arXiv subjects

Wei Huang

Publications and source records attributed to Wei Huang.

At least 163 records · Page 9Linked to original sources

arXiv2025

Rate Optimization for Downlink URLLC via Pinching Antenna Arrays

This work studies an ultra-reliable and low-latency communications (uRLLC) downlink system using pinching antennas which are realized by activating small dielectric particles along a dielectric waveguide. Our goal is to maximize the data rate by optimizing the positions of the pinching antennas. By proposing a compact and cost-efficient antenna architecture and formulating a finite blocklength-based optimization model, we derive a closed-form solution for the optimal antenna placement under quality-of-service (QoS) and antenna spacing constraints. Meanwhile, a phase-alignment strategy is integrated into the design, enabling coherent signal superposition across the array. Simulation results confirm significant rate improvements over conventional antenna systems while satisfying uRLLC requirements, making the proposed design well-suited for compact and latency-critical future applications.

eess.SP↗

arXiv2025

DPF-CM: A Data Processing Framework with Privacy-Preserving Vector Databases for Chinese Medical LLMs Training and Deployment

Current open-source training pipelines for Chinese medical language models predominantly emphasize optimizing training methodologies to enhance the performance of large language models (LLMs), yet lack comprehensive exploration into training data processing. To address this gap, we propose DPF-CM, a holistic Data Processing Framework for Chinese Medical LLMs training and deployment. DPF-CM comprises two core modules. The first module is a data processing pipeline tailored for model training. Beyond standard data processing operations, we (1) introduce a chained examples context-learning strategy to generate question-oriented instructions to mitigate the lack of instruction content, and (2) implement an ensemble-based filtering mechanism for preference data curation that averages multiple reward models to suppress noisy samples. The second module focuses on privacy preservation during model deployment. To prevent privacy risks from the inadvertent exposure of training data, we propose a Privacy Preserving Vector Database (PPVD) approach, which involves model memory search, high-risk database construction, secure database construction, and match-and-replace, four key stages to minimize privacy leakage during inference collectively. Experimental results show that DPF-CM significantly improves model accuracy, enabling our trained Chinese medical LLM to achieve state-of-the-art performance among open-source counterparts. Moreover, the framework reduces training data privacy leakage by 27%.

cs.LG↗

arXiv2025

Activation Steering Meets Preference Optimization: Defense Against Jailbreaks in Vision Language Models

Vision Language Models (VLMs) have demonstrated impressive capabilities in integrating visual and textual information for understanding and reasoning, but remain highly vulnerable to adversarial attacks. While activation steering has emerged as a promising defence, existing approaches often rely on task-specific contrastive prompts to extract harmful directions, which exhibit suboptimal performance and can degrade visual grounding performance. To address these limitations, we propose \textit{Sequence-Level Preference Optimization} for VLM (\textit{SPO-VLM}), a novel two-stage defense framework that combines activation-level intervention with policy-level optimization to enhance model robustness. In \textit{Stage I}, we compute adaptive layer-specific steering vectors from diverse data sources, enabling generalized suppression of harmful behaviors during inference. In \textit{Stage II}, we refine these steering vectors through a sequence-level preference optimization process. This stage integrates automated toxicity assessment, as well as visual-consistency rewards based on caption-image alignment, to achieve safe and semantically grounded text generation. The two-stage structure of SPO-VLM balances efficiency and effectiveness by combining a lightweight mitigation foundation in Stage I with deeper policy refinement in Stage II. Extensive experiments shown SPO-VLM enhances safety against attacks via activation steering and preference optimization, while maintaining strong performance on benign tasks without compromising visual understanding capabilities. We will release our code, model weights, and evaluation toolkit to support reproducibility and future research. \textcolor{red}{Warning: This paper may contain examples of offensive or harmful text and images.}

cs.CV↗

arXiv2025

Energy Detection over Composite $κ-μ$ Shadowed Fading Channels with Inverse Gaussian Distribution in Ultra mMTC Networks

This paper investigates the characteristics of energy detection (ED) over composite $κ$-$μ$ shadowed fading channels in ultra machine-type communication (mMTC) networks. We have derived the closed-form expressions of the probability density function (PDF) of signal-to-noise ratio (SNR) based on the Inverse Gaussian (\emph{IG}) distribution. By adopting novel integration and mathematical transformation techniques, we derive a truncation-based closed-form expression for the average detection probability for the first time. It can be observed from our simulations that the number of propagation paths has a more pronounced effect on average detection probability compared to average SNR, which is in contrast to earlier studies that focus on device-to-device networks. It suggests that for 6G mMTC network design, we should consider enhancing transmitter-receiver placement and antenna alignment strategies, rather than relying solely on increasing the device-to-device average SNR.

eess.SP↗

arXiv2025

Intelligent Shanghai Typhoon Model (ISTM): A generative probabilistic emulator for typhoon hybrid modeling

To address the systematic underestimation of typhoon intensity in artificial intelligence weather prediction (AIWP) models, we propose the Intelligent Shanghai Typhoon Model (ISTM): a unified regional-to-typhoon generative probabilistic forecasting system based on a two-stage UNet-Diffusion framework. ISTM learns a downscaling mapping from 4 years of 25 km ERA5 reanalysis to a 9 km high resolution typhoon reanalysis dataset, enabling the generation of kilometer-scale near-surface variables and maximum radar reflectivity from coarse resolution fields. The evaluation results show that the two-stage UNet-Diffusion model significantly outperforms both ERA5 and the baseline UNet regression in capturing the structure and intensity of surface winds and precipitation. After fine-tuning, ISTM can effectively map AIFS forecasts, an advanced AIWP model, to high-resolution forecasts from AI-physics hybrid Shanghai Typhoon Model, substantially enhancing typhoon intensity predictions while preserving track accuracy. This positions ISTM as an efficient AI emulator of hybrid modeling system, achieving fast and physically consistent downscaling. The proposed framework establishes a unified pathway for the co-evolution of AIWP and physics-based numerical models, advancing next-generation typhoon forecasting capabilities.

physics.ao-ph↗

arXiv2025

Data Pruning by Information Maximization

In this paper, we present InfoMax, a novel data pruning method, also known as coreset selection, designed to maximize the information content of selected samples while minimizing redundancy. By doing so, InfoMax enhances the overall informativeness of the coreset. The information of individual samples is measured by importance scores, which capture their influence or difficulty in model learning. To quantify redundancy, we use pairwise sample similarities, based on the premise that similar samples contribute similarly to the learning process. We formalize the coreset selection problem as a discrete quadratic programming (DQP) task, with the objective of maximizing the total information content, represented as the sum of individual sample contributions minus the redundancies introduced by similar samples within the coreset. To ensure practical scalability, we introduce an efficient gradient-based solver, complemented by sparsification techniques applied to the similarity matrix and dataset partitioning strategies. This enables InfoMax to seamlessly scale to datasets with millions of samples. Extensive experiments demonstrate the superior performance of InfoMax in various data pruning tasks, including image classification, vision-language pre-training, and instruction tuning for large language models. Code is available at https://github.com/hrtan/InfoMax.

cs.CV↗

arXiv2025

Provably Transformers Harness Multi-Concept Word Semantics for Efficient In-Context Learning

Transformer-based large language models (LLMs) have displayed remarkable creative prowess and emergence capabilities. Existing empirical studies have revealed a strong connection between these LLMs' impressive emergence abilities and their in-context learning (ICL) capacity, allowing them to solve new tasks using only task-specific prompts without further fine-tuning. On the other hand, existing empirical and theoretical studies also show that there is a linear regularity of the multi-concept encoded semantic representation behind transformer-based LLMs. However, existing theoretical work fail to build up an understanding of the connection between this regularity and the innovative power of ICL. Additionally, prior work often focuses on simplified, unrealistic scenarios involving linear transformers or unrealistic loss functions, and they achieve only linear or sub-linear convergence rates. In contrast, this work provides a fine-grained mathematical analysis to show how transformers leverage the multi-concept semantics of words to enable powerful ICL and excellent out-of-distribution ICL abilities, offering insights into how transformers innovate solutions for certain unseen tasks encoded with multiple cross-concept semantics. Inspired by empirical studies on the linear latent geometry of LLMs, the analysis is based on a concept-based low-noise sparse coding prompt model. Leveraging advanced techniques, this work showcases the exponential 0-1 loss convergence over the highly non-convex training dynamics, which pioneeringly incorporates the challenges of softmax self-attention, ReLU-activated MLPs, and cross-entropy loss. Empirical simulations corroborate the theoretical findings.

cs.LG↗

arXiv2025

Provable In-Context Vector Arithmetic via Retrieving Task Concepts

In-context learning (ICL) has garnered significant attention for its ability to grasp functions/tasks from demonstrations. Recent studies suggest the presence of a latent task/function vector in LLMs during ICL. Merullo et al. (2024) showed that LLMs leverage this vector alongside the residual stream for Word2Vec-like vector arithmetic, solving factual-recall ICL tasks. Additionally, recent work empirically highlighted the key role of Question-Answer data in enhancing factual-recall capabilities. Despite these insights, a theoretical explanation remains elusive. To move one step forward, we propose a theoretical framework building on empirically grounded hierarchical concept modeling. We develop an optimization theory, showing how nonlinear residual transformers trained via gradient descent on cross-entropy loss perform factual-recall ICL tasks via vector arithmetic. We prove 0-1 loss convergence and show the strong generalization, including robustness to concept recombination and distribution shifts. These results elucidate the advantages of transformers over static embedding predecessors. Empirical simulations corroborate our theoretical insights.

cs.LG↗

arXiv2025

Hybrid Generative Semantic and Bit Communications in Satellite Networks: Trade-offs in Latency, Generation Quality, and Computation

As satellite communications play an increasingly important role in future wireless networks, the issue of limited link budget in satellite systems has attracted significant attention in current research. Although semantic communications emerge as a promising solution to address these constraints, it introduces the challenge of increased computational resource consumption in wireless communications. To address these challenges, we propose a multi-layer hybrid bit and generative semantic communication framework which can adapt to the dynamic satellite communication networks. Furthermore, to balance the semantic communication efficiency and performance in satellite-to-ground transmissions, we introduce a novel semantic communication efficiency metric (SEM) that evaluates the trade-offs among latency, computational consumption, and semantic reconstruction quality in the proposed framework. Moreover, we utilize a novel deep reinforcement learning (DRL) algorithm group relative policy optimization (GRPO) to optimize the resource allocation in the proposed network. Simulation results demonstrate the flexibility of our proposed transmission framework and the effectiveness of the proposed metric SEM, illustrate the relationships among various semantic communication metrics.

cs.IT↗

arXiv2025

WOCD: A Semi-Supervised Method for Overlapping Community Detection Using Weak Cliques

Overlapping community detection (OCD) is a fundamental graph data analysis task for extracting graph patterns. Traditional OCD methods can be broadly divided into node clustering and link clustering approaches, both of which rely solely on link information to identify overlapping communities. In recent years, deep learning-based methods have made significant advancements for this task. However, existing GNN-based approaches often face difficulties in effectively integrating link, attribute, and prior information, along with challenges like limited receptive fields and over-smoothing, which hinder their performance on complex overlapping community detection. In this paper, we propose a Weak-clique based Overlapping Community Detection method, namely WOCD, which incorporates prior information and optimizes the use of link information to improve detection accuracy. Specifically, we introduce pseudo-labels within a semi-supervised framework to strengthen the generalization ability, making WOCD more versatile. Furthermore, we initialize pseudo-labels using weak cliques to fully leverage link and prior information, leading to better detection accuracy. Additionally, we employ a single-layer Graph Transformer combined with GNN, which achieves significant performance improvements while maintaining efficiency. We evaluate WOCD on eight real-world attributed datasets, and the results demonstrate that it outperforms the state-of-the-art semi-supervised OCD method by a significant margin in terms of accuracy.

cs.SI↗

arXiv2025

GaRe: Relightable 3D Gaussian Splatting for Outdoor Scenes from Unconstrained Photo Collections

We propose a 3D Gaussian splatting-based framework for outdoor relighting that leverages intrinsic image decomposition to precisely integrate sunlight, sky radiance, and indirect lighting from unconstrained photo collections. Unlike prior methods that compress the per-image global illumination into a single latent vector, our approach enables simultaneously diverse shading manipulation and the generation of dynamic shadow effects. This is achieved through three key innovations: (1) a residual-based sun visibility extraction method to accurately separate direct sunlight effects, (2) a region-based supervision framework with a structural consistency loss for physically interpretable and coherent illumination decomposition, and (3) a ray-tracing-based technique for realistic shadow simulation. Extensive experiments demonstrate that our framework synthesizes novel views with competitive fidelity against state-of-the-art relighting solutions and produces more natural and multifaceted illumination and shadow effects.

cs.CV↗

arXiv2025

ND1 centers in diamond for long-term data storage in extreme conditions

Practically feasible long-term data storage under extreme conditions is an unsolved problem in modern data storage systems. This study introduces a novel approach using ND1 centers in diamonds for high-density, three-dimensional optical data storage. By employing near-infrared femtosecond laser pulses, we demonstrate the creation of sub-micron ND1 defect sites with precise spatial control, enabling efficient data encoding as luminescent ''pits." The ND1 centers exhibit robust photoluminescence in the UV spectrum, driven by three-photon absorption, which intrinsically provides a 3D reading of the data. Remarkably, these centers remain stable under extreme electric and magnetic fields, temperatures ranging from 4 K to 500 K, and corrosive chemical environments, with no degradation observed over extended periods. A reading speed of 500 MBits/s, limited by the lifetime of the photoluminescence, surpasses conventional Blu-ray technology while maintaining compatibility with existing optical data storage infrastructure. Our findings highlight diamond-based ND1 centers as a promising medium for durable, high-capacity data storage, capable of preserving critical information for millions of years, even under harsh conditions.

physics.optics↗

arXiv2025

EmbRACE-3K: Embodied Reasoning and Action in Complex Environments

Recent advanced vision-language models(VLMs) have demonstrated strong performance on passive, offline image and video understanding tasks. However, their effectiveness in embodied settings, which require online interaction and active scene understanding remains limited. In such scenarios, an agent perceives the environment from a first-person perspective, with each action dynamically shaping subsequent observations. Even state-of-the-art models such as GPT-4o, Claude 3.5 Sonnet, and Gemini 2.5 Pro struggle in open-environment interactions, exhibiting clear limitations in spatial reasoning and long-horizon planning. To address this gap, we introduce EmRACE-3K, a dataset of over 3,000 language-guided tasks situated in diverse, photorealistic environments constructed using Unreal Engine and the UnrealCV-Zoo framework. The tasks encompass a wide range of embodied challenges, including navigation, object manipulation, and multi-stage goal execution. Each task unfolds as a multi-step trajectory, pairing first-person visual observations with high-level instructions, grounded actions, and natural language rationales that express the agent's intent at every step. Using EmRACE-3K, we establish a benchmark to evaluate the embodied reasoning capabilities of VLMs across three key dimensions: Exploration, Dynamic Spatial-Semantic Reasoning, and Multi-stage Goal Execution. In zero-shot settings, all models achieve success rates below 20%, underscoring the challenge posed by our benchmark and the current limitations of VLMs in interactive environments. To demonstrate the utility of EmRACE-3K, we further fine-tune Qwen2.5-VL-7B using supervised learning followed by reinforcement learning. This approach yields substantial improvements across all three challenge categories, highlighting the dataset's effectiveness in enabling the development of embodied reasoning capabilities.

cs.CV↗

arXiv2025

2D Instance Editing in 3D Space

Generative models have achieved significant progress in advancing 2D image editing, demonstrating exceptional precision and realism. However, they often struggle with consistency and object identity preservation due to their inherent pixel-manipulation nature. To address this limitation, we introduce a novel "2D-3D-2D" framework. Our approach begins by lifting 2D objects into 3D representation, enabling edits within a physically plausible, rigidity-constrained 3D environment. The edited 3D objects are then reprojected and seamlessly inpainted back into the original 2D image. In contrast to existing 2D editing methods, such as DragGAN and DragDiffusion, our method directly manipulates objects in a 3D environment. Extensive experiments highlight that our framework surpasses previous methods in general performance, delivering highly consistent edits while robustly preserving object identity.

cs.CV↗

arXiv2025

A New and Efficient Debiased Estimation of General Treatment Models by Balanced Neural Networks Weighting

Estimation and inference of treatment effects under unconfounded treatment assignments often suffer from bias and the `curse of dimensionality' due to the nonparametric estimation of nuisance parameters for high-dimensional confounders. Although debiased state-of-the-art methods have been proposed for binary treatments under particular treatment models, they can be unstable for small sample sizes. Moreover, directly extending them to general treatment models can lead to computational complexity. We propose a balanced neural networks weighting method for general treatment models, which leverages deep neural networks to alleviate the curse of dimensionality while retaining optimal covariate balance through calibration, thereby achieving debiased and robust estimation. Our method accommodates a wide range of treatment models, including average, quantile, distributional, and asymmetric least squares treatment effects, for discrete, continuous, and mixed treatments. Under regularity conditions, we show that our estimator achieves rate double robustness and $\sqrt{N}$-asymptotic normality, and its asymptotic variance achieves the semiparametric efficiency bound. We further develop a statistical inference procedure based on weighted bootstrap, which avoids estimating the efficient influence/score functions. Simulation results reveal that the proposed method consistently outperforms existing alternatives, especially when the sample size is small. Applications to the 401(k) dataset and the Mother's Significant Features dataset further illustrate the practical value of the method for estimating both average and quantile treatment effects under binary and continuous treatments, respectively.

stat.ME↗

arXiv2025

A Multimodal In Vitro Diagnostic Method for Parkinson's Disease Combining Facial Expressions and Behavioral Gait Data

Parkinson's disease (PD), characterized by its incurable nature, rapid progression, and severe disability, poses significant challenges to the lives of patients and their families. Given the aging population, the need for early detection of PD is increasing. In vitro diagnosis has garnered attention due to its non-invasive nature and low cost. However, existing methods present several challenges: 1) limited training data for facial expression diagnosis; 2) specialized equipment and acquisition environments required for gait diagnosis, resulting in poor generalizability; 3) the risk of misdiagnosis or missed diagnosis when relying on a single modality. To address these issues, we propose a novel multimodal in vitro diagnostic method for PD, leveraging facial expressions and behavioral gait. Our method employs a lightweight deep learning model for feature extraction and fusion, aimed at improving diagnostic accuracy and facilitating deployment on mobile devices. Furthermore, we have established the largest multimodal PD dataset in collaboration with a hospital and conducted extensive experiments to validate the effectiveness of our proposed method.

cs.CV↗

arXiv2025

Radii of spherical timelike geodesics in Kerr-Newman black holes

The existence, radii and radial stability of the equatorial and non-equatorial (particularly, the polar) spherical orbits are discussed for particles with different conserved energy. The radii of these orbits generally are solutions of a quintic polynomial equation with four dimensionless parameters. For the case with $γ=1$, we obtain the analytical expressions for the radii of the polar, equatorial and general orbits. The radial stability of the orbits outside the event horizon is also discussed. In the $(u, w, β)$ space, a no-orbit surface is found. When the parameters lies on this surface there is no orbit outside the event horizon, otherwise there is always one spherical orbit outside the event horizon. For the cases with $γ\neq1$, we focus on the study of polar and equatorial orbits. For polar orbits with $0<γ<1$, a boundary surface in $(u, w, γ)$ space is identified which determines the existence of spherical polar orbits outside the event horizon. Numerical results of the radii and radial stability of the polar orbits are shown for examples with specific values of $γ$. For polar orbits with $γ>1$, it is found that there is always one unstable orbit outside the event horizon. For equatorial orbits with $0<γ<1$, in each rotating case (prograde case and retrograde case), a boundary surface in $(u, w, γ)$ space is also identified which divides the parameter space into two regions: one region with two orbits (one stable and the other unstable) and the other with no orbit outside the event horizon. Parameters on the boundary surface correspond to ISCOs. An analytical formula for the ISCOs is derived by choosing $(w,γ)$ as independent variables. For equatorial orbits with $γ>1$, it is found that there is always one unstable orbit outside the event horizon.

gr-qc↗

arXiv2025

Generalization Bound of Gradient Flow through Training Trajectory and Data-dependent Kernel

Gradient-based optimization methods have shown remarkable empirical success, yet their theoretical generalization properties remain only partially understood. In this paper, we establish a generalization bound for gradient flow that aligns with the classical Rademacher complexity bounds for kernel methods-specifically those based on the RKHS norm and kernel trace-through a data-dependent kernel called the loss path kernel (LPK). Unlike static kernels such as NTK, the LPK captures the entire training trajectory, adapting to both data and optimization dynamics, leading to tighter and more informative generalization guarantees. Moreover, the bound highlights how the norm of the training loss gradients along the optimization trajectory influences the final generalization performance. The key technical ingredients in our proof combine stability analysis of gradient flow with uniform convergence via Rademacher complexity. Our bound recovers existing kernel regression bounds for overparameterized neural networks and shows the feature learning capability of neural networks compared to kernel methods. Numerical experiments on real-world datasets validate that our bounds correlate well with the true generalization gap.

cs.LG↗