arXiv subjects

Jinu Jeong

Publications and source records attributed to Jinu Jeong.

3 recordsLinked to original sources

arXiv2026

DiffGLE: Differentiable Coarse-Grained Dynamics using Generalized Langevin Equation

Capturing dynamical fidelity at the coarse-grained scale remains a central challenge in systematic molecular modelling. The generalized Langevin equation, rooted in the Mori--Zwanzig formalism, provides a principled framework for representing the memory friction and stochastic forces induced by eliminated microscopic degrees of freedom. In practice, however, parameterising its memory kernel is difficult because the exact kernel is a history-dependent projected-dynamics object coupled by the fluctuation--dissipation theorem to coloured random forces that are not directly observable from ordinary trajectories. Here we combine differentiable simulation with a coloured-noise ansatz to learn non-Markovian memory kernels in a top-down manner. The random force is represented by a trainable filter whose autocorrelation defines the friction memory, enforcing fluctuation--dissipation consistency by construction. The filter is optimised by backpropagating through coarse-grained generalized Langevin trajectories to match reference velocity autocorrelation functions, avoiding explicit projected-force reconstruction. We demonstrate the approach on bulk water, bulk carbon dioxide, and a single particle star-polymer memory benchmark. Across all these systems, the proposed framework accurately reproduces the target dynamical correlations, demonstrating that differentiable simulation enables direct optimisation of non-Markovian memory kernels from time-correlation observables.

cond-mat.soft↗

arXiv2025

Generative Quasi-Continuum Modeling of Confined Fluids at the Nanoscale

We present a data-efficient, multiscale framework for predicting the density profiles of confined fluids at the nanoscale. While accurate density estimates require prohibitively long timescales that are inaccessible by ab initio molecular dynamics (AIMD) simulations, machine-learned molecular dynamics (MLMD) offers a scalable alternative, enabling the generation of force predictions at ab initio accuracy with reduced computational cost. However, despite their efficiency, MLMD simulations remain constrained by femtosecond timesteps, which limit their practicality for computing long-time averages needed for accurate density estimation. To address this, we propose a conditional denoising diffusion probabilistic model (DDPM) based quasi-continuum approach that predicts the long-time behavior of force profiles along the confinement direction, conditioned on noisy forces extracted from a limited AIMD dataset. The predicted smooth forces are then linked to continuum theory via the Nernst-Planck equation to reveal the underlying density behavior. We test the framework on water confined between two graphene nanoscale slits and demonstrate that density profiles for channel widths outside of the training domain can be recovered with ab initio accuracy. Compared to AIMD and MLMD simulations, our method achieves orders-of-magnitude speed-up in runtime and requires significantly less training data than prior works.

physics.comp-ph↗

arXiv2024

Water Isotope Separation using Deep Learning and a Catalytically Active Ultrathin Membrane

Water isotope separation, specifically separating heavy from light water, is a socially significant issue due to the usage of heavy water in applications such as nuclear magnetic resonance, nuclear power, and spectroscopy. Separation of heavy water from light water is difficult due to very similar physical and chemical properties between the isotopes. We show that a catalytically active ultrathin membrane (e.g., a nanopore in MoS2) can enable chemical exchange processes and physicochemical mechanisms that lead to efficient separation of deuterium from hydrogen, quantified as the D2O and deuterium separation ratio of 4.5 and 1.73, respectively. The separation process is inherently multiscale in nature with the shorter times representing chemical exchange processes and the longer timescales representing the transport phenomena. To bridge the timescales, we employ a deep learning methodology which uses short time scale ab-initio molecular dynamics data for training and extends the timescales to classical molecular dynamics regime to demonstrate isotope separation and reveal the underlying complex physicochemical processes.

physics.chem-ph↗