arXiv subjects

Hao Xu

Publications and source records attributed to Hao Xu.

At least 163 records · Page 9Linked to original sources

arXiv2025

Camel: Energy-Aware LLM Inference on Resource-Constrained Devices

Most Large Language Models (LLMs) are currently deployed in the cloud, with users relying on internet connectivity for access. However, this paradigm faces challenges such as network latency, privacy concerns, and bandwidth limits. Thus, deploying LLMs on edge devices has become an important research focus. In edge inference, request latency is critical as high latency can impair real-time tasks. At the same time, edge devices usually have limited battery capacity, making energy consumption another major concern. Balancing energy consumption and inference latency is essential. To address this, we propose an LLM inference energy management framework that optimizes GPU frequency and batch size to balance latency and energy consumption. By effectively managing the exploration-exploitation dilemma in configuration search, the framework finds the optimal settings. The framework was implemented on the NVIDIA Jetson AGX Orin platform, and a series of experimental validations were conducted. Results demonstrate that, compared to the default configuration, our framework reduces energy delay product (EDP) by 12.4%-29.9%, achieving a better balance between energy consumption and latency.

cs.NI↗

arXiv2025

DreamPainter: Image Background Inpainting for E-commerce Scenarios

Although diffusion-based image genenation has been widely explored and applied, background generation tasks in e-commerce scenarios still face significant challenges. The first challenge is to ensure that the generated products are consistent with the given product inputs while maintaining a reasonable spatial arrangement, harmonious shadows, and reflections between foreground products and backgrounds. Existing inpainting methods fail to address this due to the lack of domain-specific data. The second challenge involves the limitation of relying solely on text prompts for image control, as effective integrating visual information to achieve precise control in inpainting tasks remains underexplored. To address these challenges, we introduce DreamEcom-400K, a high-quality e-commerce dataset containing accurate product instance masks, background reference images, text prompts, and aesthetically pleasing product images. Based on this dataset, we propose DreamPainter, a novel framework that not only utilizes text prompts for control but also flexibly incorporates reference image information as an additional control signal. Extensive experiments demonstrate that our approach significantly outperforms state-of-the-art methods, maintaining high product consistency while effectively integrating both text prompt and reference image information.

cs.CV↗

arXiv2025

Guided Depth Map Super-Resolution via Multi-Scale Fusion U-shaped Mamba Network

Depth map super-resolution technology aims to improve the spatial resolution of low-resolution depth maps and effectively restore high-frequency detail information. Traditional convolutional neural network has limitations in dealing with long-range dependencies and are unable to fully model the global contextual information in depth maps. Although transformer can model global dependencies, its computational complexity and memory consumption are quadratic, which significantly limits its ability to process high-resolution depth maps. In this paper, we propose a multi-scale fusion U-shaped Mamba (MSF-UM) model, a novel guided depth map super-resolution framework. The core innovation of this model is to integrate Mamba's efficient state-space modeling capabilities into a multi-scale U-shaped fusion structure guided by a color image. The structure combining the residual dense channel attention block and the Mamba state space module is designed, which combines the local feature extraction capability of the convolutional layer with the modeling advantage of the state space model for long-distance dependencies. At the same time, the model adopts a multi-scale cross-modal fusion strategy to make full use of the high-frequency texture information from the color image to guide the super-resolution process of the depth map. Compared with existing mainstream methods, the proposed MSF-UM significantly reduces the number of model parameters while achieving better reconstruction accuracy. Extensive experiments on multiple publicly available datasets validate the effectiveness of the model, especially showing excellent generalization ability in the task of large-scale depth map super-resolution.

cs.CV↗

arXiv2025

Hessian spectrum at the global minimum of the spherical pure-like mixed p-spin glasses

We study the large $N$-dimensional limit of the Hessian spectrum at the global minimum of some subclasses of the spherical mixed $p$-spin models. Specifically, we show that its empirical spectral measure converges in probability to a shifted and rescaled semicircle law and does not have outliers. Our method follows the second moment approach developed recently in \cite{BSZ20}, from which the ground state energy can be derived for the $pure$-$like$ mixed $p$-spin model. By analyzing the complexity function with given radial derivative and energy, we derive the convergence of the Hessian spectrum from the vanishing mean number of critical points. For the $1$-RSB model, the ground state energy was explicitly computed in \cite{huang2023constructive}. Combined with the complexity function of local maxima with given radial derivative obtained in \cite{belius2022complexity}, this allows us to obtain the corresponding results more directly. Our result extends those corresponding results in the regime of topology trivialization.

math-ph↗

arXiv2025

Achievable Rates for a Distributed Antenna System with No Channel State Information at the Central Processor

A recent trend in wireless communications considers the migration of traditional monolithic base stations to the so-called disaggregated architecture, where radio units (RUs) implement only the low-level physical layer functionalities such as demodulation, and A/D conversion, while the high-level physical layer, such as channel decoding, is implemented as software-defined functions running on general-purpose hardware in some remote central processing unit (CP). The corresponding information theoretic model for the uplink (from the wireless users to the CP) is a multiaccess-relay channel with primitive oblivious relays. The relays (RUs) are oblivious, as they are agnostic of the users codebooks, and primitive, since the fronthaul links (from RUs to CP) are error-free with limited capacity. This class of networks has been intensely studied in the information theoretic literature, where several approximated or exact (under certain conditions) capacity results have been derived. In particular, in the Gaussian case, the model has been analyzed for fixed and known channel state. This paper is motivated by the fact that, in practice, the channel state is a random process, and it is estimated at the base station side through uplink pilot symbols sent by the users. The pilot dimension may take up a large portion of the channel coherence block, i.e., the number of symbols over which the channel state remains approximately constant. Hence, sending both pilot and data symbols from the relays to the CP may require a significant overhead, especially when the fronthaul capacity is small. As a prototypical problem, we consider the ergodic achievable rate for a diamond network formed by a single user and two relays where the channel state is known at the relays, but not known at the CP.

cs.IT↗

arXiv2025

Momentum-Resolved Probing of Lorentz-Violating Dispersion Relations via Unruh-DeWitt Detector

Inspired by quantum gravity frameworks predicting Planck-scale deviations from Lorentz invariance, we probe Lorentz symmetry violation via modified dispersion relations $ω_{|\textbf{k}|}$. Departing from conventional approaches, we employ an Unruh-DeWitt detector to probe energy-dependent modifications to the dispersion relations. Two key methodological advances are introduced: (i) a generalized formulation for detector acceleration without assuming specific dispersion relations, and (ii) a momentum-resolved detection paradigm enabling spectral decomposition of $ω_{|\textbf{k}|}$ through localized momentum-shell integration. Analysis of deviations reveals disruption of the thermal spectrum under significant departures from the Lorentz invariance, while small perturbative regimes manifest as phase-modulated thermal distributions. By restricting detector-field interactions to narrow spectral windows and performing iterative Taylor expansions around reference momenta $|\textbf{k}_0|$, we derive coefficients encoding derivatives of $ω_{|\textbf{k}|}$, reconstructing its global profile via momentum-space tomography. Our approach offers a scalable method to test Lorentz symmetry violation across energy scales, and establishes a foundation for experimental verification of Planck-scale relics through high-precision spectral measurements.

gr-qc↗

arXiv2025

Disorder-induced spin excitation continuum and spin-glass ground state in the inverse spinel CuGa$_2$O$_4$

Spinel-structured compounds serve as prototypical examples of highly frustrated systems, and are promising candidates for realizing the long-sought quantum spin liquid (QSL) state. However, structural disorder is inevitable in many real QSL candidates and its impact remains a topic of intense debate. In this work, we conduct comprehensive investigations on CuGa$_2$O$_4$, a spinel compound with significant structural disorder, focusing on its thermodynamic properties and spectroscopic behaviors. No long-range magnetic order is observed down to $\sim$80 mK, as evidenced by magnetic susceptibility, specific heat and elastic neutron scattering measurements. More intriguingly, inelastic neutron scattering experiments reveal a broad gapless continuum of magnetic excitations around the Brillouin zone boundary, resembling the magnetic excitation spectra expected for a QSL. Nevertheless, a spin-freezing transition at $T_{\rm{f}} \approx $ 0.88 K is identified from the cusp in the dc susceptibility curves, where a bifurcation between zero-field-cooling and field-cooling curves occurs. Furthermore, ac susceptibility measurements show a peak close to $T_{\rm{f}}$ at low frequency, which shifts to higher temperature with increasing frequency. These results are evident that CuGa$_2$O$_4$ has spin-glass ground state, consistent with the establishment of short-range order inferred from the specific heat measurements. Collectively, these results illustrate the crucial role of disorder in defining the excitation spectrum out of the disordered ground state. Our findings shed light onto the broader class of AB$_2$O$_4$ spinels and advance our understanding of the spin dynamics in magnetically disordered systems.

cond-mat.str-el↗

arXiv2025

Transformer based Collaborative Reinforcement Learning for Fluid Antenna System (FAS)-enabled 3D UAV Positioning

In this paper, a novel Three dimensional (3D) positioning framework of fluid antenna system (FAS)-enabled unmanned aerial vehicles (UAVs) is developed. In the proposed framework, a set of controlled UAVs cooperatively estimate the real-time 3D position of a target UAV. Here, the active UAV transmits a measurement signal to the passive UAVs via the reflection from the target UAV. Each passive UAV estimates the distance of the active-target-passive UAV link and selects an antenna port to share the distance information with the base station (BS) that calculates the real-time position of the target UAV. As the target UAV is moving due to its task operation, the controlled UAVs must optimize their trajectories and select optimal antenna port, aiming to estimate the real-time position of the target UAV. We formulate this problem as an optimization problem to minimize the target UAV positioning error via optimizing the trajectories of all controlled UAVs and antenna port selection of passive UAVs. Here, an attention-based recurrent multi-agent reinforcement learning (AR-MARL) scheme is proposed, which enables each controlled UAV to use the local Q function to determine its trajectory and antenna port while optimizing the target UAV positioning performance without knowing the trajectories and antenna port selections of other controlled UAVs. Different from current MARL methods, the proposed method uses a recurrent neural network (RNN) that incorporates historical state-action pairs of each controlled UAV, and an attention mechanism to analyze the importance of these historical state-action pairs, thus improving the global Q function approximation accuracy and the target UAV positioning accuracy. Simulation results show that the proposed AR-MARL scheme can reduce the average positioning error by up to 17.5% and 58.5% compared to the VD-MARL scheme and the proposed method without FAS.

cs.NI↗

arXiv2025

3DPortraitGAN: Learning One-Quarter Headshot 3D GANs from a Single-View Portrait Dataset with Diverse Body Poses

3D-aware face generators are typically trained on 2D real-life face image datasets that primarily consist of near-frontal face data, and as such, they are unable to construct one-quarter headshot 3D portraits with complete head, neck, and shoulder geometry. Two reasons account for this issue: First, existing facial recognition methods struggle with extracting facial data captured from large camera angles or back views. Second, it is challenging to learn a distribution of 3D portraits covering the one-quarter headshot region from single-view data due to significant geometric deformation caused by diverse body poses. To this end, we first create the dataset 360°-Portrait-HQ (360°PHQ for short) which consists of high-quality single-view real portraits annotated with a variety of camera parameters (the yaw angles span the entire 360° range) and body poses. We then propose 3DPortraitGAN, the first 3D-aware one-quarter headshot portrait generator that learns a canonical 3D avatar distribution from the 360°PHQ dataset with body pose self-learning. Our model can generate view-consistent portrait images from all camera angles with a canonical one-quarter headshot 3D representation. Our experiments show that the proposed framework can accurately predict portrait body poses and generate view-consistent, realistic portrait images with complete geometry from all camera angles.

cs.CV↗

arXiv2025

OmniPart: Part-Aware 3D Generation with Semantic Decoupling and Structural Cohesion

The creation of 3D assets with explicit, editable part structures is crucial for advancing interactive applications, yet most generative methods produce only monolithic shapes, limiting their utility. We introduce OmniPart, a novel framework for part-aware 3D object generation designed to achieve high semantic decoupling among components while maintaining robust structural cohesion. OmniPart uniquely decouples this complex task into two synergistic stages: (1) an autoregressive structure planning module generates a controllable, variable-length sequence of 3D part bounding boxes, critically guided by flexible 2D part masks that allow for intuitive control over part decomposition without requiring direct correspondences or semantic labels; and (2) a spatially-conditioned rectified flow model, efficiently adapted from a pre-trained holistic 3D generator, synthesizes all 3D parts simultaneously and consistently within the planned layout. Our approach supports user-defined part granularity, precise localization, and enables diverse downstream applications. Extensive experiments demonstrate that OmniPart achieves state-of-the-art performance, paving the way for more interpretable, editable, and versatile 3D content.

cs.CV↗

arXiv2025

A Deep Unfolding Framework for Diffractive Snapshot Spectral Imaging

Snapshot hyperspectral imaging systems acquire spectral data cubes through compressed sensing. Recently, diffractive snapshot spectral imaging (DSSI) methods have attracted significant attention. While various optical designs and improvements continue to emerge, research on reconstruction algorithms remains limited. Although numerous networks and deep unfolding methods have been applied on similar tasks, they are not fully compatible with DSSI systems because of their distinct optical encoding mechanism. In this paper, we propose an efficient deep unfolding framework for diffractive systems, termed diffractive deep unfolding (DDU). Specifically, we derive an analytical solution for the data fidelity term in DSSI, ensuring both the efficiency and the effectiveness during the iterative reconstruction process. Given the severely ill-posed nature of the problem, we employ a network-based initialization strategy rather than non-learning-based methods or linear layers, leading to enhanced stability and performance. Our framework demonstrates strong compatibility with existing state-of-the-art (SOTA) models, which effectively address the initialization and prior subproblem. Extensive experiments validate the superiority of the proposed DDU framework, showcasing improved performance while maintaining comparable parameter counts and computational complexity. These results suggest that DDU provides a solid foundation for future unfolding-based methods in DSSI.

eess.IV↗

arXiv2025

Subpixel correction of diffraction pattern shifts in ptychography via automatic differentiation

Ptychography, a coherent diffraction imaging technique, has become an indispensable tool in materials characterization, biological imaging, and nanostructure analysis due to its capability for high-resolution, lensless reconstruction of complex-valued images. In typical workflows, raw diffraction patterns are commonly cropped to isolate the valid central region before reconstruction. However, if the crop is misaligned from the diffraction pattern's zero-order, reconstruction may suffer from slower convergence, phase wrapping, and reduced image fidelity. These issues are further exacerbated in experimental configurations involving reflective geometries or broadband illumination, where incorrect cropping introduces systematic preprocessing errors that compromise the entire ptychographic inversion. To address this challenge, we present an approach based on automatic differentiation (AD), where the cropping shift is treated as an optimizable parameter within the reconstruction framework. By integrating shift correction into the backpropagation loop, our method simultaneously refines the object, probe, and shift positions without requiring manual tuning. Simulation results demonstrate that, even with initial offsets ranging up to 5 pixels, the proposed method achieves subpixel correction, with an average deviation below 0.5 pixels. Experiments in the extreme ultraviolet (EUV) regime further validate the method's robustness and effectiveness. This AD-based strategy enhances the automation and robustness of ptychographic reconstructions, and is adaptable to diverse experimental conditions.

physics.optics↗

arXiv2025

Specific Absorption Rate-Aware Multiuser MIMO Assisted by Fluid Antenna System

With the development of the upcoming sixth-generation (6G) wireless networks, there is a pressing need for innovative technologies capable of satisfying heightened performance indicators. Fluid antenna system (FAS) is proposed recently as a promising technique to achieve higher data rates and more diversity gains by dynamically changing the positions of the antennas to form a more desirable channel. However, worries regarding the possibly harmful effects of electromagnetic (EM) radiation emitted by devices have arisen as a result of the rapid evolution of advanced techniques in wireless communication systems. Specific absorption rate (SAR) is a widely adopted metric to quantify EM radiation worldwide. In this paper, we investigate the SAR-aware multiuser multiple-input multiple-output (MIMO) communications assisted by FAS. In particular, a two-layer iterative algorithm is proposed to minimize the SAR value under signal-to-interference-plus-noise ratio (SINR) and FAS constraints. Moreover, the minimum weighted SINR maximization problem under SAR and FAS constraints is studied by finding its relationship with the SAR minimization problem. Simulation results verify that the proposed SAR-aware FAS design outperforms the adaptive backoff and fixed-position antenna designs.

eess.SP↗

arXiv2025

Advanced Deep Learning Techniques for Automated Segmentation of Type B Aortic Dissections

Purpose: Aortic dissections are life-threatening cardiovascular conditions requiring accurate segmentation of true lumen (TL), false lumen (FL), and false lumen thrombosis (FLT) from CTA images for effective management. Manual segmentation is time-consuming and variable, necessitating automated solutions. Materials and Methods: We developed four deep learning-based pipelines for Type B aortic dissection segmentation: a single-step model, a sequential model, a sequential multi-task model, and an ensemble model, utilizing 3D U-Net and Swin-UnetR architectures. A dataset of 100 retrospective CTA images was split into training (n=80), validation (n=10), and testing (n=10). Performance was assessed using the Dice Coefficient and Hausdorff Distance. Results: Our approach achieved superior segmentation accuracy, with Dice Coefficients of 0.91 $\pm$ 0.07 for TL, 0.88 $\pm$ 0.18 for FL, and 0.47 $\pm$ 0.25 for FLT, outperforming Yao et al. (1), who reported 0.78 $\pm$ 0.20, 0.68 $\pm$ 0.18, and 0.25 $\pm$ 0.31, respectively. Conclusion: The proposed pipelines provide accurate segmentation of TBAD features, enabling derivation of morphological parameters for surveillance and treatment planning

eess.IV↗

arXiv2025

Ancient Script Image Recognition and Processing: A Review

Ancient scripts, e.g., Egyptian hieroglyphs, Oracle Bone Inscriptions, and Ancient Greek inscriptions, serve as vital carriers of human civilization, embedding invaluable historical and cultural information. Automating ancient script image recognition has gained importance, enabling large-scale interpretation and advancing research in archaeology and digital humanities. With the rise of deep learning, this field has progressed rapidly, with numerous script-specific datasets and models proposed. While these scripts vary widely, spanning phonographic systems with limited glyphs to logographic systems with thousands of complex symbols, they share common challenges and methodological overlaps. Moreover, ancient scripts face unique challenges, including imbalanced data distribution and image degradation, which have driven the development of various dedicated methods. This survey provides a comprehensive review of ancient script image recognition methods. We begin by categorizing existing studies based on script types and analyzing respective recognition methods, highlighting both their differences and shared strategies. We then focus on challenges unique to ancient scripts, systematically examining their impact and reviewing recent solutions, including few-shot learning and noise-robust techniques. Finally, we summarize current limitations and outline promising future directions. Our goal is to offer a structured, forward-looking perspective to support ongoing advancements in the recognition, interpretation, and decipherment of ancient scripts.

cs.CV↗

arXiv2025

Towards Fine-Grained Emotion Understanding via Skeleton-Based Micro-Gesture Recognition

We present our solution to the MiGA Challenge at IJCAI 2025, which aims to recognize micro-gestures (MGs) from skeleton sequences for the purpose of hidden emotion understanding. MGs are characterized by their subtlety, short duration, and low motion amplitude, making them particularly challenging to model and classify. We adopt PoseC3D as the baseline framework and introduce three key enhancements: (1) a topology-aware skeleton representation specifically designed for the iMiGUE dataset to better capture fine-grained motion patterns; (2) an improved temporal processing strategy that facilitates smoother and more temporally consistent motion modeling; and (3) the incorporation of semantic label embeddings as auxiliary supervision to improve the model generalization. Our method achieves a Top-1 accuracy of 67.01\% on the iMiGUE test set. As a result of these contributions, our approach ranks third on the official MiGA Challenge leaderboard. The source code is available at \href{https://github.com/EGO-False-Sleep/Miga25_track1}{https://github.com/EGO-False-Sleep/Miga25\_track1}.

cs.CV↗

arXiv2025

Fast Point Cloud Geometry Compression with Context-based Residual Coding and INR-based Refinement

Compressing a set of unordered points is far more challenging than compressing images/videos of regular sample grids, because of the difficulties in characterizing neighboring relations in an irregular layout of points. Many researchers resort to voxelization to introduce regularity, but this approach suffers from quantization loss. In this research, we use the KNN method to determine the neighborhoods of raw surface points. This gives us a means to determine the spatial context in which the latent features of 3D points are compressed by arithmetic coding. As such, the conditional probability model is adaptive to local geometry, leading to significant rate reduction. Additionally, we propose a dual-layer architecture where a non-learning base layer reconstructs the main structures of the point cloud at low complexity, while a learned refinement layer focuses on preserving fine details. This design leads to reductions in model complexity and coding latency by two orders of magnitude compared to SOTA methods. Moreover, we incorporate an implicit neural representation (INR) into the refinement layer, allowing the decoder to sample points on the underlying surface at arbitrary densities. This work is the first to effectively exploit content-aware local contexts for compressing irregular raw point clouds, achieving high rate-distortion performance, low complexity, and the ability to function as an arbitrary-scale upsampling network simultaneously.

cs.CV↗

arXiv2025

Spectral Efficiency Maximization for DMA-enabled Multiuser MISO with Statistical CSI

Dynamic metasurface antennas (DMAs) offer the potential to achieve large-scale antenna arrays with low power consumption and reduced hardware costs, making them a promising technology for future communication systems. This paper investigates the spectral efficiency (SE) of DMA-enabled multiuser multiple-input single-output (MISO) systems in both uplink and downlink transmissions, using only statistical channel state information (CSI) to maximize the ergodic sum rate of multiple users. For the uplink system, we consider two decoding rules: minimum mean square error (MMSE) with and without successive interference cancellation (SIC). For both decoders, we derive closed-form surrogates to substitute the original expressions of ergodic sum rate and formulate tractable optimization problems for designing DMA weights. Then, a weighted MMSE (WMMSE)-based algorithm is proposed to maximize the ergodic sum rate. For the downlink system, we derive an approximate expression for the ergodic sum rate and formulate a hybrid analog/digital beamforming optimization problem that jointly optimizes the digital precoder and DMA weights. A penalty dual decomposition (PDD)-based algorithm is proposed by leveraging the fractional programming framework. Numerical results validate the accuracy of the derived surrogates and highlight the superiority of the proposed algorithms over baseline schemes. It is shown that these algorithms are effective across various DMA settings and are particularly well-suited for system design in fast time-varying channels.

cs.IT↗