arXiv subjects

Zezhong Zhang

Publications and source records attributed to Zezhong Zhang.

At least 19 recordsLinked to original sources

arXiv2026

Updated all-electron Dirac--Fock densities and an element-adaptive parameterisation of scattering factors and potentials for neutral atoms

Updated reference data and an analytic parameterisation of elastic electron and X-ray scattering are presented for all 118 neutral atoms. The reference electron densities for the multi-electron elements $Z=2$--$118$ are computed with the relativistic B-spline Dirac--Fock code atomx, while hydrogen is constructed from the exact relativistic one-electron Dirac $1s$ solution; the electron scattering factor $f_e(g)$, X-ray scattering factor $f_x(g)$ and radial moments are derived from these densities. The reported parameterisation extends the fixed-size Lobato--Van Dyck hydrogenic expansion while retaining closed-form expressions for $f_x(g)$, $\rho(r)$, the electrostatic potential $V(r)$ and the projected potential $V(R)$. These extensions are an element-adaptive basis size $n_t(Z)$, a simultaneous real- and reciprocal-space fit, an exact $\langle r^4\rangle$ constraint in place of the non-relativistic Kato cusp, and a charge-carrying Dirac--Pad\'e basis term that adds a polynomial-times-exponential shape channel without replacing the hydrogenic basis by a tabulated Dirac radial function or assigning the term to a physical shell. For the same total parameter count, the Dirac--Pad\'e-enriched basis improves on the parameter-matched non-relativistic basis for 111 of the 118 elements, lowering the mean total cost by 39\%. Relative to a controlled fixed five-term refit on the same reference grid and objective, the element-adaptive bases improve the median reciprocal-space deviations by about three to four orders of magnitude and resolve shell structure in $4\pi r^2\rho(r)$ that the fixed five-term basis cannot. The largest changes occur near the nucleus and in the reciprocal-space tail beyond the legacy 12 inverse angstroms range, which is directly relevant to quantitative high-angle scattering and electron-diffraction measurements.

physics.comp-ph↗

arXiv2026

Pruning-Aware Multi-Cluster Co-Inference for Large AI Models in AI-RANs

The increasing scale and computational demands of large artificial intelligence models (LAIMs) present significant challenges for efficient inference in resource-constrained distributed environments. In this paper, we propose a multi-cluster LAIM co-inference framework, where an edge server equipped with multiple graphics processing units (GPUs) coordinates multiple user clusters to execute inference tasks collaboratively. Within each cluster, devices capture data from diverse perspectives and employ lightweight on-device LAIMs to extract local features. These features are then transmitted to the edge server, where they are aggregated and fused to generate a more accurate inference outcome. To reveal the fundamental trade-off between model pruning and collaborative inference performance, we develop a theoretical framework that characterizes the impact of pruning ratios and device contributions using rate-distortion theory and partial information decomposition. Based on this analysis, we formulate a joint optimization problem that determines the model pruning ratio, the task scheduling strategy, the bandwidth allocation, and the transmission power, with the goal of minimizing the inference distortion while satisfying the constraints of latency, energy consumption, and server capacity. Extensive simulation results demonstrate that the proposed framework significantly outperforms existing benchmark schemes, achieving superior inference accuracy and resource efficiency in multi-cluster edge intelligence networks.

cs.DC↗

arXiv2026

IEnSF: Iterative Ensemble Score Filter for Reducing Error in Posterior Score Estimation in Nonlinear Data Assimilation

The Ensemble Score Filter (EnSF) is a score-based diffusion model approach for solving high-dimensional and nonlinear data assimilation problems. While initial applications of EnSF to the Lorenz-96 model and the quasi-geostrophic system showed potential, the current method employs a heuristic weighted sum to combine the prior and the likelihood score functions. This introduces a structural error into the estimation of the posterior score function in the nonlinear setting. This work addresses this challenge by developing an iterative ensemble score filter (IEnSF) that applies an iterative algorithm as an outer loop around the reverse-time stochastic differential equation solver. When the state dynamics or the observation operator is nonlinear, the iterative algorithm can gradually reduce the posterior score estimation error by improving the accuracy of approximating the conditional expectation of the likelihood score function. The number of iterations required depends on the distance between the prior and posterior distributions. Numerical experiments demonstrate that the IEnSF algorithm substantially reduces the error in posterior score estimation in the nonlinear setting and thus improves the accuracy of tracking high-dimensional dynamical systems.

math.NA↗

arXiv2026

OSP-Next: Efficient High-Quality Video Generation with Sparse Sequence Parallelism, HiF8 Quantization, and Reinforcement Learning

Diffusion Transformers achieve strong video generation quality, but the quadratic cost of full attention limits efficiency. We introduce OSP-Next, an efficient text-to-video generation model that integrates sparse attention, parallelism, quantization, and reinforcement learning. OSP-Next uses a hybrid full-sparse attention architecture, where the sparse component is implemented with Skiparse-2D Attention. This fixed-pattern mechanism applies token-wise and group-wise sparse attention along spatial dimensions, leveraging locality while maintaining native compatibility with FlashAttention kernels. Based on the local equivalence of rearrangement in Skiparse-2D Attention, we further propose Sparse Sequence Parallelism (SSP), which partitions subsequences across ranks and switches sparse patterns through a single All-to-All communication. Compared with Ulysses Sequence Parallelism (SP), SSP provides a native parallel strategy for sparse attention and reduces communication volume by 75%. OSP-Next also incorporates HiF8 quantization to enable stable joint training with 8-bit quantization and sparse fine-tuning, and applies Mix-GRPO post-training to improve the performance of the sparse model. Experiments show that OSP-Next achieves a VBench total score of 83.73%, surpassing the Wan2.1 baseline. Under the 5-second 720P and 5-second 768P settings, OSP-Next achieves up to 1.64$\times$ single-GPU speedup and over 1.52$\times$ eight-GPU speedup on NVIDIA H200 GPUs. In addition, with only a 0.4% drop in VBench total score, OSP-Next-HiF8 achieves 1.69$\times$ and 2.27$\times$ speedups under the two settings on a single Ascend 950PR, demonstrating the efficiency and performance of OSP-Next across hardware platforms.

cs.CV↗

arXiv2026

Global Attention with Linear Complexity for Exascale Generative Data Assimilation in Earth System Prediction

Accurate Earth system prediction requires state inference from incomplete observations, but conventional two-stage data assimilation (DA) is computationally prohibitive because repeated PDE-based ensemble forecasts, observation updates, and intermediate data movement limit ensemble size at high resolution. We introduce STORM, a one-stage generative AI framework that reformulates DA as diffusion-based Bayesian posterior sampling, replacing online PDE ensemble forecasts with scalable AI inference. It further combines a spatiotemporal transformer with a global-attention algorithm that reduces complexity from quadratic to linear through scalable gradient propagation, enabling high-resolution, long-context Earth modeling. STORM scales to 74,400 GPUs on Frontier with 96--99\% strong-scaling efficiency and up to 6 ExaFLOPs sustained BF16 throughput, while enabling 32,768-member ensembles for uncertainty quantification in 34 seconds on 4,096 GPUs. It scales to 20 billion spatiotemporal tokens and 177,000 temporal frames. Hurricane tracking and long-term climate reanalysis demonstrate improved accuracy, including benefits from longer temporal context and recovery of temperature extremes missed by forecast-only predictions.

cs.LG↗

arXiv2026

EMS-FL: Federated Tuning of Mixture-of-Experts in Satellite-Terrestrial Networks via Expert-Driven Model Splitting

The rapid advancement of large AI models imposes stringent demands on data volume and computational resources. Federated learning, though designed to exploit distributed data and computational resources, faces data shortage from limited network coverage and computational constraints from edge devices. To address these issues, both the mixture-of-experts (MoE) and satellite-terrestrial network (STN) provide promising solutions, offering lightweight computation overhead and broad coverage, respectively. However, the satellite-ground relative motion results in intermittent connectivity, hindering conventional federated learning that relies on model synchronization across devices. To leverage the coverage of STN while preserving training efficiency, we propose EMS-FL, an expert-driven model splitting and federated learning method. EMS-FL assigns each device cluster only the experts highly correlated to their local data. Through non-overlapping expert assignments, asynchronous local learning is further proposed, where each device cluster trains its assigned experts consecutively and only uploads local parameters to the satellite during connected phases for aggregation and model updates. Consequently, EMS-FL effectively reduces the training overhead and achieves both faster convergence and higher accuracy compared with conventional federated learning. Rigorous convergence analysis is provided to theoretically characterize the learning performance. Furthermore, comprehensive experiments are conducted using public datasets and large models, validating the superiority of EMS-FL.

cs.NI↗

arXiv2026

BeamVLM for Low-altitude Economy: Generative Beam Prediction via Vision-language Models

For low-altitude economy (LAE), fast and accurate beam prediction between high-mobility unmanned aerial vehicles (UAVs) and ground base stations is of paramount importance, which ensures seamless coverage and reliable communications. However, existing deep learning-based beam prediction methods lack high-level semantic understanding of dynamic environments, resulting in poor generalization. On the other hand, the emerging large language model (LLM) based approaches show promise in enhancing generalization, but they typically lack rich environmental perception, thereby failing to capture fine-grained spatial semantics essential for precise beam alignment. To tackle these limitations, we propose in this correspondence a novel end-to-end generative framework for beam prediction, called BeamVLM, which treats beam prediction as a vision question answering task capitalizing on powerful existing vision-language models (VLMs). By projecting raw visual patches directly into the language domain and judiciously designing an instructional prompt, the proposed BeamVLM enables the VLM to jointly reason over UAV trajectories and environmental context. Last, experimental results on real-world datasets demonstrate that the proposed BeamVLM outperforms state-of-the-art methods in prediction accuracy and also exhibits superior generalization for other scenarios such as vehicle-to-infrastructure (V2I) beam prediction.

cs.NI↗

arXiv2026

RadioGen3D: 3D Radio Map Generation via Adversarial Learning on Large-Scale Synthetic Data

Radio maps are essential for efficient radio resource management in future 6G and low-altitude networks. While deep learning (DL) techniques have emerged as an efficient alternative to conventional ray-tracing for radio map estimation (RME), most existing DL approaches are confined to 2D near-ground scenarios. They often fail to capture essential 3D signal propagation characteristics and antenna polarization effects, primarily due to the scarcity of 3D data and training challenges. To address these limitations, we present the RadioGen3D framework. First, we propose an efficient data synthesis method to generate high-quality 3D radio map data. By establishing a parametric target model that captures 2D ray-tracing and 3D channel fading characteristics, we derive realistic coefficient combinations from minimal real measurements, enabling the construction of a large-scale synthetic dataset, Radio3DMix. Utilizing this dataset, we propose a 3D model training scheme based on a conditional generative adversarial network (cGAN), yielding a 3D U-Net capable of accurate RME under diverse input feature combinations. Experimental results demonstrate that RadioGen3D surpasses all baselines in both estimation accuracy and speed. Furthermore, fine-tuning experiments verify its strong generalization capability via successful knowledge transfer.

cs.LG↗

arXiv2026

Error estimates of a training-free diffusion model for high-dimensional sampling

Score-based diffusion models are a powerful class of generative models, but their practical use often depends on training neural networks to approximate the score function. Training-free diffusion models provide an attractive alternative by exploiting analytically tractable score functions, and have recently enabled supervised learning of efficient end-to-end generative samplers. Despite their empirical success, the training-free diffusion models lack rigorous and numerically verifiable error estimates. In this work, we develop a comprehensive error analysis for a class of training-free diffusion models used to generate labeled data for supervised learning of generative samplers. By exploiting the availability of the exact score function for Gaussian mixture models, our analysis avoids propagating score-function approximation errors through the reverse-time diffusion process and recovers classical convergence rates for ODE discretization schemes, such as first-order convergence for the Euler method. Moreover, the resulting error bounds exhibit favorable dimension dependence, scaling as $O(d)$ in the $\ell_2$ norm and $O(\log d)$ in the $\ell_\infty$ norm. Importantly, the proposed error estimates are fully numerically verifiable with respect to both time-step size and dimensionality, thereby bridging the gap between theoretical analysis and observed numerical behavior.

math.NA↗

arXiv2025

Bohrium + SciMaster: Building the Infrastructure and Ecosystem for Agentic Science at Scale

AI agents are emerging as a practical way to run multi-step scientific workflows that interleave reasoning with tool use and verification, pointing to a shift from isolated AI-assisted steps toward \emph{agentic science at scale}. This shift is increasingly feasible, as scientific tools and models can be invoked through stable interfaces and verified with recorded execution traces, and increasingly necessary, as AI accelerates scientific output and stresses the peer-review and publication pipeline, raising the bar for traceability and credible evaluation. However, scaling agentic science remains difficult: workflows are hard to observe and reproduce; many tools and laboratory systems are not agent-ready; execution is hard to trace and govern; and prototype AI Scientist systems are often bespoke, limiting reuse and systematic improvement from real workflow signals. We argue that scaling agentic science requires an infrastructure-and-ecosystem approach, instantiated in Bohrium+SciMaster. Bohrium acts as a managed, traceable hub for AI4S assets -- akin to a HuggingFace of AI for Science -- that turns diverse scientific data, software, compute, and laboratory systems into agent-ready capabilities. SciMaster orchestrates these capabilities into long-horizon scientific workflows, on which scientific agents can be composed and executed. Between infrastructure and orchestration, a \emph{scientific intelligence substrate} organizes reusable models, knowledge, and components into executable building blocks for workflow reasoning and action, enabling composition, auditability, and improvement through use. We demonstrate this stack with eleven representative master agents in real workflows, achieving orders-of-magnitude reductions in end-to-end scientific cycle time and generating execution-grounded signals from real workloads at multi-million scale.

cs.AI↗

arXiv2025

Mapping vacancy and bonding electron distributions around aluminium nanovoids

All materials have defects and many contain nanostructures, both of which disrupt chemical bonding - the basis of materials properties. No experimental measurements of bonding electron distributions associated with defects and nanostructures have ever been possible. We present a method enabling such measurements and interrogate nanovoids surrounded by vacancies - the most fundamental of nanostructures and defects - in aluminium. We measure the volume of a vacancy with 3% uncertainty and map vacancy concentrations surrounding nanovoids with nanometre resolution in three dimensions where previously only two-dimensional mapping was possible. We discover that radiation-damaged voids can "heal". Our bonding measurements are depth-resolved, vacancy-sensitive, and agree with density functional theory. This work opens bonding electron density measurements to inhomogeneous nanostructured multi-phased materials so that the electronic origins of phenomena such as strengthening, weakening, interface functionality, solute diffusion and phase transformations within them may be revealed.

cond-mat.mtrl-sci↗

arXiv2025

Interfacial Control of Orbital Occupancy and Spin State in LaCoO$_3$

Transition metal oxides exhibit a wide range of tunable electronic properties arising from the complex interplay of charge, spin, and lattice degrees of freedom, governed by their $d$ orbital configurations, making them particularly interesting for oxide electronics and (electro)catalysis. Perovskite oxide heterointerfaces offer a promising route to engineer these orbital states. In this work, we tune the Co $3d$ orbital occupancy in LaCoO$_3$ from a partial $d^7$ to a partial $d^5$ state through interfacial engineering with LaTiO$_3$, LaMnO$_3$, LaAlO$_3$ and LaNiO$_3$. Using X-ray absorption spectroscopy combined with charge transfer multiplet calculations, we identify differences in the Co valence and spin state for the series of oxide heterostructures. LaTiO$_3$ and LaMnO$_3$ interfaces result in interfacial charge transfer towards LaCoO$_3$, resulting in a partial $d^7$ orbital occupancy, while a LaNiO$_3$ interface results in a partial Co $d^5$ occupancy. Strikingly, a LaAlO$_3$ spacer layer between LaNiO$_3$ and LaCoO$_3$ results in a Co $d^6$ low spin state. These results indicate that the Co spin state, like the valence state, is governed by the interfacial environment. High-resolution scanning transmission electron microscopy imaging reveals a clear connection between strain and spin configuration, emphasizing the importance of structural control at oxide interfaces. Overall, this work demonstrates that interfacial engineering simultaneously governs orbital occupancy and spin state in correlated oxides, advancing spin-engineering strategies in correlated oxides and offering new insights for the rational design of functional oxide heterostructures.

cond-mat.mtrl-sci↗

arXiv2025

Sub-unit cell engineering of CrVO$_3$ superlattice thin films

Ordered corundum oxides introduce new prospects in the field of functional oxides thin films, complementing the more widely studied class of ABO$_3$ perovskites. In this work, we take advantage of the layer-by-layer growth regime to fabricate epitaxial CrVO$_3$ superlattice thin films with atomic-scale accuracy on the periodic arrangement of Cr and V layers. By means of X-ray diffraction, scanning transmission electron microscopy and Raman spectroscopy, we confirm the thickness control in the sub-unit cell scale, alternating 3, 2 or 1 single atomic layers of Cr$_2$O$_3$ and V$_2$O$_3$. For the first time, we stabilize the ilmenite phase of CrVO$_3$ (space group R-3) and compare the functional properties of the thin film with those calculated by density functional theory. This novel approach to the growth of ordered corundum oxides opens the path towards the stabilization of new complex oxides with tailored properties by varying the composition and the superlattice period, ultimately broadening the family of functional rhombohedral oxides.

cond-mat.mtrl-sci↗

arXiv2025

Uni-AIMS: AI-Powered Microscopy Image Analysis

This paper presents a systematic solution for the intelligent recognition and automatic analysis of microscopy images. We developed a data engine that generates high-quality annotated datasets through a combination of the collection of diverse microscopy images from experiments, synthetic data generation and a human-in-the-loop annotation process. To address the unique challenges of microscopy images, we propose a segmentation model capable of robustly detecting both small and large objects. The model effectively identifies and separates thousands of closely situated targets, even in cluttered visual environments. Furthermore, our solution supports the precise automatic recognition of image scale bars, an essential feature in quantitative microscopic analysis. Building upon these components, we have constructed a comprehensive intelligent analysis platform and validated its effectiveness and practicality in real-world applications. This study not only advances automatic recognition in microscopy imaging but also ensures scalability and generalizability across multiple application domains, offering a powerful tool for automated microscopic analysis in interdisciplinary research. A online application is made available for researchers to access and evaluate the proposed automated analysis service.

eess.IV↗

arXiv2025

RadioDiff-3D: A 3D$\times$3D Radio Map Dataset and Generative Diffusion Based Benchmark for 6G Environment-Aware Communication

Radio maps (RMs) serve as a critical foundation for enabling environment-aware wireless communication, as they provide the spatial distribution of wireless channel characteristics. Despite recent progress in RM construction using data-driven approaches, most existing methods focus solely on pathloss prediction in a fixed 2D plane, neglecting key parameters such as direction of arrival (DoA), time of arrival (ToA), and vertical spatial variations. Such a limitation is primarily due to the reliance on static learning paradigms, which hinder generalization beyond the training data distribution. To address these challenges, we propose UrbanRadio3D, a large-scale, high-resolution 3D RM dataset constructed via ray tracing in realistic urban environments. UrbanRadio3D is over 37$\times$3 larger than previous datasets across a 3D space with 3 metrics as pathloss, DoA, and ToA, forming a novel 3D$\times$33D dataset with 7$\times$3 more height layers than prior state-of-the-art (SOTA) dataset. To benchmark 3D RM construction, a UNet with 3D convolutional operators is proposed. Moreover, we further introduce RadioDiff-3D, a diffusion-model-based generative framework utilizing the 3D convolutional architecture. RadioDiff-3D supports both radiation-aware scenarios with known transmitter locations and radiation-unaware settings based on sparse spatial observations. Extensive evaluations on UrbanRadio3D validate that RadioDiff-3D achieves superior performance in constructing rich, high-dimensional radio maps under diverse environmental dynamics. This work provides a foundational dataset and benchmark for future research in 3D environment-aware communication. The dataset is available at https://github.com/UNIC-Lab/UrbanRadio3D.

cs.LG↗

arXiv2025

Exact Conditional Score-Guided Generative Modeling for Amortized Inference in Uncertainty Quantification

We propose an efficient framework for amortized conditional inference by leveraging exact conditional score-guided diffusion models to train a non-reversible neural network as a conditional generative model. Traditional normalizing flow methods require reversible architectures, which can limit their expressiveness and efficiency. Although diffusion models offer greater flexibility, they often suffer from high computational costs during inference. To combine the strengths of both approaches, we introduce a two-stage method. First, we construct a training-free conditional diffusion model by analytically deriving an exact score function under a Gaussian mixture prior formed from samples of the underlying joint distribution. This exact conditional score model allows us to efficiently generate noise-labeled data, consisting of initial diffusion Gaussian noise and posterior samples conditioned on various observation values, by solving a reverse-time ordinary differential equation. Second, we use this noise-labeled data to train a feedforward neural network that maps noise and observations directly to posterior samples, eliminating the need for reversibility or iterative sampling at inference time. The resulting model provides fast, accurate, and scalable conditional sampling for high-dimensional and multi-modal posterior distributions, making it well-suited for uncertainty quantification tasks, e.g., parameter estimation of complex physical systems. We demonstrate the effectiveness of our approach through a series of numerical experiments.

cs.CE↗

arXiv2025

Diffusion-based supervised learning of generative models for efficient sampling of multimodal distributions

We propose a hybrid generative model for efficient sampling of high-dimensional, multimodal probability distributions for Bayesian inference. Traditional Monte Carlo methods, such as the Metropolis-Hastings and Langevin Monte Carlo sampling methods, are effective for sampling from single-mode distributions in high-dimensional spaces. However, these methods struggle to produce samples with the correct proportions for each mode in multimodal distributions, especially for distributions with well separated modes. To address the challenges posed by multimodality, we adopt a divide-and-conquer strategy. We start by minimizing the energy function with initial guesses uniformly distributed within the prior domain to identify all the modes of the energy function. Then, we train a classifier to segment the domain corresponding to each mode. After the domain decomposition, we train a diffusion-model-assisted generative model for each identified mode within its support. Once each mode is characterized, we employ bridge sampling to estimate the normalizing constant, allowing us to directly adjust the ratios between the modes. Our numerical examples demonstrate that the proposed framework can effectively handle multimodal distributions with varying mode shapes in up to 100 dimensions. An application to Bayesian inverse problem for partial differential equations is also provided.

stat.ML↗

arXiv2024

GenAI4UQ: A Software for Inverse Uncertainty Quantification Using Conditional Generative Models

We introduce GenAI4UQ, a software package for inverse uncertainty quantification in model calibration, parameter estimation, and ensemble forecasting in scientific applications. GenAI4UQ leverages a generative artificial intelligence (AI) based conditional modeling framework to address the limitations of traditional inverse modeling techniques, such as Markov Chain Monte Carlo methods. By replacing computationally intensive iterative processes with a direct, learned mapping, GenAI4UQ enables efficient calibration of model input parameters and generation of output predictions directly from observations. The software's design allows for rapid ensemble forecasting with robust uncertainty quantification, while maintaining high computational and storage efficiency. GenAI4UQ simplifies the model training process through built-in auto-tuning of hyperparameters, making it accessible to users with varying levels of expertise. Its conditional generative framework ensures versatility, enabling applicability across a wide range of scientific domains. At its core, GenAI4UQ transforms the paradigm of inverse modeling by providing a fast, reliable, and user-friendly solution. It empowers researchers and practitioners to quickly estimate parameter distributions and generate model predictions for new observations, facilitating efficient decision-making and advancing the state of uncertainty quantification in computational modeling. (The code and data are available at https://github.com/patrickfan/GenAI4UQ).

cs.LG↗