arXiv subjects

Noam Bernstein

Publications and source records attributed to Noam Bernstein.

At least 19 recordsLinked to original sources

arXiv2026

Dilute-Limit Defect Displacements Enabled by Brillouin-Zone Sampling

Defects and their interaction with the semiconductor host lattice play an essential role in a variety of technologies. When a defect transitions between two electronic states, the lattice distorts in response. Large displacements occur near the defect, inducing displacements on neighboring atoms, and so on, producing small displacements on atoms hundreds of {\AA}ngstr\"oms from the defect. Describing these displacements accurately is challenging for first-principles calculations due to limited supercell sizes. Here we demonstrate a procedure to efficiently obtain dilute-limit atomic displacements, using information obtained in typical modest supercells. In our approach, the atomic displacements are first obtained in a small supercell and converted into a force difference under a harmonic potential energy surface. The force difference and phonon modes at different ${\bf q}$-points are then unfolded into the Born-von K\'arm\'an supercell to obtain the dilute-limit atomic displacements. We critically analyze the convergence behavior of the force difference and study possible electron density differences that give rise to those forces, arguing that modest supercells are sufficient for internal transitions and bound-exciton transitions. Two example applications of our approach are given: (1) we calculate the luminescence spectrum of the NV center in diamond and the T center in Si and (2) we obtain dilute-limit configuration coordinate diagrams for these defects. In particular, we find that coupling to acoustic phonon modes reduces the accepting-mode frequency in the configuration coordinate diagram. Our work provides the missing ingredients necessary to address truly dilute-limit transitions at defects.

cond-mat.mtrl-sci↗

arXiv2026

Fine-tuning MLIP foundation models: strategies for accuracy and transferability

Adapting machine-learned interatomic potential (MLIP) foundation models to specialised tasks through fine-tuning is an increasingly important practice, yet systematic guidance on when and how to fine-tune is currently limited. We evaluate seven fine-tuning strategies -- naive full-parameter updates, two layer-freezing variants, Low-Rank Adaptation (LoRA), multihead replay, pseudolabelled replay, and replay combined with LoRA -- across five chemically diverse benchmarks (aqueous NaCl, ice polymorphs, S$_\mathrm{N}$2 reactions, SPICE biomolecules, and lithium electrolytes), three generations of foundation models, and training sets spanning five orders of magnitude. To support this evaluation we implement three capabilities in the MACE codebase: LoRA adapted for equivariant message-passing architectures, including both scalar and equivariant linear layers; pseudolabelled replay, which decouples the replay data source from the original pretraining corpus; and model-aware atomic reference energy (E0) reestimation for fine-tuning workflows. We find that foundation model quality, correct E0 initialisation, and well-chosen hyperparameters are prerequisites whose impact routinely exceeds that of the fine-tuning strategy itself. Once these prerequisites are met, most strategies achieve strong target-task accuracy, consistently surpassing models trained from scratch. The practical distinction depends on deployment scope: naive fine-tuning offers the best convergence for single-system applications, while multihead replay -- with either original or pseudolabelled data -- is the only approach tested that consistently preserves out-of-distribution robustness, maintaining both pretraining-distribution accuracy for broader deployment and many-body short-range repulsion.

physics.chem-ph↗

arXiv2026

Eigenstate condensation in quantum systems with finite-dimensional Hilbert spaces

Random quantum states drawn from the Haar ensemble with a constraint on the energy expectation value $E_{\mathrm{av}} = \langle \psi | H | \psi\rangle$ display \textit{eigenstate condensation}: for $E_{\mathrm{av}}$ below a critical value $E_c$, they develop macroscopic overlap with the ground state. We study eigenstate condensation in systems with finite-dimensional Hilbert spaces. These systems display three phases: a ground-state phase, in which energy-constrained random states have macroscopic overlap with the ground state; a high-temperature phase, in which they have exponentially small overlap with each energy eigenstate; and an anti-ground-state phase, in which they have macroscopic overlap with the most highly excited state. In local spin systems the ground-state and anti-ground-state phases approach the middle of the spectrum as $1/[\text{system size}]$, but -- because the condensation phase transitions have exponential, rather than polynomial, finite-size scaling -- the crossover becomes exponentially sharp in system size and the high-temperature phase is best understood as an extended phase.

quant-ph↗

arXiv2025

Machine Learning Phonon Spectra for Fast and Accurate Optical Lineshapes of Defects

The optical properties of defects in solids produce rich physics, from gemstone coloration to single-photon emission for quantum networks. Essential to describing optical transitions is electron-phonon coupling, which can be predicted from first principles but requires computationally expensive evaluation of all phonon modes in simulation cells containing hundreds of atoms. We demonstrate that this bottleneck can be overcome using machine learning interatomic potentials with negligible accuracy loss. A key finding is that atomic relaxation data from routine first-principles calculations suffice as a dataset for fine-tuning, though additional data can further improve models. The efficiency of this approach enables studies of defect vibrational properties with high-level theory. We fine-tune to hybrid functional calculations to obtain highly accurate spectra, comparing with explicit calculations and experiments for various defects. Notably, we resolve fine details of local vibrational mode coupling in the luminescence spectrum of the T center in Si, a prominent quantum defect.

cond-mat.mtrl-sci↗

arXiv2024

From GAP to ACE to MACE

The Gaussian approximation potential (GAP) machine-learning-inspired functional form was the first to be used for a general-purpose interatomic potential. The atomic cluster expansion (ACE), previously the subject of a KIM Review, and its multilayer neural-network extension (MACE) have joined GAP among the methods widely used for machine-learning interatomic potentials. Here I review extensions to the original GAP formalism, as well as ACE and MACE-based frameworks that maintain the good features and mitigate the limitations of the original GAP approach.

cond-mat.mtrl-sci↗

arXiv2023

A foundation model for atomistic materials chemistry

Atomistic simulations of matter, especially those that leverage first-principles (ab initio) electronic structure theory, provide a microscopic view of the world, underpinning much of our understanding of chemistry and materials science. Over the last decade or so, machine-learned force fields have transformed atomistic modeling by enabling simulations of ab initio quality over unprecedented time and length scales. However, early ML force fields have largely been limited by: (i) the substantial computational and human effort of developing and validating potentials for each particular system of interest; and (ii) a general lack of transferability from one chemical system to the next. Here we show that it is possible to create a general-purpose atomistic ML model, trained on a public dataset of moderate size, that is capable of running stable molecular dynamics for a wide range of molecules and materials. We demonstrate the power of the MACE-MP-0 model - and its qualitative and at times quantitative accuracy - on a diverse set of problems in the physical sciences, including properties of solids, liquids, gases, chemical reactions, interfaces and even the dynamics of a small protein. The model can be applied out of the box as a starting or "foundation" model for any atomistic system of interest and, when desired, can be fine-tuned on just a handful of application-specific data points to reach ab initio accuracy. Establishing that a stable force-field model can cover almost all materials changes atomistic modeling in a fundamental way: experienced users get reliable results much faster, and beginners face a lower barrier to entry. Foundation models thus represent a step towards democratising the revolution in atomic-scale modeling that has been brought about by ML force fields.

physics.chem-ph↗

arXiv2023

ACEpotentials.jl: A Julia Implementation of the Atomic Cluster Expansion

We introduce ACEpotentials.jl, a Julia-language software package that constructs interatomic potentials from quantum mechanical reference data using the Atomic Cluster Expansion (Drautz, 2019). As the latter provides a complete description of atomic environments, including invariance to overall translation and rotation as well as permutation of like atoms, the resulting potentials are systematically improvable and data efficient. Furthermore, the descriptor's expressiveness enables use of a linear model, facilitating rapid evaluation and straightforward application of Bayesian techniques for active learning. We summarize the capabilities of ACEpotentials.jl and demonstrate its strengths (simplicity, interpretability, robustness, performance) on a selection of prototypical atomistic modelling workflows.

physics.comp-ph↗

arXiv2023

wfl Python Toolkit for Creating Machine Learning Interatomic Potentials and Related Atomistic Simulation Workflows

Predictive atomistic simulations are increasingly employed for data intensive high throughput studies that take advantage of constantly growing computational resources. To handle the sheer number of individual calculations that are needed in such studies, workflow management packages for atomistic simulations have been developed for a rapidly growing user base. These packages are predominantly designed to handle computationally heavy ab initio calculations, usually with a focus on data provenance and reproducibility. However, in related simulation communities, e.g. the developers of machine learning interatomic potentials (MLIPs), the computational requirements are somewhat different: the types, sizes, and numbers of computational tasks are more diverse, and therefore require additional ways of parallelization and local or remote execution for optimal efficiency. In this work, we present the atomistic simulation and MLIP fitting workflow management package wfl and Python remote execution package ExPyRe to meet these requirements. With wfl and ExPyRe, versatile Atomic Simulation Environment based workflows that perform diverse procedures can be written. This capability is based on a low-level developer-oriented framework, which can be utilized to construct high level functionality for user-friendly programs. Such high level capabilities to automate machine learning interatomic potential fitting procedures are already incorporated in wfl, which we use to showcase its capabilities in this work. We believe that wfl fills an important niche in several growing simulation communities and will aid the development of efficient custom computational tasks.

physics.comp-ph↗

arXiv2022

Nested sampling for physical scientists

We review Skilling's nested sampling (NS) algorithm for Bayesian inference and more broadly multi-dimensional integration. After recapitulating the principles of NS, we survey developments in implementing efficient NS algorithms in practice in high-dimensions, including methods for sampling from the so-called constrained prior. We outline the ways in which NS may be applied and describe the application of NS in three scientific fields in which the algorithm has proved to be useful: cosmology, gravitational-wave astronomy, and materials science. We close by making recommendations for best practice when using NS and by summarizing potential limitations and optimizations of NS.

stat.CO↗

arXiv2019

Structural transitions in dense disordered silicon from quantum-accurate ultra-large-scale simulations

Structurally disordered materials continue to pose fundamental questions, including that of how different disordered phases ("polyamorphs") can coexist and transform from one to another. As a widely studied case, amorphous silicon (a-Si) forms a fourfold-coordinated, covalent random network at ambient conditions, but much higher-coordinated, metallic-like phases under pressure. However, a detailed mechanistic understanding of the liquid-amorphous and amorphous-amorphous transitions in silicon has been lacking, due to intrinsic limitations of even the most advanced experimental and computational techniques. Here, we show how machine-learning (ML)-driven simulations can break through this long-standing barrier, affording a comprehensive, quantum-accurate, and fully atomistic description of all relevant liquid and amorphous phases of silicon. Combining a model system size of 100,000 atoms (ten-nanometre length scale) with a prediction accuracy of a few meV per atom, our simulations reveal a remarkable, three-step transformation sequence for a-Si under increasing external pressure. First, up to 10-11 GPa, polyamorphic low- and high-density amorphous (LDA and HDA) regions are found to coexist, rather than appearing sequentially. Then, we observe a structural collapse into a distinct, very-high-density amorphous (VHDA) phase at 12-13 GPa, reminiscent of the dense liquid but being formed at a much lower temperature. Finally, our simulations indicate the transient nature of this VHDA phase: it rapidly nucleates crystallites at 13-16 GPa, ultimately leading to the formation of a poly-crystalline, simple-hexagonal structure, consistent with experiments but not seen in earlier simulations.

cond-mat.mtrl-sci↗

arXiv2019

Impact of biaxial and uniaxial strain on V$_2$O$_3$

Using first-principles calculations we determine the role of compressive and tensile uniaxial and equibiaxial strain on the structural, electronic and magnetic properties of V$_2$O$_3$. We find that compressive strain increases the energy cost to transition from the high-temperature paramagnetic metallic phase to the low-temperature antiferromagnetic insulating phase. This shift in the energy difference can be explained by changes in the V-V bond lengths that are antiferromagnetically aligned in the low temperature structure. The insights that we have obtained provide a microscopic explanation for the shifts in the metal-insulator transition temperature that have been observed in experiments of V$_2$O$_3$ films grown on different substrates.

cond-mat.mtrl-sci↗

arXiv2019

Machine-learned Interatomic Potentials for Alloys and Alloy Phase Diagrams

We introduce machine-learned potentials for Ag-Pd to describe the energy of alloy configurations over a wide range of compositions. We compare two different approaches. Moment tensor potentials (MTP) are polynomial-like functions of interatomic distances and angles. The Gaussian Approximation Potential (GAP) framework uses kernel regression, and we use the Smooth Overlap of Atomic Positions (SOAP) representation of atomic neighbourhoods that consists of a complete set of rotational and permutational invariants provided by the power spectrum of the spherical Fourier transform of the neighbour density. Both types of potentials give excellent accuracy for a wide range of compositions and rival the accuracy of cluster expansion, a benchmark for this system. While both models are able to describe small deformations away from the lattice positions, SOAP-GAP excels at transferability as shown by sensible transformation paths between configurations, and MTP allows, due to its lower computational cost, the calculation of compositional phase diagrams. Given the fact that both methods perform as well as cluster expansion would but yield off-lattice models, we expect them to open new avenues in computational materials modeling for alloys.

cond-mat.mtrl-sci↗

arXiv2019

De novo exploration and self-guided learning of potential-energy surfaces

Interatomic potential models based on machine learning (ML) are rapidly developing as tools for materials simulations. However, because of their flexibility, they require large fitting databases that are normally created with substantial manual selection and tuning of reference configurations. Here, we show that ML potentials can be built in a largely automated fashion, exploring and fitting potential-energy surfaces from the beginning (de novo) within one and the same protocol. The key enabling step is the use of a configuration-averaged kernel metric that allows one to select the few most relevant structures at each step. The resulting potentials are accurate and robust for the wide range of configurations that occur during structure searching, despite only requiring a relatively small number of single-point DFT calculations on small unit cells. We apply the method to materials with diverse chemical nature and coordination environments, marking a milestone toward the more routine application of ML potentials in physics, chemistry, and materials science.

cond-mat.mtrl-sci↗

arXiv2019

The role of defects in the metal-insulator transition in VO$_{2}$ and V$_2$O$_3$

The vanadates VO$_2$ and V$_2$O$_3$ are prototypical examples of strongly correlated materials that exhibit a metal-insulator transition. While the phase transitions in these materials have been studied extensively, there is a limited understanding of how the properties of these materials are affected by the presence of defects and doping. In this study we investigate the impact of native point defects in the form of Frenkel defects on the structural, magnetic and electronic properties of VO$_2$ and V$_2$O$_3$, using first-principles calculations. In VO$_2$ the vanadium Frenkel pairs lead to a non-trivial insulating state. The unpaired vanadium interstitial bonds to a single dimer, which leads to a trimer that has one singlet state and one localized single-electron $S=1/2$ state. The unpaired broken dimer created by the vanadium vacancy also has a localized $S=1/2$ state. Thus, the insulating state is created by the singlet dimers, the trimer and the two localized $S=1/2$ states. Oxygen Frenkel pairs, on the other hand, lead to a metallic state in VO$_2$, but are expected to be present in much lower concentrations. In contrast, the Frenkel defects in V$_2$O$_3$ do not directly suppress the insulating character of the material. However, the disorder created by defects in V$_2$O$_3$ alters the local magnetic moments and in turn reduces the energy cost of a transition between the insulating and conducting phases of the material. We also find self-trapped small polarons in V$_2$O$_3$, which has implications for transport properties in the insulating phase.

cond-mat.mtrl-sci↗

arXiv2018

Quantifying Chemical Structure and Atomic Energies in Amorphous Silicon Networks

Amorphous materials are coming within reach of realistic computer simulations, but new approaches are needed to fully understand their intricate atomic structures. Here, we show how machine-learning (ML)-based techniques can give new, quantitative chemical insight into the atomic-scale structure of amorphous silicon (a-Si). Based on a similarity function ("kernel"), we define a structural metric that unifies the description of nearest- and next-nearest-neighbor environments in the amorphous state. We apply this to an ensemble of a-Si networks, generated in melt-quench simulations with an ML-based interatomic potential, in which we tailor the degree of ordering by varying the quench rates down to $10^{10}$ K/s (leading to a structural model that is lower in energy than the established WWW network). We then show how "machine-learned" atomic energies permit a chemical interpretation, associating coordination defects in a-Si with distinct energetic stability regions. The approach is straightforward and inexpensive to apply to arbitrary structural models, and it is therefore expected to have more general significance for developing a quantitative understanding of the amorphous state.

cond-mat.mtrl-sci↗

arXiv2018

Uncovering the mechanism of the impurity-selective Mott transition in paramagnetic V$_{2}$O$_{3}$

While the phase diagrams of the one- and multi-orbital Hubbard model have been well studied, the physics of real Mott insulators is often much richer, material dependent, and poorly understood. In the prototype Mott insulator V$_{2}$O$_{3}$, chemical pressure was initially believed to explain why the paramagnetic-metal to antiferromagnetic-insulator transition temperature is lowered by Ti doping while Cr doping strengthens correlations, eventually rendering the high-temperature phase paramagnetic insulating. However, this scenario has been recently shown both experimentally and theoretically to be untenable. Based on full structural optimization, we demonstrate via the charge self-consistent combination of density functional theory and dynamical mean-field theory that changes in the V$_{2}$O$_{3}$ phase diagram are driven by defect-induced local symmetry breakings resulting from dramatically different couplings of Cr and Ti dopants to the host system. This finding emphasizes the high sensitivity of the Mott metal-insulator transition to the local environment and the importance of accurately accounting for the one-electron Hamiltonian, since correlations crucially respond to it.

cond-mat.str-el↗

arXiv2018

Machine learning a general purpose interatomic potential for silicon

The success of first principles electronic structure calculation for predictive modeling in chemistry, solid state physics, and materials science is constrained by the limitations on simulated length and time scales due to computational cost and its scaling. Techniques based on machine learning ideas for interpolating the Born-Oppenheimer potential energy surface without explicitly describing electrons have recently shown great promise, but accurately and efficiently fitting the physically relevant space of configurations has remained a challenging goal. Here we present a Gaussian Approximation Potential for silicon that achieves this milestone, accurately reproducing density functional theory reference results for a wide range of observable properties, including crystal, liquid, and amorphous bulk phases, as well as point, line, and plane defects. We demonstrate that this new potential enables calculations that would be extremely expensive with a first principles electronic structure method, such as finite temperature phase boundary lines, self-diffusivity in the liquid, formation of the amorphous by slow quench, and dynamic brittle fracture. We show that the uncertainty quantification inherent to the Gaussian process regression framework gives a qualitative estimate of the potential's accuracy for a given atomic configuration. The success of this model shows that it is indeed possible to create a useful machine-learning-based interatomic potential that comprehensively describes a material, and serves as a template for the development of such models in the future.

cond-mat.mtrl-sci↗

arXiv2018

Realistic atomistic structure of amorphous silicon from machine-learning-driven molecular dynamics

Amorphous silicon (a-Si) is a widely studied non-crystalline material, and yet the subtle details of its atomistic structure are still unclear. Here, we show that accurate structural models of a-Si can be obtained by harnessing the power of machine-learning algorithms to create interatomic potentials. Our best a-Si network is obtained by cooling from the melt in molecular-dynamics simulations, at a rate of 10$^{11}$ K/s (that is, on the 10 ns timescale). This structure shows a defect concentration of below 2% and agrees with experiments regarding excess energies, diffraction data, as well as $^{29}$Si solid-state NMR chemical shifts. We show that this level of quality is impossible to achieve with faster quench simulations. We then generate a 4,096-atom system which correctly reproduces the magnitude of the first sharp diffraction peak (FSDP) in the structure factor, achieving the closest agreement with experiments to date. Our study demonstrates the broader impact of machine-learning interatomic potentials for elucidating accurate structures and properties of amorphous functional materials.

cond-mat.mtrl-sci↗