition replaces the previous HITRAN edition of 2016 (including its updates during the intervening years). All five components of HITRAN have undergone major updates. In particular, the extent of the updates in the HITRAN2020 edition range from updating a few lines of specific molecules to complete replacements of the lists, and also the introduction of additional isotopologues and new (to HITRAN) molecules: SO, CH3F, GeH4, CS2, CH3I and NF3. Many new vibrational bands were added, extending the spectral coverage and completeness of the line lists. Also, the accuracy of the parameters for major atmospheric absorbers has been increased substantially, often featuring sub-percent uncertainties. Broadening parameters associated with the ambient pressure of water vapor were introduced to HITRAN for the first time and are now available for several molecules. The HITRAN2020 edition continues to take advantage of the relational structure and efficient interface available at www.hitran.org and the HITRAN Application Programming Interface (HAPI). The functionality of both tools has been extended for the new edition.