The paper introduces a method to construct phase‑field models directly from ab initio data by projecting molecular dynamics onto species‑density fields using the Mori‑Zwanzig formalism. Neural networks parameterize the resulting non‑local free energy and mobility, trained on short MD trajectories generated with machine‑learning interatomic potentials. Demonstrations on an iron‑boron melt and hydrogen‑helium mixtures show the approach can predict pressure‑dependent stability, immiscibility boundaries, and large‑scale droplet dynamics beyond conventional atomistic simulations.
By Mengyi Chen, Peichen Zhong, Zihan Zhang, Qianxiao Li
The paper presents a neural operator that learns the Kohn–Sham map, directly predicting electron density from the Kohn–Sham potential without orbital diagonalization. Using a domain‑invariant SE(3)‑equivariant Fourier neural operator trained on 8,504 molecules and solids, the model achieves quasi‑linear scaling self‑consistent field (SCF) convergence across diverse systems—including organic molecules, insulators, and metals—while reproducing Kohn–Sham DFT accuracy for densities, spectra, and structural observables. This enables large‑scale simulations, such as magnesium dislocation densities with 82,500 valence electrons, on a single GPU.
By Danish Khan, Maurice D. Hanisch, Nikolai Argatoff, Evan Xie, Sandeep Sharma, Anima Anandkumar
arXiv:2609.39090v1 Announce Type: cross
Abstract: Data-driven machine learning (ML) techniques have become an essential tool in many domains of science. Their application to atomistic simulations of...
By J\"org Behler, Michele Ceriotti, Cecilia Clementi, G\'abor Cs\'anyi, Alin-Marin Elena, Aditi Krishnapriyan, Joseph W. Abbott, Fabio Affinito, Albert P. Bart\'ok, Ilyes Batatia, Filippo Bigi, Florian N. Br\"unig, Yannick Calvino Alonso, Giuseppe Carleo, Aur\'elie Champagne, Stefan Chmiela, Marc L. Descoteaux, Ralf Drautz, Alexandra Farcas, Meng Gao, Rohit Goswami, Michael F. Herbst, Christian Holm, James R. Kermode, Alexander L. M. Knoll, Tobias Kreiman, Hoang-Thien Luu, Yury Lysogorskiy, Mihai-Cosmin Marinica, Rocco Meli, Klaus-Robert M\"uller, Frank No\'e, Mohamadhosein Nosratjoo, Simon Olsson, Christoph Ortner, Aldo S. Pasos-Trejo, Anyang Peng, Eric Qu, Andrea Rizzi, Mariana Rossi, Bassem Sboui, Gregor N. C. Simm, Alexandre Tkatchenko, Jacopo Venturin, O. Anatole von Lilienfeld, William C. Witt, Brandon M. Wood, Tigany Zarrouk, Fabian Zills
Data-driven machine learning (ML) techniques have become an essential tool in many domains of science. Their application to atomistic simulations of matter is particularly widespread and impactful. Th...
The study uses large‑scale, quantum‑accurate reactive simulations powered by active learning and deep equivariant neural network potentials to investigate the coupled reactions and diffusion at the interface of a symmetric solid‑state battery cell. Unsupervised clustering of local atomic environments reveals a previously unreported crystalline disordered phase, Li₂S₀.₇₂P₀.₁₄Cl₀.₁₄, in the solid‑electrolyte interphase (SEI), and explains experimental observations of SEI formation and lithium creep mechanisms that drive dendrite initiation. The work demonstrates a parameter‑free digital twin capable of providing atomistic insights into complex heterogeneous processes in solid‑state electrochemistry.
By Jingxuan Ding, Laura Zichi, Matteo Carli, Menghang Wang, Albert Musaelian, Yu Xie, Boris Kozinsky
The paper investigates whether machine learning models can uncover physical patterns in atomistic data without relying on traditional physics-based inductive biases such as geometric locality or graph structures. By training a general-purpose architecture on molecular simulation data, the authors demonstrate that the model autonomously learns interatomic interaction strengths resembling classical electrostatics and identifies interaction cutoffs aligned with established physical models. The study also reports predictable neural scaling behavior and competitive accuracy on certain metrics compared to physics-informed architectures, suggesting that explicit priors may only be necessary when empirically justified.
By Tobias Kreiman, Yutong Bai, Fadi Atieh, Elizabeth Weaver, Eric Qu, Aditi S. Krishnapriyan
arXiv:2607. 19198v1 Announce Type: cross Abstract: Amorphous materials exhibit exceptional mechanical and functional properties, yet their rugged energy landscapes are notoriously difficult to sample.
By Mouyang Cheng, Denis Blessing, Botao Yu, Gerhard Neumann, Mingda Li, Carles Domingo-Enrich, Yuanqi Du
arXiv:2606. 14498v1 Announce Type: cross Abstract: Predicting the Kohn-Sham Hamiltonian with machine learning can accelerate density functional theory while retaining access to molecular orbitals, energy levels, and electronic-structure observables that energy-only surrogates cannot resolve.
By Yunhong Lou, Xihang Yue, Xinran Wei, Tianqi Deng, Linchao Zhu
The paper introduces a neural‑network based local‑box chemical kinetics solver for exoplanet atmospheres, employing a residual flow‑map architecture. It achieves microsecond‑scale inference with percent‑level accuracy across a wide range of temperatures, pressures, time steps, and compositional variations, outperforming other machine‑learning models and handling the extreme stiffness of atmospheric chemistry. The surrogate model offers a flexible, efficient alternative to classical solvers for state‑to‑state flow‑map problems in numerical simulations.
By Isaac Malsky, Xi Zhang, Tiffany Kataria, Matthew Graham, Ziyu Huang, Boris Bonev, Shang-Min Tsai, Elspeth K. H. Lee
arXiv:2607. 29158v1 Announce Type: cross Abstract: We introduce implicit machine learning force fields (I-MLFFs), which replace explicit stacks of neural network layers with self-consistent fixed-point equations.
By Johannes Mae{\ss}, Leon Werner, J. Thorben Frank, Winfried Ripken, Martin Michajlow, Joshua Futterer, Klaus-Robert M\"uller, Stefan Chmiela
arXiv:2507. 03853v2 Announce Type: replace Abstract: We introduce OrbitAll, a geometry- and physics-informed deep learning framework that encodes any molecular system with arbitrary charges, spins, and environmental effects using electronic structure information.
By Beom Seok Kang, Vignesh C. Bhethanabotla, Amin Tavakoli, Maurice D. Hanisch, Arimitsu Horikawa-Strakovsky, Miguel Nouman, Danish Khan, William A. Goddard III, Anima Anandkumar
arXiv:2606. 04100v1 Announce Type: new Abstract: Machine learning interatomic potentials (MLIPs) enable efficient and accurate atomistic simulations but depend critically on the quality and diversity of the training data.
By Joanna Zou, Fraser Birks, Dallas Foster, Youssef Marzouk