arXiv subjects

Amrita Goswami

Publications and source records attributed to Amrita Goswami.

10 recordsLinked to original sources

arXiv2026

Typed Component Algebras for Simulated Annealing and Markov-Chain Monte Carlo

Simulated annealing (SA) and fixed-temperature Markov-chain Monte Carlo (MCMC) run the same Metropolis-Hastings kernel over a tempered objective, but the variants appear as separate monolithic drivers, so improving one ingredient requires rewriting and re-verifying a whole solver. The shared kernel becomes a typed algebra of five components (objective, cooling schedule, neighborhood, move kernel, and acceptance rule) whose four local composition laws the construction checks; a single Sampler step then runs any point of the algebra. A surrogate proposal, a fitted generalized-Langevin thermostat, a quasi-Monte Carlo polish, or a noise-aware acceptance rule is implemented once and becomes available to every classical, fast, generalized, Hamiltonian, or parallel-tempered driver that shares the interface. The same typing carries the correctness artifacts: SymPy-checked reductions of Generalized SA to its Boltzmann, fast, and Metropolis limits (the reductions surfaced a sign error that had stood in the visiting-distribution literature for three decades); a TLA+ specification model-checked for four safety and two liveness properties; and a three-channel finite-precision audit showing that fixing one channel of the acceptance path does not let float16 reproduce float64 basin selection. The implementation is the open-source Rust-and-Python package anneal, with an Array-API/DLPack device boundary and a portfolio optimizer whose only argument is a budget. On the CUTEst collection under a shared work-unit budget it reaches the best observed basin on more problems than a budget-matched CMA-ES restart heuristic, while carrying the almost-sure convergence and regret guarantees that heuristic lacks. Every reported number and figure regenerates from the reproducibility package with its pinned environment.

cs.SE↗

arXiv2026

ChemFit: A framework for automated high-dimensional model parameter optimization

The parameterization of simulation-based models is a central yet laborious task in computational chemistry and physics, often driven by human intuition and manual iteration. Automating this task necessitates the definition of suitable objective functions, which tend to be expensive to evaluate, noisy, non-differentiable, or composed of heterogeneous contributions originating from separate sets of simulations. Gradient-free and black-box optimization algorithms are powerful tools which are particularly well-suited to minimizing such objective functions. Here, we introduce ChemFit, a flexible Python framework for the definition, composition, and massively concurrent evaluation of simulation-based objective functions, which is designed to operate in conjunction with these algorithms. We demonstrate the broad applicability of this approach by using ChemFit for three representative examples of increasing complexity and real-world relevance. First, we obtain the parameters of the Lennard-Jones potential for liquid argon from experimental measurements of the density. Second, we parameterize a polarizable and flexible potential energy function to reproduce the structure of small H$_2$O clusters obtained from density functional theory calculations. Finally, we tune a small subset of the parameters of a residue-level coarse-grained protein force-field, with the goal to reproduce the experimental critical solution temperature of the low complexity domain of the wild-type hnRNPA1 sequence and an arginine-enriched mutant of this protein. hnRNPA1 is an RNA-binding protein linked to amyotrophic lateral sclerosis. Together, these examples illustrate how ChemFit enables scalable, reproducible, and optimizer-agnostic parameter fitting for broadly applicable multiscale models.

physics.chem-ph↗

arXiv2026

Viscosity as a Smoking Gun for Complex Formation in Solution: Fe$^{2+}$ and Mg$^{2+}$ Chlorides as Examples

Electrolyte solutions at high concentration are indispensable and yet poorly understood. In particular, the extent of speciation -- the formation of complexes composed of multiple species -- in concentrated ionic solutions is very challenging to obtain theoretically and experimentally, but can have a strong effect on solution properties. The literature is rife with contradictory estimates of speciation from experiments. We find that speciation affects transport properties, and is therefore, a prerequisite to accurately model concentrated solutions. We turn this to our advantage by showing that the viscosity can be used to determine the extent of complexation in concentrated aqueous solutions. Results of simulations as well as experimental measurements are presented. The atomistic Madrid-2019 force-field is extended to model FeCl$_2$. Solutions of FeCl$_2$ and MgCl$_2$ are compared and the observed difference in viscosity explained by more complexation in the former, a conclusion supported by recently reported X-ray absorption and neutron scattering experiments.

physics.chem-ph↗

arXiv2025

Flowy: High performance probabilistic lava emplacement prediction

Lava emplacement is a complex physical phenomenon, affected by several factors. These include, but are not limited to features of the terrain, the lava settling process, the effusion rate or total erupted volume, and the probability of effusion from different locations. One method, which has been successfully employed to predict lava flow emplacement and forecast the inundated area and final lava thickness, is the MrLavaLoba method from Vitturi et al. The MrLavaLoba method has been implemented in their code of the same name. Here, we introduce Flowy, a new computational tool that implements the MrLavaLoba method in a more efficient manner. New fast algorithms have been incorporated for all performance critical code paths, resulting in a complete overhaul of the implementation. When compared to the MrLavaLoba code, Flowy exhibits a significant reduction in runtime -- between 100 to 400 times faster -- depending on the specific input parameters. The accuracy and the probabilistic convergence of the model outputs are not compromised, maintaining high fidelity in generating possible lava flow paths and deposition characteristics. We have validated Flowy's performance and reliability through comprehensive unit-testing and a real-world eruption scenario. The source code is freely available on GitHub, facilitating transparency, reproducibility and collaboration within the geoscientific community.

physics.geo-ph↗

arXiv2024

Evidence of sharp transitions between octahedral and capped trigonal prism states of the solvation shell of Fe$^{+3}$(aq)

The structure of the solvation shell of aqueous Fe$^{+3}$ ion has been a subject of controversy due to discrepancies between experiments and different levels of theory. We address this issue by performing simulations for a wide range of ion concentrations, using various empirical potential energy functions, as well as density functional theory calculations of selected configurations. The solvation shell undergoes abrupt transitions between two states: an octahedral (OH) state with 6-fold coordination, and a capped trigonal prism (CTP) state with 7-fold coordination. The lifetime of these states is concentration dependent. In dilute $\mathrm{FeCl_3}$ solutions, the lifetime of the two states is similar ($\approx 1$ ns). However, the lifetime of the OH state increases with ion concentration, while that of the CTP state decreases slightly. When a uniform negative background charge is used instead of explicit counterions, the lifetime of the OH state is greatly overestimated. These findings underscore the need for further experimental measurements as well as high-level simulations over sufficiently long timescales and low concentration.

physics.chem-ph↗

arXiv2021

Homogeneous Nucleation of Sheared Liquids: Advances and Insights from Simulations and Theory

One of the most ubiquitous and technologically important phenomena in nature is the nucleation of homogeneous flowing systems. The microscopic effects of shear on a nucleating system are still imperfectly understood, although in recent years a consistent picture has emerged. The opposing effects of shear can be split into two major contributions for simple liquids: increase of the energetic cost of nucleation, and enhancement of the kinetics. In this perspective, we describe the latest computational and theoretical techniques which have been developed over the past two decades. We collate and unify the overarching influences of shear, temperature, and supersaturation on the process of homogeneous nucleation. Experimental techniques and capabilities are discussed, against the backdrop of results from simulations and theory. Although we primarily focus on simple liquids, we also touch upon the sheared nucleation of more complex systems, including glasses and polymer melts. We speculate on the promising directions and possible advances that could come to fruition in the future.

cond-mat.soft↗

arXiv2021

Universal Nucleation Behaviour of Sheared Systems

Using molecular simulations and a modified Classical Nucleation Theory, we study the nucleation, under flow, of a variety of liquids: different water models, Lennard-Jones and hard sphere colloids. Our approach enables us to analyze a wide range of shear rates inaccessible to brute-force simulations. Our results reveal that the variation of the nucleation rate with shear is universal. A simplified version of the theory successfully captures the non-monotonic temperature dependence of the nucleation behavior, which is shown to originate from the violation of the Stokes-Einstein relation.

cond-mat.soft↗

arXiv2020

Seeding Method for Ice Nucleation under Shear

Hydrodynamic flow can have complex and far-reaching consequences on the rate of homogenous nucleation. We present a general formalism for calculating the nucleation rates of simply sheared systems. We have derived an extension to the conventional Classical Nucleation Theory, explicitly embodying the shear rate. Seeded Molecular Dynamics simulations form the backbone of our approach. The framework can be used for moderate supercoolings, at which temperatures brute-force methods are practically infeasible. The competing energetic and kinetic effects of shear arise naturally from the equations. We show how the theory can be used to identify shear regimes of ice nucleation behaviour for the mW water model, unifying disparate trends reported in the literature. At each temperature, we define a crossover shear rate in the limit of $1000-10,000 \ s^{-1}$, beyond which the nucleation rate increases steadily upto a maximum, at the optimal shear rate. For $235$, $240$, $255$ and $260 \ K$, the optimal shear rates are in the range of $\approx 10^6-10^7 \ s^{-1}$. For very high shear rates beyond $10^8 \ s^{-1}$, nucleation is strongly inhibited. Our results indicate that the shear-dependent nucleation rate curves have a non-monotonic dependence on temperature.

physics.comp-ph↗

arXiv2020

d-SEAMS: Deferred Structural Elucidation Analysis for Molecular Simulations

Structural analyses are an integral part of computational research on nucleation and supercooled water, whose accuracy and efficiency can impact the validity and feasibility of such studies. The underlying molecular mechanisms of these often elusive and computationally expensive processes can be inferred from the evolution of ice-like structures, determined using appropriate structural analysis techniques. We present d-SEAMS, a free and open-source post-processing engine for the analysis of molecular dynamics trajectories, which is specifically able to qualitatively classify ice structures, in both strong confinement and bulk systems. For the first time, recent algorithms for confined ice structure determination have been implemented, along with topological network criteria for bulk ice structure determination. Recognizing the need for customization in structural analysis, d-SEAMS has a unique code architecture, built with `nix`, employing a `YAML`-`Lua` scripting pipeline. The software has been designed to be user-friendly and easy to extend. The engine outputs are compatible with popular graphics software suites, allowing for immediate visual insights into the systems studied. We demonstrate the features of d-SEAMS by using it to analyze nucleation in the bulk regime and for quasi-one and quasi-two-dimensional systems. Structural time evolution and quantitative metrics are determined for heterogenous ice nucleation on a silver-exposed $β$-AgI surface, homogenous ice nucleation, flat monolayer square ice formation and freezing of an ice nanotube.

physics.comp-ph↗

arXiv2019

A General Topological Network Criterion for Exploring the Structure of Icy Nanoribbons and Monolayers

We develop intuitive metrics for quantifying complex nucleating systems under confinement. These are shown to arise naturally from the analysis of the topological ring network, and are amenable for use as order parameters for such systems. Drawing inspiration from qualitative visual inspection, we introduce a general topological criterion for elucidating the ordered structures of confined water, using a graph theoretic approach. Our criterion is based on primitive rings, and reinterprets the hydrogen-bond-network in terms of these primitives. This approach has no a priori assumptions, except the hydrogen bond definition, and may be used as an exploratory tool for the automated discovery of new ordered phases. We demonstrate the versatility of our criterion by applying it to analyse well-known monolayer ices. Our methodology is then extended to identify the building blocks of one-dimensional $n$-sided prismatic nanoribbon ices.

physics.chem-ph↗