Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.20779

Enabling large language model agents for high-fidelity many-body perturbation theory calculations

Benran Zhang·Byeongchan Lee·Zien Zhu·Chih-En Hsu·Zhenfa Zheng·Hung-Chung Hsueh·Young Woo Choi·Mauro Del Ben·Zhenglu Li
View original source ↗Read source document ↗Export metadata ↓

Abstract

Many-body perturbation theory (MBPT) methods, including the GW approximation, the GW plus Bethe-Salpeter equation (GW-BSE) method, and GW perturbation theory (GWPT), among others, provide accurate first-principles descriptions of excited-state properties and multiparticle interactions of materials. However, their use remains limited due to complex multi-software workflows, interdependent physical parameters, and substantial theoretical expertise. Large language model (LLM) agents, which couple a language model to software tools, show potential in automating computational tasks, but MBPT remains an unaddressed challenge. We show that LLM agents tested in our work, including frontier models such as Opus 4.7 and GPT-5.5, do not reliably construct MBPT workflows in one attempt, suggesting that correct setup requires expert knowledge and consistency that current models lack. Across tested model-harness combinations, out-of-the-box agents average 54%, 36%, and 19% on GW, GW-BSE, and GWPT workflow benchmarks, with strong model dependence and large run-to-run variation. Errors typically originate from cross-stage inconsistencies that surface only several stages after they are introduced. We therefore develop AI-Donkey, an agentic framework that combines expert-curated knowledge with deterministic execution to enable high-fidelity GW, GW-BSE, and GWPT calculations and analyses. With AI-Donkey, all tested agents, from frontier proprietary models to open-weight models, produce complete, runnable MBPT workflows with no observed run-to-run variation. We further demonstrate autonomous calculations and post-processing on 1D, 2D, and bulk materials, including the many-body enhancement of Fröhlich electron-phonon coupling in polar semiconductors. Our approach provides a general framework for knowledge engineering to enable reliable agentic workflows in computational materials science.

Explore related subjects

cond-mat.mtrl-sci

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Benran Zhang, Byeongchan Lee, Zien Zhu, Chih-En Hsu, Zhenfa Zheng, Hung-Chung Hsueh, Young Woo Choi, Mauro Del Ben, Zhenglu Li. 2026-10-14. Enabling large language model agents for high-fidelity many-body perturbation theory calculations. https://arxiv.org/abs/2610.20779

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Impact of electron--spin coupling on exchange coupling parameters: a nonperturbative approach

Exchange coupling parameters $J_{ij}$ in the Heisenberg model are crucial for describing magnetic behavior at the atomic level. In magnetic materials, spin fluctuations can be accompanied by a self-consistent electronic response, i.e., electron--spin coupling in the sense of electronic feedback to finite spin rotations. However, the quantitative importance of this coupling in extracting reliable $J_{ij}$ has not been fully clarified. Here, using fully self-consistent, nonperturbative evaluations, we show that finite-angle spin rotations induce such electronic feedback and quantify how strongly it renormalizes the extracted $J_{ij}$. We examine systems of both fundamental and practical interest, including perovskite SrMnO$_3$, Nd-based permanent-magnet compounds (Nd$_2$Fe$_{14}$B and Nd$_2$Co$_{14}$B), and elemental 3$d$ transition metals. The nonperturbative approach yields exchange couplings that remain consistent over a wide range of rotation angles. Moreover, when the sampled configurations represent the magnetic disorder near the transition, spin models parameterized in this way give reasonable agreement with experimental magnetic phase-transition temperatures. Overall, our results demonstrate the quantitative impact of electron--spin coupling on the exchange coupling parameters $J_{ij}$ and provide a nonperturbative route to effective $J_{ij}$ for a specified regime of magnetic disorder.

cond-mat.mtrl-sci↗
arXiv2026

Benchmarking and advancing machine-learning interatomic potentials for 3C-SiC across equilibrium and extreme conditions

Several machine-learning interatomic potentials (MLIPs) have been developed for cubic silicon carbide (3C-SiC). Here, we benchmark seven existing MLIPs against a common set of structural, thermal, defect, radiation-damage, disordered-phase, and computational-performance tests. We also introduce two complementary MLIPs trained on DFT-labeled crystalline, liquid, amorphous, and defective configurations. TGAP, a Gaussian approximation potential based on the \verb|soap_turbo| descriptor, provides consistently high accuracy across the sampled configurational space. MTP, a custom moment tensor potential, retains good accuracy across these regimes while offering substantially lower computational cost, GPU acceleration, and access to multi-million-atom simulations. As an application-level test, both potentials remain stable in displacement cascades spanning short-range repulsion, transient disorder, and recrystallization. Overall, MTP emerges as a particularly strong accuracy--cost compromise for large-scale and strongly non-equilibrium simulations of 3C-SiC.

cond-mat.mtrl-sci↗
arXiv2026

Loss of altermagnetic order and smooth restoration of Kramers' spin degeneracy with increasing temperature in CrSb and MnTe

We describe how thermally induced spin fluctuations modify the electronic structures of two prototypical altermagnets, CrSb and MnTe, via application of the disordered local moment picture. For both materials, our self-consistent, ab initio calculations demonstrate that local magnetic moments persist on Cr and Mn atoms in their paramagnetic states, necessitating a spin-polarised description of the electronic structure even above the Néel temperature, $T_\mathrm{N}$. Moreover, Kramers' spin degeneracy, which is broken for both materials in their altermagnetic ground states, is shown to be smoothly restored - on the average - as the local moments thermally disorder. In metallic CrSb, this occurs at temperatures well below $T_\mathrm{N}$ and the signature effects of its altermagnetism are lost as the magnetic disorder induces heavy smearing of strongly dispersive electronic states around the Fermi energy. By contrast, in semiconducting MnTe, with its band gap largely unaffected by magnetic disorder, the spin degeneracy only returns at temperatures close to and above $T_\mathrm{N}$. We quantify the temperature dependence of the altermagnetic order parameter and the underlying electronic structures of both materials, with implications for their spin transport and spectroscopic properties.

cond-mat.mtrl-sci↗