arXiv subjects

Hao Wu

Publications and source records attributed to Hao Wu.

At least 181 records · Page 10Linked to original sources

arXiv2026

Quantitative coronary calcification analysis for prediction of myocardial ischemia using non-contrast CT calcium scoring

Non-contrast computed tomography calcium scoring (CTCS) is widely recognized as an effective tool for cardiovascular risk stratification. This study aimed to develop a novel machine learning framework for predicting myocardial ischemia from routine non-contrast CTCS scans using quantitative coronary calcium assessment. This study analyzed 1,375 patients who underwent both non-contrast CTCS and regadenoson stress cardiac positron emission tomography myocardial perfusion imaging within one year at University Hospitals Cleveland Medical Center. A total of 74 variables, including clinical variables, Agatston score, and calcium-omics features, were evaluated. Relevant features were identified using XGBoost with Shapley Additive exPlanations (SHAP). Predictive models were trained and evaluated using 5-fold cross-validation. Among 987 patients, 89 (9%) were positive for myocardial ischemia. The final model incorporated the Agatston score, eight calcium-omics features, and age. The proposed model achieved a precision of 98.9+/-3.0%, sensitivity of 79.2+/-8.4, and F1 score of 87.7+/-5.3%. The addition of calcium-omics features significantly improved predictive performance compared with models using clinical variables alone or clinical variables with the Agatston score (p<0.05). Interestingly, the number of calcified arteries, despite being the lowest-ranked feature based on SHAP analysis, showed the strongest association with myocardial ischemia in logistic regression analysis (odds ratio: 3.63, 95% confidence interval: 2.80-4.77, p<0.00001). We developed a machine learning approach for predicting myocardial ischemia using routinely acquired non-contrast CTCS scans. Calcium-omics features provided incremental predictive value beyond conventional risk factors and Agatston scoring and may support more accessible cardiovascular risk stratification.

cs.LG↗

arXiv2026

Differential-Integral Neural Operator for Long-Term Turbulence Forecasting

Accurately forecasting the long-term evolution of turbulence represents a grand challenge in scientific computing and is crucial for applications ranging from climate modeling to aerospace engineering. Existing deep learning methods, particularly neural operators, often fail in long-term autoregressive predictions, suffering from catastrophic error accumulation and a loss of physical fidelity. This failure stems from their inability to simultaneously capture the distinct mathematical structures that govern turbulent dynamics: local, dissipative effects and global, non-local interactions. In this paper, we propose the {\textbf{\underline{D}}}ifferential-{\textbf{\underline{I}}}ntegral {\textbf{\underline{N}}}eural {\textbf{\underline{O}}}perator (\method{}), a novel framework designed from a first-principles approach of operator decomposition. \method{} explicitly models the turbulent evolution through parallel branches that learn distinct physical operators: a local differential operator, realized by a constrained convolutional network that provably converges to a derivative, and a global integral operator, captured by a Transformer architecture that learns a data-driven global kernel. This physics-based decomposition endows \method{} with exceptional stability and robustness. Through extensive experiments on the challenging 2D Kolmogorov flow benchmark, we demonstrate that \method{} significantly outperforms state-of-the-art models in long-term forecasting. It successfully suppresses error accumulation over hundreds of timesteps, maintains high fidelity in both the vorticity fields and energy spectra, and establishes a new benchmark for physically consistent, long-range turbulence forecast.

cs.LG↗

arXiv2026

C-ReD: A Comprehensive Chinese Benchmark for AI-Generated Text Detection Derived from Real-World Prompts

Recently, large language models (LLMs) are capable of generating highly fluent textual content. While they offer significant convenience to humans, they also introduce various risks, like phishing and academic dishonesty. Numerous research efforts have been dedicated to developing algorithms for detecting AI-generated text and constructing relevant datasets. However, in the domain of Chinese corpora, challenges remain, including limited model diversity and data homogeneity. To address these issues, we propose C-ReD: a comprehensive Chinese Real-prompt AI-generated Detection benchmark. Experiments demonstrate that C-ReD not only enables reliable in-domain detection but also supports strong generalization to unseen LLMs and external Chinese datasets-addressing critical gaps in model diversity, domain coverage, and prompt realism that have limited prior Chinese detection benchmarks. We release our resources at https://github.com/HeraldofLight/C-ReD.

cs.CL↗

arXiv2026

SERE: A Stabilized Element-Wise Method for Downlink Rate Estimation in Clustered Cell-Free Networks

Clustered cell-free networks have emerged as a promising architecture for sixth generation ultra-dense wireless communication systems by enabling local cooperation among base stations while controlling system complexity. For resource allocation and performance optimization of such networks, accurate and efficient estimation of the ergodic achievable downlink rate is a fundamental prerequisite. Existing rate estimation approaches mainly rely on computationally prohibitive Monte Carlo simulations or adopt random matrix theory-based methods, which have been well-developed for conventional cellular and cell-free networks. However, existing RMT-based methods have not addressed the unique inter-subnetwork interference in clustered cell-free networks, and therefore lack an efficient solution for accurate downlink rate estimation under both regularized zero-forcing and zero-forcing precoding. In this paper, we propose a stabilized element-wise rate estimation method for downlink rate estimation in clustered cell-free networks. We establish the diagonal element-wise convergence of resolvent matrices, which enables the derivation of deterministic equivalents for inter-subnetwork interference and the downlink ergodic rate. We further introduce a stabilized variable transformation to address the numerical instability when the regularization parameter is very small, hereby enabling a unified formulation applicable to both regularized zero-forcing and zero-forcing precoding. Simulation results show that the proposed method achieves a relative error below 6% while significantly reducing computational complexity compared with the Monte Carlo simulation.

cs.IT↗

arXiv2026

Achieving $α$-Fairness in Clustered Cell-Free Networking: A Tight Relaxation Approach

Clustered cell-free networking has emerged as a promising architecture to balance the high performance of cell-free massive MIMO and the scalability of traditional cellular systems. However, achieving fairness across subnetworks remains a critical yet largely unsolved challenge. This paper investigates the fairness problem in clustered cell-free networking and proposes a unified and tunable alpha-fairness scheme that effectively balances overall spectral efficiency and inter-subnetwork fairness. Using the closed-form deterministic equivalent of the ergodic sum capacity, we reformulate the combinatorial clustering problem as a continuous optimization problem. Leveraging the concavity/convexity properties of the alpha-fair objective, we classify the problem into four distinct cases according to the value of alpha. For each case, we establish the exact equivalence between the original integer program and its continuous relaxation, and develop efficient algorithms with guaranteed convergence. Extensive simulations show that the proposed scheme achieves up to 11% improvement in Jain's fairness index and 45% gain in minimum subnetwork capacity, with only a negligible 5% reduction in aggregate throughput.

cs.IT↗

arXiv2026

ElasticDiT: Efficient Diffusion Transformers via Elastic Architecture and Sparse Attention for High-Resolution Image Generation on Mobile Devices

The Diffusion Transformer (DiT) architecture is the state-of-the-art paradigm for high-fidelity image generation, underpinning models like Stable Diffusion-3 and FLUX.1. However, deploying these models on resource-constrained mobile devices entails prohibitive computational and memory overhead. While efficiency-driven approaches like Linear-DiT and static pruning alleviate bottlenecks, they often incur quality degradation. Unlike cloud environments, mobile constraints require a single-model paradigm that dynamically balances fidelity and latency. We introduce ElasticDiT, which achieves this dynamic trade-off by adjusting spatial compression ratios and DiT block depths. By integrating Shift Sparse Block Attention (SSBA) and a Tiny DWT-Distilled VAE (T-DVAE), ElasticDiT reduces inference latency and memory footprint while maintaining image quality. Experiments confirm that ElasticDiT effectively covers a wide range of fidelity-latency trade-offs within a single set of parameters. By jointly adjusting compression and depth, a single ElasticDiT model can be reconfigured on-the-fly to outperform task-specific baselines. Specifically, our flex lite variant achieves an HPS of 32.87, surpassing the Flux model, while maintaining competitive quality at 84.16 percent average sparsity through SSBA. Furthermore, the plug-and-play T-DVAE provides SD3-level reconstruction with only 1/8x the computational cost of standard VAEs, and Flow-GRPO boosts semantic alignment (GenEval: 66.93 to 73.62). These results demonstrate that ElasticDiT offers a versatile, hardware-adaptive solution that eliminates the need for multiple specialized models, providing a promising path for future high-resolution image generation on mobile devices.

cs.CV↗

arXiv2026

MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents

Existing memory systems for embodied agents typically inject retrieved memory as static context at episode start, a paradigm we term Ahead-of-time Monolithic Memory Injection (AMMI). However, this static design quickly becomes misaligned with the agent's evolving state and may degrade lightweight executors below the no-memory baseline. To address this, we propose MemCompiler, which reframes memory utilization as State-Conditioned Memory Compilation. A learned Memory Compiler reads a structured Brief State capturing the agent's current execution state and dynamically selects and compiles only relevant memory into executable guidance. This guidance is delivered through a text channel and a latent Soft-Mem channel that preserves perceptual information not expressible in text. Across Alf World, EmbodiedBench, and ScienceWorld, MemCompiler consistently improves over no-memory across open-source backbones (up to +129%), matches or approaches frontier closed-source systems, and reduces per-step latency by 60%, demonstrating that state-aware memory compilation improves both effectiveness and efficiency.

cs.RO↗

arXiv2026

Weight Hybrid Architecture of Rydberg-Atomic Sensors

Rydberg atomic quantum receivers have been seen as novel radio frequency measurements and the high sensitivity to a large range of frequencies makes it attractive for communications reception. However, their performance can be significantly degraded by hardware-induced noise, particularly the noise from laser, which impacts the overall system noise floor and exhibits correlation. To address this challenge, this paper proposes a weight hybrid (WH) architecture for Rydberg-atomic sensors, a novel four-channel combining scheme designed for atomic sensors operating in correlated noise environments. By jointly processing dual signal channels and dual noise reference channels, the WH architecture effectively mitigates noise contributions from lasers and other hardware components. All channels are optimally combined via maximum likelihood estimation within an expectation maximization framework, enabling robust signal extraction under correlated noise. Moreover, the proposed WH architecture is universal and can be readily extended to other types of Rydberg receivers to achieve consistent performance improvements.

eess.SP↗

arXiv2026

Usability as a Weapon: Attacking the Safety of LLM-Based Code Generation via Usability Requirements

Large Language Models (LLMs) are increasingly used for automated software development, making their ability to preserve secure coding practices critical. In practice, however, many security requirements are implicit or underspecified, whereas usability requirements are explicit and high-signal. This asymmetry motivates our investigation of usability pressure as a practical attack surface: realistic usability-oriented requirements (e.g., new features, performance constraints, or simplicity demands) can cause coding LLMs to satisfy explicit usability goals while silently dropping implicit security constraints -- a form of reward hacking. We formalize this threat as UPAttack and propose U-SPLOIT, an automated framework to craft UPAttack that (i) selects tasks where a model is initially secure, (ii) synthesizes usability pressures by identifying usability rewards of insecure alternatives across three vectors (Functionality, Implementation, Trade-off), and (iii) verifies security regression via both existing test cases and dynamically generated exploit payloads. Across 75 seed scenarios (25 CWEs x 3 cases), spanning multiple languages (Python, C, and JavaScript), U-SPLOIT achieves attack success rates up to 98.1% on multiple state-of-the-art models (e.g., GPT-5.2-chat and Gemini-3-Flash-Preview).

cs.CR↗

arXiv2026

Deep set based operator learning with uncertainty quantification

Learning operators from data is central to scientific machine learning. While DeepONets are widely used for their ability to handle complex domains, they require fixed sensor numbers and locations, lack mechanisms for uncertainty quantification, and are thus limited in practical applicability. Recent permutation-invariant extensions, such as the Variable-Input Deep Operator Network, relax these sensor constraints but still rely on sufficiently dense observations and cannot capture uncertainties arising from incomplete measurements or from operators with inherent randomness. To address these challenges, we propose UQ-SONet, a permutation-invariant operator learning framework with built-in uncertainty quantification. Our model integrates a set transformer embedding to handle sparse and variable sensor locations, and employs a conditional variational autoencoder to approximate the conditional distribution of the solution operator. By minimizing the negative ELBO, UQ-SONet provides principled uncertainty estimation while maintaining predictive accuracy. Numerical experiments on deterministic and stochastic PDEs, including the Navier-Stokes equation, demonstrate the robustness and effectiveness of the proposed framework.

cs.LG↗

arXiv2026

SLQ: Bridging Modalities via Shared Latent Queries for Retrieval with Frozen MLLMs

Multimodal Large Language Models (MLLMs) possess intrinsic reasoning and world-knowledge capabilities, yet adapting them for dense retrieval remains challenging. Existing approaches rely on invasive parameter updates, such as full fine-tuning and LoRA, which may disrupt the pre-trained semantic space and impair the structured knowledge essential for reasoning. To address this, we propose SLQ, a parameter-efficient tuning framework that adapts MLLMs for retrieval while keeping the backbone entirely frozen. SLQ introduces a small set of Shared Latent Queries that are appended to both text and image tokens, leveraging the model's native causal attention to aggregate multimodal context into a unified embedding space. Furthermore, to better evaluate retrieval beyond superficial pattern matching, we construct KARR-Bench, a benchmark designed for knowledge-aware reasoning retrieval. Extensive experiments show that SLQ outperforms full fine-tuning and LoRA on COCO and Flickr30K, while achieving competitive performance on MMEB and yielding substantial gains on KARR-Bench, validating that preserving the pre-trained representations via non-invasive adaptation is an effective strategy for MLLM-based retrieval. The code is available under: https://github.com/CnFaker/SLQ.

cs.CV↗

arXiv2026

Dual-Correction Physics-Informed Neural Networks for Hemodynamic Reconstruction from Sparse Data

Quantifying hemodynamics in the curved segments of the intracranial internal carotid artery is a core challenge in diagnosing vascular stenosis. Conventional full-field imaging, such as 4D Flow MRI, is costly and difficult to widely promote. Meanwhile, reconstructing full-field fluid information from easily accessible and non-invasive sparse measurement data (such as transcranial Doppler ultrasound/computed tomography angiography) is essentially a highly challenging ill-posed inverse problem. To overcome the severe optimization difficulties and generalization failures of conventional physics-informed neural networks (PINNs) in highly tortuous geometries, we propose a dual-correction physics-informed neural network (DCP-INN) framework taking into account a causal decoupling strategy. The proposed DCP-INN model utilizes a diamond-shaped main network to capture low-frequency trends in physical evolution, and employs a parallel wide-deep correction network to compensate for high-frequency residuals resulting from complex geometric shapes. Furthermore, the framework introduces a high-order physical loss function based on Taylor expansion to enhance local continuity under extremely sparse data constraints. To validate the proposed method, we performed computational evaluations on realistic vascular geometries with significant tortuosity. The results demonstrate that the method effectively mitigates optimization challenges and significantly reduces flow field reconstruction error. This study not only achieves physically credible and robust flow field reconstruction in complex morphologies but also provides a highly promising algorithmic foundation for building low-cost, high-resolution personalized cardiovascular digital twins in future.

physics.med-ph↗

arXiv2026

InterCoG: Towards Spatially Precise Image Editing with Interleaved Chain-of-Grounding Reasoning

Emerging unified editing models have demonstrated strong capabilities in general object editing tasks. However, it remains a significant challenge to perform fine-grained editing in complex multi-entity scenes, particularly those where targets are not visually salient and require spatial reasoning. To this end, we propose InterCoG, a novel text-vision Interleaved Chain-of-Grounding reasoning framework for fine-grained image editing in complex real-world scenes. The key insight of InterCoG is to first perform object position reasoning solely within text that includes spatial relation details to explicitly deduce the location and identity of the edited target. It then conducts visual grounding via highlighting the editing targets with generated bounding boxes and masks in pixel space, and finally rewrites the editing description to specify the intended outcomes. To further facilitate this paradigm, we propose two auxiliary training modules: multimodal grounding reconstruction supervision and multimodal grounding reasoning alignment to enforce spatial localization accuracy and reasoning interpretability, respectively. We also construct GroundEdit-45K, a dataset comprising 45K grounding-oriented editing samples with detailed reasoning annotations, and GroundEdit-Bench for grounding-aware editing evaluation. Extensive experiments substantiate the superiority of our approach in highly precise edits under spatially intricate and multi-entity scenes.

cs.CV↗

arXiv2026

Tyche: One Step Flow for Efficient Probabilistic Weather Forecasting

Probabilistic weather forecasting requires not only accurate trajectories, but calibrated distributions over plausible atmospheric futures. Recent data-driven systems have achieved remarkable deterministic skill, and diffusion-based ensemble forecasters have substantially improved sample realism and uncertainty quantification. However, their inference cost scales with forecast horizon, ensemble size, and the number of denoising steps required for each transition, making large operational ensembles expensive. To address this, we present Tyche, a one-step conditional flow model for efficient probabilistic weather forecasting. Tyche models the conditional forecast distribution with a destination-aware average-velocity flow that maps Gaussian noise directly to future weather states in a single function evaluation (1-NFE). To make this one-step transport learnable in high-dimensional geophysical fields, we derive a JVP-regularized rectification objective that enforces temporal self-consistency across source and destination flow timesteps without explicitly forming Jacobians. The transport field is parameterized by an isotropic Swin-style transformer that preserves fine-scale spatial structure while remaining scalable on global grids. To improve ensemble reliability under autoregressive forecasting, we further introduce a rollout-based finetuning stage with curriculum CRPS calibration supervision. Experiments on ERA5 at 1.5$^\circ$ and 6-hour resolution show that our Tyche, using merely a single NFE, matches or exceeds the forecast skill and calibration of state-of-the-art multi-step generative baselines and the operational ECMWF IFS ensemble.

cs.LG↗

arXiv2026

eLLM: Elastic Memory Management Framework for Efficient LLM Serving

Large Language Models are increasingly being deployed in datacenters. Serving these models requires careful memory management, as their memory usage includes static weights, dynamic activations, and key-value caches. While static weights are constant and predictable, dynamic components such as activations and KV caches change frequently during runtime, presenting significant challenges for efficient memory management. Modern LLM serving systems typically handle runtime memory and KV caches at distinct abstraction levels: runtime memory management relies on static tensor abstractions, whereas KV caches utilize a page table-based virtualization layer built on top of the tensor abstraction. This virtualization dynamically manages KV caches to mitigate memory fragmentation. However, this dual-level approach fundamentally isolates runtime memory and KV cache management, resulting in suboptimal memory utilization under dynamic workloads, which can lead to a nearly 20% drop in throughput. To address these limitations, we propose eLLM, an elastic memory management framework inspired by the classical memory ballooning mechanism in operating systems. The core components of eLLM include: (1) Virtual Tensor Abstraction, which decouples the virtual address space of tensors from the physical GPU memory, creating a unified and flexible memory pool; (2) an Elastic Memory Mechanism that dynamically adjusts memory allocation through runtime memory inflation and deflation, leveraging CPU memory as an extensible buffer; and (3) a Lightweight Scheduling Strategy employing SLO-aware policies to optimize memory utilization and effectively balance performance trade-offs under stringent SLO constraints. Comprehensive evaluations demonstrate that eLLM significantly outperforms state-of-the-art systems, 2.32x higher decoding throughput, and supporting 3x larger batch sizes for 128K-token inputs.

cs.DC↗

arXiv2026

MHD modeling of magnetic flux evolution around solar maximum by the coronal model COCONUT

In this paper, we simulate the magnetic flux evolution at different heliocentric distances during two solar-maximum Carrington rotations (CRs) using the time-evolving coronal magnetohydrodynamic (MHD) model COCONUT to investigate the ``open flux problem". The simulated open magnetic flux (OMF) near the solar surface is comparable to that derived from \textit{in situ} observations by PSP and WIND satellites, and is about 5 times larger than that derived from SDO coronal hole (CH) observations, and the variation in the simulated radial solar wind speed is consistent with the evolution of the OMF evaluated around the corresponding solar disk center. We find that the OMF is reduced by up to $45\%$ from 1.01~$R_s$ to 0.1~AU and increases with a higher-resolution mesh. The OMF decreases mainly within 3~$R_s$, where the closed magnetic flux drops more rapidly, from about $60\%$ of the total magnetic flux at 1.01~$R_s$ to about $4\%$ at 3~$R_s$. Moderate adjustment of the heating source term can effectively regulate the simulated OMF. Preprocessing the photospheric magnetograms with a potential field solver that removes many high-order spherical harmonic components reduces the OMF in the low corona, while having little impact beyond 3~$R_s$. Additionally, the ratio of the maximum to the minimum OMF can reach 1.4 during a single solar maximum CR. These findings highlight the necessity of considering higher grid resolution, more realistic heating mechanisms, and the time-evolving regime of coronal MHD modeling when further addressing the ``open flux problem".

astro-ph.SR↗

arXiv2026

STARRY: Spatial-Temporal Action-Centric World Modeling for Robotic Manipulation

Robotic manipulation requires reasoning about future spatial-temporal interactions and geometric constraints, yet existing Vision-Language-Action (VLA) policies often leave predictive representation weakly coupled with action execution, causing failures in tasks requiring precise spatial-temporal coordination. We propose STARRY, a world-model-enhanced action-generation policy that aligns spatial-temporal prediction and action generation by jointly denoising future spatial-temporal latents and actions through a unified diffusion process. To bridge 2D visual tokens and 3D metric control, STARRY introduces Geometry-Aware Selective Attention Modulation (GASAM), which converts predicted depth and end-effector geometry into token-aligned weights for selective action-attention modulation. On RoboTwin 2.0, STARRY achieves 93.82% / 93.30% average success under Clean and Randomized settings across 50 bimanual tasks. Real-world experiments show that STARRY improves average success from 42.5% to 70.8% compared with $π_{0.5}$. These results demonstrate the effectiveness of action-centric spatial-temporal world modeling for spatially and temporally demanding robotic manipulation.

cs.RO↗

arXiv2026

AblateCell: A Reproduce-then-Ablate Agent for Virtual Cell Repositories

Systematic ablations are essential to attribute performance gains in AI Virtual Cells, yet they are rarely performed because biological repositories are under-standardized and tightly coupled to domain-specific data and formats. While recent coding agents can translate ideas into implementations, they typically stop at producing code and lack a verifier that can reproduce strong baselines and rigorously test which components truly matter. We introduce AblateCell, a reproduce-then-ablate agent for virtual cell repositories that closes this verification gap. AblateCell first reproduces reported baselines end-to-end by auto-configuring environments, resolving dependency and data issues, and rerunning official evaluations while emitting verifiable artifacts. It then conducts closed-loop ablation by generating a graph of isolated repository mutations and adaptively selecting experiments under a reward that trades off performance impact and execution cost. Evaluated on three single-cell perturbation prediction repositories (CPA, GEARS, BioLORD), AblateCell achieves 88.9% (+29.9% to human expert) end-to-end workflow success and 93.3% (+53.3% to heuristic) accuracy in recovering ground-truth critical components. These results enable scalable, repository-grounded verification and attribution directly on biological codebases.

cs.AI↗