arXiv subjects

Yuxiang Gao

Publications and source records attributed to Yuxiang Gao.

At least 19 recordsLinked to original sources

arXiv2026

A Strategy Toward Room Temperature Topological Hall Effect via Local Moment Magnetism

Topological spin textures in local moment systems hold great promise for technological applications due to their large magnetic moments, strong spin-orbit coupling (SOC), and high tunability. Finding new spin textures that are stable near room temperature is paramount to maximizing their potential for applications. Here, we provide a strategy for realizing topological spin textures at high temperatures by identifying rare earth ($R$) magnets ordering at or near room temperature. We demonstrate the feasibility of this strategy in one of these magnets, hexagonal Gd$_5$Pb$_3$, which orders at $T_C$ = 285 K. The indication for topological spin textures comes from topological Hall effect (THE), which, in Gd$_5$Pb$_3$, occurs between $T$ = 100 - 200 K, an order of magnitude higher temperature than in other reported $R$-based systems. Our results present an opportunity to explore the role of SOC, anisotropic exchange, geometric frustration, and magnetic interactions in stabilizing topological spin textures, and provide a pathway toward realizing them near room temperature in $R$-based magnets.

cond-mat.mtrl-sci↗

arXiv2026

4D-WAM: Infusing Spatiotemporal Awareness into World Action Models through Trajectory Fields

Building on recent advances in world models, World Action Models (WAMs) jointly model video prediction and action generation. However, they typically represent videos in 2D pixel space, creating a representation gap with 3D space in which robotic actions are executed. Recent 3D approaches introduce 3D information, but fail to fully exploit the dynamics of 3D structures. In this work, we propose 4D-WAM, a model-agnostic training strategy that injects spatiotemporal knowledge from 3D trajectory fields into WAMs through representation alignment. To this end, we introduce two complementary objectives: 1) motion alignment, which aligns temporal feature variations across adjacent frames and encourages the model to build local 4D awareness during training, and 2) destination alignment, which guides the model to infer the final destination from the source frame by minimizing the gap between their attention-like similarity distributions. Together, these objectives provide both local motion supervision and long-horizon goal guidance, enabling WAMs to learn trajectory-level spatiotemporal representations. Extensive in-distribution and out-of-distribution experiments across different base models demonstrate the model's improvements in spatial understanding, execution precision, robustness, generalization, and versatility.

cs.RO↗

arXiv2026

Is Forward Prediction Enough? Physical State Grounding for JEPA World Models

Learning structured and control-relevant latent representations remains a key challenge for world models. Recent JEPA-based world models learn action-conditioned predictive latent dynamics from observation sequences. However, their forward-prediction objectives do not explicitly enforce reliable identifiability of robot-centric physical state from individual latents or state changes from latent pairs, which can limit downstream planning and policy performance. We propose PSG-JEPA, a physically grounded JEPA world model that shapes its latent space with two complementary grounding objectives beyond forward prediction: grounding individual latents in robot proprioceptive state, and grounding latent pairs in multi-horizon joint-angle changes. Both objectives are applied only during training, leaving the inference architecture and computational cost unchanged. To comprehensively evaluate PSG-JEPA, we conduct experiments at three levels: (1) latent identifiability via probing, (2) goal-conditioned planning on frozen latents, and (3) policy learning in simulation and on a real robot. Experiments demonstrate that our PSG-JEPA consistently outperforms state-of-the-art latent world-model baselines at all three levels.

cs.RO↗

arXiv2026

DFM-VLA: Iterative Action Refinement for Robot Manipulation via Discrete Flow Matching

Vision-Language-Action (VLA) models that encode actions using a discrete tokenization scheme have been widely adopted for robotic manipulation, but existing decoding paradigms remain fundamentally limited. Whether actions are decoded sequentially by autoregressive VLAs or in parallel by discrete diffusion VLAs, once a token is generated, it is typically fixed and cannot be revised in subsequent iterations. Consequently, early token errors cannot be effectively corrected later. We propose DFM-VLA, a discrete flow matching VLA that iteratively refines action tokens. DFM-VLA models a token-level probability velocity field that dynamically updates the full action sequence across refinement iterations. We investigate two approaches to constructing the velocity field: an auxiliary velocity-head formulation and an embedding-guided formulation. To further improve prediction accuracy, we introduce a metric-aligned action tokenizer (MAAT) tailored to the coarse-to-fine nature of DFM, together with a two-stage decoding strategy. Extensive experiments on CALVIN, LIBERO, LIBERO-Plus, and real-world manipulation tasks demonstrate the effectiveness of our approach. Our project is available at https://chris1220313648.github.io/DFM-VLA/.

cs.RO↗

arXiv2026

DyPES-VLA: Learning Shared Dynamics Priors and Embodiment-Specific Control for Cross-Embodiment Manipulation

Vision-Language-Action (VLA) models have become a powerful paradigm for robot manipulation, but training a single generalist policy for heterogeneous robot embodiments remains an open problem. Existing methods have two main limitations. First, they underuse dynamics priors shared across diverse visual and interaction data, limiting cross-embodiment transfer. Second, they require extensive manual preprocessing to convert embodiment-specific actions into a common format. To overcome these limitations, we propose DyPES-VLA, a cross-embodiment VLA that learns shared Dynamics Priors and Embodiment-Specific control. First, we learn shared dynamics priors by training the vision-language model (VLM) with a future-prediction objective on cross-embodiment data, driving the shared query representation to capture object motion, contact, and interaction-induced scene changes. Second, an embodiment-specific Mixture-of-Experts (MoE) action head translates these shared dynamics priors into executable controls directly in each embodiment's native action space, without manually pre-aligning heterogeneous actions into a common format. This head shares attention layers to capture common temporal action structures, while its embodiment-specific feed-forward experts resolve the unique kinematic constraints and control semantics of distinct embodiments. As a generalist policy, our \ourmethod achieves state-of-the-art performance across simulation and real-world evaluations, reaching 98.0% success on LIBERO, 59.25% on RoboCasa-GR1, and 89.02% on RoboTwin~2.0.

cs.RO↗

arXiv2026

Data-model Coevolution as the Architectural Principle for AI-Native Materials Databases

AI-native approaches are reshaping computational materials discovery into iterative data-model coevolution cycles. However, most existing materials databases remain fundamentally data-centric, where predictive models remain external to database state and data growth is decoupled from model updating. Here we formalize data-model coevolution as the architectural basis of AI-native materials databases, where data and predictive models evolve through endogenous generation-evaluation-refinement cycles. Using the Li-P-S ternary as a demonstrative prototype, we generated approximately 70,000 candidate structures, more than 10,000 of which satisfy the stable-unique-novel (S.U.N.) criterion, achieving rapid saturation of local chemical environments together with stabilization of energy distributions. We autonomously found chemically plausible phases and motifs outside the Materials Project (MP) and Alexandria databases, including a stable Li$_2$PS$_3$ phase, the (PS$_3$)$_3^{3-}$ trimer, the (P$_3$S$_8$)$^{3-}$ ring, two isomers of the (P$_2$S$_8$)$^{2-}$ ring, and polymeric (PS$_4$)$_n^{n-}$ chains. Within two to three iterations, the integrated predictive models converged to high precision under a low first-principles cost, and the resulting data-model state can be directly queried for atomistic and electronic-structure properties within the same unified framework. Data-model states can be reused and extended across related chemical systems, enabling scalable and continuous accumulation of computational materials knowledge. These results demonstrate data-model coevolution as a practical architectural principle for AI-era materials data infrastructure.

cond-mat.mtrl-sci↗

arXiv2026

Uniaxial strain tuned magnetism of the altermagnet candidate h-FeS

Altermagnets are collinear magnetic materials with 'alter'nating local crystalline environments, characterized by joint spin and crystalline symmetries that enable ferromagnetic-like transport properties but with vanishing net magnetization. Hexagonal FeS (h-FeS) is a recently identified altermagnet candidate that shows a spontaneous anomalous Hall effect (AHE) accompanied by a tiny net magnetization. Here, we show that both the spontaneous AHE and magnetization can be effectively suppressed by an in-plane compressive strain. Since neutron diffraction measurements show that the applied uniaxial strain only modifies the in-plane domain population but does not affect the in-plane magnetic structure, the major effect of the applied strain is to tune the small $c$-axis ferromagnetic moment. Our results demonstrate a strong correlation between the tiny net magnetization and the spontaneous AHE in h-FeS, and show that uniaxial strain provides an effective knob to tune both properties in this altermagnet candidate for spintronic applications.

cond-mat.mtrl-sci↗

arXiv2026

Atomically-sharp magnetic soliton in the square-net lattice EuRhAl$_{4}$Si$_{2}$

Topological spin textures are hallmark manifestations of competing interactions in magnetic matter. Their effective description by nonlinear field theories reflects an energetic frustration that destabilizes uniform order while selecting finite-size, topologically nontrivial configurations as stationary states. Among the most extreme realizations are atomically-sharp domain wall excitations, namely one-dimensional (1D) magnetic solitons, which represent the ultimate scaling limit of magnetic textures. Such solitons may emerge in magnetic systems where effective exchange interactions compete directly with uniaxial magnetic anisotropy. Here we show that the square-net rare earth compound EuRhAl$_{4}$Si$_{2}$ realizes a very susceptible regime where the magnetic anisotropy competes with highly frustrated exchange interactions stabilizing a rare ferrimagnetic $\uparrow\uparrow\downarrow$ state that, under applied magnetic field, supports the formation of atomically-sharp soliton defects. We confirm the bulk response of the 1D magnetic solitons via magnetization and electrical transport measurements. We establish both the zero- and in-field $\uparrow\uparrow\downarrow$ order via neutron diffraction, while magnetic force microscopy visualizes its real-space evolution into a stripe-like array. To elucidate the microscopic origin of the soliton, we relate the Ruderman-Kittel-Kasuya-Yosida (RKKY)-driven exchange interactions and the magnetic anisotropy through density functional theory, and we construct an effective 1D $J_{1}$-$J_{2}$-$K$ model whose atomistic spin dynamics simulations reproduce the observed soliton states as a function of external field. Our results demonstrate that EuRhAl$_{4}$Si$_{2}$ hosts atomically-sharp, field-driven 1D magnetic solitons, providing a new platform for studying 1D topological excitations at the atomic length scale.

cond-mat.str-el↗

arXiv2025

Pre-training, fine-tuning, and distillation (PFD): Automatically generating machine learning force fields from universal models

Universal force fields generalizable across the periodic table represent a new trend in computational materials science. However, the applications of universal force fields in material simulations are limited by their slow inference speed and the lack of first-principles accuracy. Instead of building a single model simultaneously satisfying these characteristics, a strategy that quickly generates material-specific models from the universal model may be more feasible. Here, we propose a new workflow pattern, PFD (Pre-training, Fine-tuning, and Distillation), which automatically generates machine-learning force fields for specific materials from a pre-trained universal model through fine-tuning and distillation. By fine-tuning the pre-trained model, our PFD workflow generates force fields with first-principles accuracy while requiring one to two orders of magnitude less training data compared to traditional methods. The inference speed of the generated force field is further improved through distillation, meeting the requirements of large-scale molecular simulations. Comprehensive testing across diverse materials including complex systems, such as amorphous carbon, interface, etc., reveals marked enhancements in training efficiency, which suggests the PFD workflow a practical and reliable approach for force field generation in computational material sciences.

cond-mat.mtrl-sci↗

arXiv2025

Quantum oscillations and anisotropic magnetoresistance in the quasi-two-dimensional Dirac nodal line superconductor $\mathrm{YbSb_2}$

Recent interest in quantum materials has focused on systems exhibiting both superconductivity and non-trivial band topology as material candidates to realize topological or unconventional superconducting states. So far, superconductivity in most topological materials has been identified as type II. In this work, we present magnetotransport studies on the quasi-two-dimensional type I superconductor $\mathrm{YbSb_2}$. Combined ab initio DFT calculations and quantum oscillation measurements confirm that $\mathrm{YbSb_2}$ is a Dirac nodal line semimetal in the normal state. The complex Fermi surface morphology is evidenced by the non-monotonic angular dependence of both the quantum oscillation amplitude and the magnetoresistance. Our results establish $\mathrm{YbSb_2}$ as a candidate material platform for exploring the interplay between band topology and superconductivity.

cond-mat.supr-con↗

arXiv2025

Fermi surface and Berry phase analysis for Dirac nodal line semimetals: cautionary tale to SrGa$_2$ and BaGa$_2$

A Berry phase of odd multiples of $π$ inferred from quantum oscillations (QOs) has often been treated as evidence for nontrivial reciprocal space topology. However, disentangling the Berry phase values from the Zeeman effect and the orbital magnetic moment is often challenging. In centrosymmetric compounds, the case is simpler as the orbital magnetic moment contribution is negligible. Although the Zeeman effect can be significant, it is usually overlooked in most studies of QOs in centrosymmetric compounds. Here, we present a detailed study on the non-magnetic centrosymmetric $\mathrm{SrGa_2}$ and $\mathrm{BaGa_2}$, which are predicted to be Dirac nodal line semimetals (DNLSs) based on density functional theory (DFT) calculations. Evidence of the nontrivial topology is found in magnetotransport measurements. The Fermi surface topology and band structure are carefully studied through a combination of angle-dependent QOs, angle-resolved photoemission spectroscopy (ARPES), and DFT calculations, where the nodal line is observed in the vicinity of the Fermi level. Strong de Haas-van Alphen fundamental oscillations associated with higher harmonics are observed in both compounds, which are well-fitted by the Lifshitz-Kosevich (LK) formula. However, even with the inclusion of higher harmonics in the fitting, we found that the Berry phases cannot be unambiguously determined when the Zeeman effect is included. We revisit the LK formula and analyze the phenomena and outcomes that were associated with the Zeeman effect in previous studies. Our experimental results confirm that $\mathrm{SrGa_2}$ and $\mathrm{BaGa_2}$ are Dirac nodal line semimetals. Additionally, we highlight the often overlooked role of spin-damping terms in Berry phase analysis.

cond-mat.mtrl-sci↗

arXiv2025

Mechanism of Anisotropic Crystallization and Phase Transitions under Van der Waals Squeezing

Mechanical confinement strategies, such as van der Waals (vdW) squeezing, have emerged as promising routes for synthesizing non-vdW two-dimensional (2D) layers, surprisingly yielding high-quality single crystals with lateral sizes approaching 100 micrometer. However, the underlying mechanisms by which such a straightforward approach overcomes the long-standing synthesis challenges of non-vdW 2D materials remains a puzzle. Here, we investigate the crystallization dynamics and phase evolution of Bi under vdW confinement through molecular dynamics (MD) simulations powered by a machine-learning force filed fine-tuned and distilled from a pre-trained model with DFT-level accuracy. We reveal that pressure-dependent layer modulation arises from a quantum confinement-driven anisotropic crystallization mechanism, in which out-of-plane layering occurs nearly two orders of magnitude faster than in-plane ordering. Two critical transitions are identified: an alpha-to-beta phase transformation at 1.64 GPa, and a subsequent collapse into a single-atomic layer at 2.19 GPa. The formation of large-area single crystals is enabled by substrate-induced orientational selection and accelerated grain boundary migration, driven by atomic diffusion at elevated temperatures. These findings resolve the mechanistic origin of high-quality 2D crystal growth under confinement and establish guiding principles for the controlled synthesis of metastable 2D single crystals, with implications for next-generation quantum and nanoelectronic devices.

cond-mat.mtrl-sci↗

arXiv2025

Uncovering coupled ionic-polaronic dynamics and interfacial enhancement in Li$_x$FePO$_4$

Understanding and controlling coupled ionic-polaronic dynamics is crucial for optimizing electrochemical performance in battery materials. However, studying such coupled dynamics remains challenging due to the intricate interplay between Li-ion configurations, polaron charge ordering, and lattice vibrations. Here, we develop a fine-tuned machine-learned force field (MLFF) for Li$_x$FePO$_4$ that captures coupled ion-polaron behavior. Our simulations reveal picosecond-scale polaron flips occurring orders of magnitude faster than Li-ion migration, featuring strong correlation to Li configurations. Notably, polaron charge fluctuations are further enhanced at Li-rich/Li-poor phase boundaries, suggesting a potential interfacial electronic conduction mechanism. These results demonstrate the capability of fine-tuned MLFFs to resolve complex coupled transport and provide insight into emergent ionic-polaronic dynamics in multivalent battery cathodes.

cond-mat.mtrl-sci↗

arXiv2025

Kramers nodal lines in intercalated TaS$_2$ superconductors

Kramers degeneracy is one fundamental embodiment of the quantum mechanical nature of particles with half-integer spin under time reversal symmetry. Under the chiral and noncentrosymmetric achiral crystalline symmetries, Kramers degeneracy emerges respectively as topological quasiparticles of Weyl fermions and Kramers nodal lines (KNLs), anchoring the Berry phase-related physics of electrons. However, an experimental demonstration for ideal KNLs well isolated at the Fermi level is lacking. Here, we establish a class of noncentrosymmetric achiral intercalated transition metal dichalcogenide superconductors with large Ising-type spin-orbit coupling, represented by In$_x$TaS$_2$, to host an ideal KNL phase. We provide evidence from angle-resolved photoemission spectroscopy with spin resolution, angle-dependent quantum oscillation measurements, and ab-initio calculations. Our work not only provides a realistic platform for realizing and tuning KNLs in layered materials, but also paves the way for exploring the interplay between KNLs and superconductivity, as well as applications pertaining to spintronics, valleytronics, and nonlinear transport.

cond-mat.supr-con↗

arXiv2025

Anomalous Electrical Transport in the Kagome Magnet YbFe$_6$Ge$_6$

Two-dimensional (2D) kagome metals offer a unique platform for exploring electron correlation phenomena derived from quantum many-body effects. Here, we report a combined study of electrical magnetotransport and neutron scattering on YbFe$_6$Ge$_6$, where the Fe moments in the 2D kagome layers exhibit an $A$-type collinear antiferromagnetic order below $T_{\rm{N}} \approx 500$ K. Interactions between the Fe ions in the layers and the localized Yb magnetic ions in between reorient the $c$-axis aligned Fe moments to the kagome plane below $T_{\rm{SR}} \approx 63$ K. Our magnetotransport measurements show an intriguing anomalous Hall effect (AHE) that emerges in the spin-reorientated collinear state, accompanied by the closing of the spin anisotropy gap as revealed from inelastic neutron scattering. The gapless spin excitations and the Yb-Fe interaction are able to support a dynamic scalar spin chirality, which explains the observed AHE. Therefore, our study demonstrates spin fluctuations may provide an additional scattering channel for the conduction electrons and give rise to AHE even in a collinear antiferromagnet.

cond-mat.str-el↗

arXiv2025

T-MSD: An improved method for ionic diffusion coefficient calculation from molecular dynamics

Ionic conductivity is a critical property of solid ionic conductors, directly influencing the performance of energy storage devices such as batteries. However, accurately calculating ionic conductivity or diffusion coefficient remains challenging due to the complex, dynamic nature of ionic motion, which often yield significant deviations, especially at room temperature. In this study, we propose an improved method, T-MSD, to enhance the accuracy and reliability of diffusion coefficient calculations. Combining time-averaged mean square displacement analysis with block jackknife resampling, this method effectively addresses the impact of rare, anomalous diffusion events and provides robust statistical error estimates from a single simulation. Applied to large-scale deep-potential molecular dynamics simulations, we show that T-MSD eliminates the need for multiple independent simulations while ensuring accurate diffusion coefficient calculations across systems of varying sizes and simulation durations. This approach offers a practical and reliable framework for precise ionic conductivity estimation, advancing the study and design of high-performance solid ionic conductors.

cond-mat.mtrl-sci↗

arXiv2025

Undamped Soliton-like Domain Wall Motion in Sliding Ferroelectrics

Sliding ferroelectricity in bilayer van der Waals materials exhibits ultrafast switching speed and fatigue resistance during the polarization switching, offering an avenue for the design of memories and neuromorphic devices. The unique polarization switching behavior originates from the distinct characteristics of domain wall (DW), which possesses broader width and faster motion compared to conventional ferroelectrics. Herein, using machine-learning-assisted molecular dynamics simulations and field theory analysis, we predict an undamped soliton-like DW motion in sliding ferroelectrics. It is found that the DW in sliding ferroelectric bilayer 3R-MoS2 exhibits uniformly accelerated motion under an external field, with its velocity ultimately reaches the relativistic-like limit due to continuous acceleration. Remarkably, the DW velocity remains constant even after the external field removal, completely deviating from the velocity breakdown observed in conventional ferroelectrics. This work provides opportunities for applications of sliding ferroelectrics in memory devices based on DW engineering.

cond-mat.mtrl-sci↗

arXiv2024

Spontaneous curvature in two-dimensional van der Waals heterostructures

Two-dimensional (2D) van der Waals (vdW) heterostructures consist of different 2D crystals with diverse properties, constituting the cornerstone of the new generation of 2D electronic devices. Yet interfaces in heterostructures inevitably break bulk symmetry and structural continuity, resulting in delicate atomic rearrangements and novel electronic structures. In this paper, we predict that 2D interfaces undergo spontaneous curvature, which means when two flat 2D layers approach each other, they inevitably experience out-of-plane curvature. Based on deep-learning-assisted large-scale molecular dynamics simulations, we observed significant out-of-plane displacements up to 3.8 angstrom in graphene/BN bilayers induced by curvature, producing a stable hexagonal moire pattern, which agrees well with experimentally observations. Additionally, the out-of-plane flexibility of 2D crystals enables the propagation of curvature throughout the system, thereby influencing the mechanical properties of the heterostructure. These findings offer fundamental insights into the atomic structure in 2D vdW heterostructures and pave the way for their applications in devices.

cond-mat.mtrl-sci↗