arXiv subjects

Jarvist Moore Frost

Publications and source records attributed to Jarvist Moore Frost.

14 recordsLinked to original sources

arXiv2026

Spindrift: Learning quantum degeneracy from thermal purity in restricted path integral Monte Carlo

Restricted path integral Monte Carlo (RPIMC) sidesteps the fluctuating Fermion sign problem by confining paths within nodal regions of a trial density matrix, thereby recovering polynomial scaling. However, this nodal surface must be provided from elsewhere; unless it is exact, it introduces a fixed-node energy error. Here we introduce \textsc{Spindrift}, a Variational Density Matrix approach that learns the many-body Fermionic density matrix from a regularised Bloch residual, evaluated on samples drawn by a restricted Worm algorithm. Motivated by the `purity' of quantum statistical mechanics at high temperature (where kinetic energy dominates), we train the density matrix along an imaginary-time (descending temperature) curriculum from an exact infinite-temperature heat-kernel starting point, learning the condensation of quantum correlations as temperature drops, through successive corrections to the previous reference. We parametrise our model with a permutation-equivariant continuous normalising flow to generate quasi-particle backflow trajectories, modulated by a symmetric Jastrow factor. This architecture guarantees exact Fermionic antisymmetry and spatial symmetry throughout training. Simulating $N=3$ interacting Fermions in a two-dimensional harmonic trap, we demonstrate stable curriculum training. The learnt velocity field smoothly deforms the nodal structure away from the free-particle reference. Open-Worm G-sector trapping provides a natural diagnostic for nodal accuracy. \textsc{Spindrift} systematically lowers the restricted energy relative to the free-particle reference across all temperatures and successfully reproduces the benchmark energy at $β=1$, establishing a stable, physics-informed framework for finite-temperature quantum Monte Carlo where the nodal structure is learnt self-consistently.

cond-mat.stat-mech↗

arXiv2026

Hydrodynamic Backflow for Easing the Fermion Sign in Finite-Temperature Electron Path Integral Simulations

Some notable technology systems, such as high-temperature superconductors and materials for controlled nuclear fusion, require an accurate description of finite-temperature quantum matter. Stochastic path integral methods are finite-temperature and numerically exact, but scale poorly with system size due the notorious Fermion sign problem. To somewhat mitigate this, we use a hydrodynamical backflow coordinate transformation. Our first approach was a continuous normalizing flow machine learning optimisation. We found this to roughly halve the statistical uncertainty at medium sign severity. Numerical issues challenged training effectively. Thus, a semi-analytic analogue was developed to estimate the optimal parameters. We do this by using a derived expression dependent on a Bosonic observable. Hence, the calculation of these values does not have a sign problem. The resulting backflow transformations reduce the problem by multiple orders of magnitude in the specific case of a harmonically trapped, two-dimensional, electron gas at finite-temperature. The total energy of the system agrees with previous, backflow untransformed, studies and we calculate energies for up to 32 electrons. The limiting factor is found to be, primarily, the $O(N^3)$ calculation of the Jacobian, stemming from the coordinate transformation of the backflow. A more thorough implementation may further improve this scaling. Even without this, a route for simulating electron systems at currently unreachable regimes is obtained.

cond-mat.str-el↗

arXiv2026

Learning the Fermion sign structure in path-integral Monte Carlo

Starting from a \emph{probabilistic numerics} approach to the Fermion sign problem in path integral Monte Carlo, we recast the arithmetic calculation of a Fermionic observable as a statistical inference problem. We develop approaches that learn the behaviour of Fermion exchange cycles binned by the conjugacy class of the permutation group (which we term `permutation family'). This extends the work of DuBois, Brown and Alder\cite{dubois2017overcoming} to inhomogeneous and more complex systems. Monte Carlo samples are used to train models for both the probability of a permutation family and the energy of this set of exchange permutations. The overall Fermionic energy is then directly inferred from these models, without using a direct ratio estimator on the Monte Carlo samples. By imposing physical understanding as inductive priors, we produce accurate and useful fits that remain robust even in regimes with severe sign problems. We generalise the linear (ideal-gas style) models of DuBois et al. with Bayesian priors that enforce the intuitive models of Feynman\cite{Feynman1953A} at their asymptotic limits. These linear models serve as the baseline for a Long Short-Term Memory (LSTM) neural network, which is tasked with learning only the residual many-body \emph{correlations} on top of the physical model. We develop active important sampling methods driven by these models, which direct the Monte Carlo chains toward undersampled permutation regions, to efficiently reduce the variance in the observable. We apply this framework to small experiments on benchmark systems: the spin-polarised uniform electron gas, and electrons in a 2D harmonic confining potential. In both cases we demonstrate that this inference-based framework can extract stable energies in regimes where direct Monte Carlo sampling fails due to the sign problem.

cond-mat.stat-mech↗

arXiv2026

Six Open Questions in Machine-Learned Interatomic Potential Foundation Models

Machine-learned interatomic potentials (MLIPs) have had a profound impact on molecular modelling in recent years, promising to resolve the long-standing tension between the scale and accuracy of simulations. There has been a proliferation of new models and designs, and recently the paradigm of ``foundational'' MLIPs has become prevalent. Broadly speaking, foundation models are trained on large diverse datasets and promise to work well for new systems with minimal updates required. However, in such a new and fast moving field, there are many unanswered questions. In this article, we set out to articulate and explore what we see as the most important among these questions. We start by developing a working definition for foundational MLIPs and use this definition to frame the subsequent open questions. Despite the rapid progress in the field of MLIP models, we believe that these are fundamental questions which will continue to define cutting edge research in MLIPs in the years to come.

cond-mat.mtrl-sci↗

arXiv2026

Quantum dynamic simulation of triplet formation in an effective model of Y6 (BTP-4F)

We construct a five-state model for photoexcitation in Y6 (BTP-4F) dimers, and then solve the non-adiabtic dynamics using the Hierarchical Equations of Motion (HEOM) method. We find that triplets are populated mainly via a transiently excited \textit{intermolecular} charge-transfer singlet to triplet Frenkel exciton route; this route is not available to the monomer. Analysis of one-particle transition density matrices suggests that the charge-transfer states are spatially distinct to the Frenkel exciton states, indicating that the large spin-orbit-coupling for this transition is due to it being permitted by an associated change in orbital character. Aggregation in Y6 therefore directly enables fast and high-yield intersystem crossing. We selenise our model dimers, significantly enhancing spin-orbit-coupling, which then accelerates this charge-transfer mediated route. Looking forwards to simulations on larger aggregates, we show that, though Marcus theory gives qualitatively correct dynamics, the long-time yields are incorrect due to it missing quantum recurrences. Instead, we show that the recently developed memory-kernel projector\cite{Gestsson2025-ez} method can produce semi-classical rates directly from the HEOM equations which lead to quantitatively correct dynamics and yields.

cond-mat.mtrl-sci↗

arXiv2026

Correcting hybrid density functionals to model Y6 and other non-fullerene acceptors

Recently developed fused-ring electron-acceptors such as Y6 (BTP-4F) have strong oscillator strength, good charge-carrier transport and a small bandgap. They therefore have enormous current technical application to organic optoelectronics, such as solar cells. To design new materials, it would be useful to predict the electronic structure accurately. Due to the large number of atoms involved in representative aggregates of these materials, we need an efficient electronic structure method. Standard density functional theory poorly describes charge-transfer states, and were typically parameterised for vacuum calculations of individual molecules. In this work we tune a range-separated hybrid functional for Y6, and characterise representative dimers extracted from the solid-state. We demonstrate that the extensive solvatochromic effects of Y6 are due, in part, to oscillator strength borrowing between the charge-transfer and Frenkel excitons. We provide an explanation for the short optimally-tuned range-separation parameter, based on the Penn model for the frequency dependent dielectric of a semiconductor. We caution that non-tuned range-separated hybrids are less accurate than global hybrids for these, and similar, materials. We show how reducing the range-separation length improves the accuracy of standard range-separation functionals, without an involved tuning process.

cond-mat.mtrl-sci↗

arXiv2025

Machine learning intermolecular transfer integrals with compact atomic cluster representations

Calculating intermolecular charge transfer integrals in organic semiconductors requires substantial computer resource for each individual calculation. We might alternatively construct a machine learning model for transfer integrals, which model the full six-degrees of freedom for the relative position of dimer pairs, trained on representative calculations for the molecules of interest. Recent developments have produced effective machine learning force fields, which model the total energy of atomic assemblies. We extend the Atomic Cluster Expansion (ACE) with the correct symmetries for transfer (kinetic-energy) integrals. Combined with a spherical harmonic basis makes, this forms a strong inductive bias and makes for a data efficient model. We introduce coarse-grained and heavy-atom representations, and assess the methodology on representative conjugated semiconductors: ethylene, thiophene, and naphthalene.

cond-mat.dis-nn↗

arXiv2024

Predicting polaron mobility in organic semiconductors with the Feynman variational approach

We extend the Feynman variational method applied to the parabolic-band Fröhlich (continuum) large polaron~\cite{Feynman1955} to a Holstein (lattice) small polaron, with a parabolic-band. This new theory shows a discrete localisation as a function of coupling strength. Having build the theory with the same quasi-particle Lagrangian as the 1955 work, we can directly use the FHIP~\cite{Feynman1962} response theory to calculate DC mobility and complex conductivity. We show that we can take matrix elements from electronic structure calculations on real materials, by modelling charge-carrier mobility in crystalline Rubrene. Good agreement is found to measurement, with a predicted mobility of $μ= 47.72$~\si{cm^2 V^{-1} s^{-1}} at $300$~\si{K}.

cond-mat.mtrl-sci↗

arXiv2024

Anharmonic phonons with Gaussian processes

We provide a method for calculating anharmonic lattice dynamics, by building a surrogate model based on Gaussian Processes (GPs). Due to the underlying Gaussian form of a GP, the model is infinitely differentiable. This allows us to train the model trained directly on forces (the derivative of PESs) reducing the evaluations required for a given accuracy. We can extend this differentiation to directly calculate second and third order force-constants using automatic differentiation (AD). For the five model materials we study, we find that the force-constants are in close agreement with a standard finite-displacement approach. Our method appears to be linear scaling in the number of atoms at predicting both second and third-order (anharmonic) force-constants.

cond-mat.mtrl-sci↗

arXiv2022

Multiple phonon modes in Feynman path-integral variational polaron mobility

The Feynman path-integral variational approach to the polaron problem\cite{Feynman1955}, along with the associated FHIP linear-response mobility theory\cite{Feynman1962}, provides a computationally amenable method to predict the frequency-resolved temperature-dependent charge-carrier mobility, and other experimental observables in polar semiconductors. We show that the FHIP mobility theory predicts non-Drude transport behaviour, and shows remarkably good agreement with the recent diagrammatic Monte-Carlo mobility simulations of Mishchenko et al.\cite{Mishchenko2019} for the abstract Fröhlich Hamiltonian. We extend this method to multiple phonon modes in the Fröhlich model action. This enables a slightly better variational solution, as inferred from the resulting energy. We carry forward this extra complexity into the mobility theory, where it shows richer structure in the frequency and temperature dependent mobility, due to the different phonon modes activating at different energies. The method provides a computationally efficient and fully quantitative method of predicting polaron mobility and response in real materials.

cond-mat.mtrl-sci↗

arXiv2019

Sparse hierarchical representation learning on molecular graphs

Architectures for sparse hierarchical representation learning have recently been proposed for graph-structured data, but so far assume the absence of edge features in the graph. We close this gap and propose a method to pool graphs with edge features, inspired by the hierarchical nature of chemistry. In particular, we introduce two types of pooling layers compatible with an edge-feature graph-convolutional architecture and investigate their performance for molecules relevant to drug discovery on a set of two classification and two regression benchmark datasets of MoleculeNet. We find that our models significantly outperform previous benchmarks on three of the datasets and reach state-of-the-art results on the fourth benchmark, with pooling improving performance for three out of four tasks, keeping performance stable on the fourth task, and generally speeding up the training process.

cs.LG↗

arXiv2018

Review: Solid-state physics of halide perovskites

Halide perovskite solar cells presented a unique opportunity to apply modern computational materials science techniques to an (initially) poorly understood new material. In this review, we recount the key understanding developed during the last five years, through a narrative review of research progress. The central enigma of the material is how it can be so defective, and yet work so well as a photovoltaic. The physical properties of the material were understood through molecular and lattice dynamic calculations, revealing the material to show large dynamic responses on a wide range of time scales. Longer length scales in the material was simulated with effective classical potentials, showing that complex domains can be generated by the interacting molecular dipoles, generating structured features in the electrostatic potential of the lattice. Relativistic electronic structure reveals unique features in the bands, which may explain observed slow recombination, and could be used in high efficiency photovoltaics. The large dielectric response of the lattice leads to a strong drive for the formation of polarons, some device physics of which are discussed. These polarons offer a possible explanation for the observed slow cooling of photoexcitations in the material.

cond-mat.mes-hall↗

arXiv2017

Calculating polaron mobility in halide perovskites

Lead halide perovskite semiconductors are soft, polar, materials. The strong driving force for polaron formation (the dielectric electron-phonon coupling) is balanced by the light band effective-masses, leading to a strongly-interacting large-polaron. A first-principles prediction of mobility would help understand the fundamental mobility limits. Theories of mobility need to consider the polaron (rather than free-carrier) state due to the strong interactions. In this material we expect that at room temperature polar-optical phonon mode scattering will dominate, and so limit mobility. We calculate the temperature-dependent polaron mobility of hybrid halide perovskites by variationally solving the Feynman polaron model with the finite-temperature free-energies of Ōsaka. This model considers a simplified effective-mass band-structure interacting with a continuum dielectric of characteristic response frequency. We parametrise the model fully from electronic-structure calculations. In methylammonium lead iodide at 300 K we predict electron and hole mobilities of 133 and 94 cm^2/V/s respectively. These are in acceptable agreement with single-crystal measurements, suggesting that the intrinsic limit of the polaron charge carrier state has been reached. Repercussions for hot-electron photo-excited states are discussed. As well as mobility, the model also exposes the dynamic structure of the polaron. This can be used to interpret impedance measurements of the charge-carrier state. We provide the phonon-drag mass-renormalisation, and scattering time constants. These could be used as parameters for larger-scale device models and band-structure dependent mobility simulations.

cond-mat.mtrl-sci↗

arXiv2017

Slow cooling of hot polarons in halide perovskite solar cells

Halide perovskites show unusual thermalisation kinetics for above bandgap photo-excitation. We explain this as a consequence of excess energy being deposited into discrete large polaron states. The cross-over between low-fluence and high-fluence `phonon bottleneck' cooling is due to a Mott transition where the polarons overlap ($n \ge 10^{18}/\mathrm{cm}^3$) and the phonon sub-populations are shared. We calculate the initial rate of cooling (thermalisation) from the scattering time in the Fröhlich polaron model to be 78 meVps$^{-1}$ for $\mathrm{CH}_3\mathrm{NH}_3\mathrm{PbI}_3$. This rapid initial thermalisation involves heat transfer into optical phonon modes coupled by a polar dielectric interaction. Further cooling to equilibrium over hundreds of picoseconds is limited by the ultra-low thermal conductivity of the perovskite lattice.

cond-mat.mtrl-sci↗