arXiv subjects

Yunzhe Jia

Publications and source records attributed to Yunzhe Jia.

3 recordsLinked to original sources

arXiv2026

Accelerating dynamic simulations of photoexcited materials and their evolution by electron-informed machine learning

Nonadiabatic coupled electron-nuclear dynamics upon electronic excitation underpin the microscopic mechanism and rational modulation of diverse photoinduced functional phenomena in materials, yet their direct first-principles simulations remain computationally demanding. Here we develop a framework for nonadiabatic excited-state machine-learning molecular dynamics (EMLMD) simulations, where the nonequilibrium electronic information upon photoexcitation such as electron temperature is rigorously calibrated from high-precision real-time time-dependent density functional theory (rt-TDDFT) benchmark simulations, enabling accurate reconstruction of excited-state potential energy surfaces (PES). This framework natively incorporates the excited-state electron-phonon couplings and intrinsically captures photoinduced phonon anharmonicity, both of which are missing in standard machine learning molecular dynamics, thus delivering first-principles-level accuracy for excited-state atomic evolutions. Large-scale EMLMD simulations resolve time- and momentum-resolved phonon dynamics in photoexcited materials, directly uncovering the competition between photogenerated coherent phonons and thermal phonons during photoinduced phase transition of bismuth. It also simultaneously resolves elusive atomic-scale microscopic dynamics and global structural rearrangement for selenium photoamorphization. Balancing high accuracy and efficiency, EMLMD offers a versatile paradigm to tackle key challenges in the study of complex excited-state molecular dynamics.

cond-mat.mtrl-sci↗

arXiv2026

Scalable photoexcitation-induced molecular dynamics with machine-learned Hamiltonians

Ultrafast photoexcitation offers a controllable route to steer structural dynamics in solids, yet predicting how nonequilibrium electronic excitation drives lattice motion across extended spatial and temporal scales remains a major computational challenge. Here we introduce time-dependent ab-initio propagation with electronic machine learning (TDAP-eML), a framework that explicitly incorporates electronic evolution into scalable simulations of photoexcitation-induced lattice dynamics. By integrating machine-learned electronic structure with atomistic propagation, TDAP-eML describes how photoexcitation reshapes the evolving energy landscapes and forces governing structural motion. Across representative examples including silicon and FeSe, the framework reproduces key photoexcited lattice responses obtained from first-principles time-dependent density functional theory calculations and captures coherent phonon dynamics together with their dependence on excitation conditions. Its computational advantage increases with system size, reaching nearly three orders of magnitude reduction in computational cost for the large systems examined. TDAP-eML thus establishes a scalable framework for coupled electronic and lattice evolution, linking nonequilibrium excitation to photoinduced forces, predictive structural dynamics, and experimentally accessible observables.

cond-mat.mtrl-sci↗

arXiv2022

A simple but strong baseline for online continual learning: Repeated Augmented Rehearsal

Online continual learning (OCL) aims to train neural networks incrementally from a non-stationary data stream with a single pass through data. Rehearsal-based methods attempt to approximate the observed input distributions over time with a small memory and revisit them later to avoid forgetting. Despite its strong empirical performance, rehearsal methods still suffer from a poor approximation of the loss landscape of past data with memory samples. This paper revisits the rehearsal dynamics in online settings. We provide theoretical insights on the inherent memory overfitting risk from the viewpoint of biased and dynamic empirical risk minimization, and examine the merits and limits of repeated rehearsal. Inspired by our analysis, a simple and intuitive baseline, Repeated Augmented Rehearsal (RAR), is designed to address the underfitting-overfitting dilemma of online rehearsal. Surprisingly, across four rather different OCL benchmarks, this simple baseline outperforms vanilla rehearsal by 9%-17% and also significantly improves state-of-the-art rehearsal-based methods MIR, ASER, and SCR. We also demonstrate that RAR successfully achieves an accurate approximation of the loss landscape of past data and high-loss ridge aversion in its learning trajectory. Extensive ablation studies are conducted to study the interplay between repeated and augmented rehearsal and reinforcement learning (RL) is applied to dynamically adjust the hyperparameters of RAR to balance the stability-plasticity trade-off online. Code is available at https://github.com/YaqianZhang/RepeatedAugmentedRehearsal

cs.LG↗