arXiv subjects

Nicholas Lubbers

Publications and source records attributed to Nicholas Lubbers.

45 records · Page 3Linked to original sources

arXiv2019

Machine Learned Hückel Theory: Interfacing Physics and Deep Neural Networks

The Hückel Hamiltonian is an incredibly simple tight-binding model famed for its ability to capture qualitative physics phenomena arising from electron interactions in molecules and materials. Part of its simplicity arises from using only two types of empirically fit physics-motivated parameters: the first describes the orbital energies on each atom and the second describes electronic interactions and bonding between atoms. By replacing these traditionally static parameters with dynamically predicted values, we vastly increase the accuracy of the extended Hückel model. The dynamic values are generated with a deep neural network, which is trained to reproduce orbital energies and densities derived from density functional theory. The resulting model retains interpretability while the deep neural network parameterization is smooth, accurate, and reproduces insightful features of the original static parameterization. Finally, we demonstrate that the Hückel model, and not the deep neural network, is responsible for capturing intricate orbital interactions in two molecular case studies. Overall, this work shows the promise of utilizing machine learning to formulate simple, accurate, and dynamically parameterized physics models.

cond-mat.dis-nn↗

arXiv2019

Machine learning for molecular dynamics with strongly correlated electrons

We use machine learning to enable large-scale molecular dynamics (MD) of a correlated electron model under the Gutzwiller approximation scheme. This model exhibits a Mott transition as a function of on-site Coulomb repulsion $U$. The repeated solution of the Gutzwiller self-consistency equations would be prohibitively expensive for large-scale MD simulations. We show that machine learning models of the Gutzwiller potential energy can be remarkably accurate. The models, which are trained with $N=33$ atoms, enable highly accurate MD simulations at much larger scales ($N\gtrsim10^{3}$). We investigate the physics of the smooth Mott crossover in the fluid phase.

cond-mat.str-el↗

arXiv2018

Inferring low-dimensional microstructure representations using convolutional neural networks

We apply recent advances in machine learning and computer vision to a central problem in materials informatics: The statistical representation of microstructural images. We use activations in a pre-trained convolutional neural network to provide a high-dimensional characterization of a set of synthetic microstructural images. Next, we use manifold learning to obtain a low-dimensional embedding of this statistical characterization. We show that the low-dimensional embedding extracts the parameters used to generate the images. According to a variety of metrics, the convolutional neural network method yields dramatically better embeddings than the analogous method derived from two-point correlations alone.

physics.comp-ph↗

arXiv2018

Earthquake catalog-based machine learning identification of laboratory fault states and the effects of magnitude of completeness

Machine learning regression can predict macroscopic fault properties such as shear stress, friction, and time to failure using continuous records of fault zone acoustic emissions. Here we show that a similar approach is successful using event catalogs derived from the continuous data. Our methods are applicable to catalogs of arbitrary scale and magnitude of completeness. We investigate how machine learning regression from an event catalog of laboratory earthquakes performs as a function of the catalog magnitude of completeness. We find that strong model performance requires a sufficiently low magnitude of completeness, and below this magnitude of completeness model performance saturates.

physics.geo-ph↗

arXiv2018

Less is more: sampling chemical space with active learning

The development of accurate and transferable machine learning (ML) potentials for predicting molecular energetics is a challenging task. The process of data generation to train such ML potentials is a task neither well understood nor researched in detail. In this work, we present a fully automated approach for the generation of datasets with the intent of training universal ML potentials. It is based on the concept of active learning (AL) via Query by Committee (QBC), which uses the disagreement between an ensemble of ML potentials to infer the reliability of the ensemble's prediction. QBC allows the presented AL algorithm to automatically sample regions of chemical space where the ML potential fails to accurately predict the potential energy. AL improves the overall fitness of ANAKIN-ME (ANI) deep learning potentials in rigorous test cases by mitigating human biases in deciding what new training data to use. AL also reduces the training set size to a fraction of the data required when using naive random sampling techniques. To provide validation of our AL approach we develop the COMP6 benchmark (publicly available on GitHub), which contains a diverse set of organic molecules. Through the AL process, it is shown that the AL-based potentials perform as well as the ANI-1 potential on COMP6 with only 10% of the data, and vastly outperforms ANI-1 with 25% the amount of data. Finally, we show that our proposed AL technique develops a universal ANI potential (ANI-1x) that provides accurate energy and force predictions on the entire COMP6 benchmark. This universal ML potential achieves a level of accuracy on par with the best ML potentials for single molecule or materials, while remaining applicable to the general class of organic molecules comprised of the elements CHNO.

physics.comp-ph↗

arXiv2017

Hierarchical modeling of molecular energies using a deep neural network

We introduce the Hierarchically Interacting Particle Neural Network (HIP-NN) to model molecular properties from datasets of quantum calculations. Inspired by a many-body expansion, HIP-NN decomposes properties, such as energy, as a sum over hierarchical terms. These terms are generated from a neural network--a composition of many nonlinear transformations--acting on a representation of the molecule. HIP-NN achieves state-of-the-art performance on a dataset of 131k ground state organic molecules, and predicts energies with 0.26 kcal/mol mean absolute error. With minimal tuning, our model is also competitive on a dataset of molecular dynamics trajectories. In addition to enabling accurate energy predictions, the hierarchical structure of HIP-NN helps to identify regions of model uncertainty.

stat.ML↗

arXiv2017

Machine Learning Predicts Laboratory Earthquakes

Forecasting fault failure is a fundamental but elusive goal in earthquake science. Here we show that by listening to the acoustic signal emitted by a laboratory fault, machine learning can predict the time remaining before it fails with great accuracy. These predictions are based solely on the instantaneous physical characteristics of the acoustical signal, and do not make use of its history. Surprisingly, machine learning identifies a signal emitted from the fault zone previously thought to be low-amplitude noise that enables failure forecasting throughout the laboratory quake cycle. We hypothesize that applying this approach to continuous seismic data may lead to significant advances in identifying currently unknown signals, in providing new insights into fault physics, and in placing bounds on fault failure times.

physics.geo-ph↗

arXiv2012

Model Independent Direct Detection Analyses

Following the construction of the general effective theory for dark matter direct detection in 1203.3542, we perform an analysis of the experimental constraints on the full parameter space of elastically scattering dark matter. We review the prescription for calculating event rates in the general effective theory and discuss the sensitivity of various experiments to additional nuclear responses beyond the spin-independent (SI) and spin-dependent (SD) couplings: an angular-momentum-dependent (LD) and spin-and-angular-momentum-dependent (LSD) response, as well as a distinction between transverse and longitudinal spin-dependent responses. We consider the effect of interference between different operators and in particular look at directions in parameter space where such cancellations lead to holes in the sensitivity of individual experiments. We explore the complementarity of different experiments by looking at the improvement of bounds when experiments are combined. Finally, our scan through parameter space shows that within the assumptions on models and on the experiments' sensitivity that we make, no elastically scattering dark matter explanation of DAMA is consistent with all other experiments at 90%, though we find points in parameter space that are ruled out only by about a factor of 2 in the cross-section.

hep-ph↗

arXiv2012

The Effective Field Theory of Dark Matter Direct Detection

We extend and explore the general non-relativistic effective theory of dark matter (DM) direct detection. We describe the basic non-relativistic building blocks of operators and discuss their symmetry properties, writing down all Galilean-invariant operators up to quadratic order in momentum transfer arising from exchange of particles of spin 1 or less. Any DM particle theory can be translated into the coefficients of an effective operator and any effective operator can be simply related to most general description of the nuclear response. We find several operators which lead to novel nuclear responses. These responses differ significantly from the standard minimal WIMP cases in their relative coupling strengths to various elements, changing how the results from different experiments should be compared against each other. Response functions are evaluated for common DM targets - F, Na, Ge, I, and Xe - using standard shell model techniques. We point out that each of the nuclear responses is familiar from past studies of semi-leptonic electroweak interactions, and thus potentially testable in weak interaction studies. We provide tables of the full set of required matrix elements at finite momentum transfer for a range of common elements, making a careful and fully model-independent analysis possible. Finally, we discuss embedding non-relativistic effective theory operators into UV models of dark matter.

hep-ph↗