arXiv subjects

Hong Zhang

Publications and source records attributed to Hong Zhang.

At least 109 records · Page 6Linked to original sources

arXiv2025

TrackingMiM: Efficient Mamba-in-Mamba Serialization for Real-time UAV Object Tracking

The Vision Transformer (ViT) model has long struggled with the challenge of quadratic complexity, a limitation that becomes especially critical in unmanned aerial vehicle (UAV) tracking systems, where data must be processed in real time. In this study, we explore the recently proposed State-Space Model, Mamba, leveraging its computational efficiency and capability for long-sequence modeling to effectively process dense image sequences in tracking tasks. First, we highlight the issue of temporal inconsistency in existing Mamba-based methods, specifically the failure to account for temporal continuity in the Mamba scanning mechanism. Secondly, building upon this insight,we propose TrackingMiM, a Mamba-in-Mamba architecture, a minimal-computation burden model for handling image sequence of tracking problem. In our framework, the mamba scan is performed in a nested way while independently process temporal and spatial coherent patch tokens. While the template frame is encoded as query token and utilized for tracking in every scan. Extensive experiments conducted on five UAV tracking benchmarks confirm that the proposed TrackingMiM achieves state-of-the-art precision while offering noticeable higher speed in UAV tracking.

cs.CV↗

arXiv2025

SuperPlace: The Renaissance of Classical Feature Aggregation for Visual Place Recognition in the Era of Foundation Models

Recent visual place recognition (VPR) approaches have leveraged foundation models (FM) and introduced novel aggregation techniques. However, these methods have failed to fully exploit key concepts of FM, such as the effective utilization of extensive training sets, and they have overlooked the potential of classical aggregation methods, such as GeM and NetVLAD. Building on these insights, we revive classical feature aggregation methods and develop more fundamental VPR models, collectively termed SuperPlace. First, we introduce a supervised label alignment method that enables training across various VPR datasets within a unified framework. Second, we propose G$^2$M, a compact feature aggregation method utilizing two GeMs, where one GeM learns the principal components of feature maps along the channel dimension and calibrates the output of the other. Third, we propose the secondary fine-tuning (FT$^2$) strategy for NetVLAD-Linear (NVL). NetVLAD first learns feature vectors in a high-dimensional space and then compresses them into a lower-dimensional space via a single linear layer. Extensive experiments highlight our contributions and demonstrate the superiority of SuperPlace. Specifically, G$^2$M achieves promising results with only one-tenth of the feature dimensions compared to recent methods. Moreover, NVL-FT$^2$ ranks first on the MSLS leaderboard.

cs.CV↗

arXiv2025

EmbodiedPlace: Learning Mixture-of-Features with Embodied Constraints for Visual Place Recognition

Visual Place Recognition (VPR) is a scene-oriented image retrieval problem in computer vision in which re-ranking based on local features is commonly employed to improve performance. In robotics, VPR is also referred to as Loop Closure Detection, which emphasizes spatial-temporal verification within a sequence. However, designing local features specifically for VPR is impractical, and relying on motion sequences imposes limitations. Inspired by these observations, we propose a novel, simple re-ranking method that refines global features through a Mixture-of-Features (MoF) approach under embodied constraints. First, we analyze the practical feasibility of embodied constraints in VPR and categorize them according to existing datasets, which include GPS tags, sequential timestamps, local feature matching, and self-similarity matrices. We then propose a learning-based MoF weight-computation approach, utilizing a multi-metric loss function. Experiments demonstrate that our method improves the state-of-the-art (SOTA) performance on public datasets with minimal additional computational overhead. For instance, with only 25 KB of additional parameters and a processing time of 10 microseconds per frame, our method achieves a 0.9\% improvement over a DINOv2-based baseline performance on the Pitts-30k test set.

cs.CV↗

arXiv2025

Collimated Hard X-Rays from Hybrid Laser and Plasma Wakefield Accelerators

We report a synergistic enhancement of betatron radiation based on the hybrid laser and plasma wakefield acceleration scheme. Quasi-phase-stable acceleration in an up-ramp plasma density first generates GeV-energy electron beams that act as a drive beam for PWFA, which then further accelerates the witness beam to GeV energies, enhancing both photon energy and flux. A full width at half maximum divergence $(6.1 \pm 1.9)\times(5.8\pm 1.6) $ mrad$^2$ of betatron radiation, a critical energy of $71 \pm 8$ keV, and an average flux of more than $10^{14}$ photons per steradian above 5 keV were all experimentally obtained thanks to this scheme, which was an order of magnitude higher than the previous reports. Quasi-three-dimensional particle-in-cell simulations were used to model the acceleration and radiation of the electrons in our experimental conditions, establishing a new paradigm for compact collimated hard X-ray sources.

physics.plasm-ph↗

arXiv2025

SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning

Recent development in Large Language Models (LLMs) and Multi-modal Large Language Models (MLLMs) have leverage Attention-based Transformer architectures and achieved superior performance and generalization capabilities. They have since covered extensive areas of traditional learning tasks. For instance, text-based tasks such as text-classification and sequence-labeling, as well as multi-modal tasks like Visual Question Answering (VQA) and Optical Character Recognition (OCR), which were previously addressed using different models, can now be tackled based on one foundation model. Consequently, the training and lightweight fine-tuning of LLMs and MLLMs, especially those based on Transformer architecture, has become particularly important. In recognition of these overwhelming needs, we develop SWIFT, a customizable one-stop infrastructure for large models. With support of over $300+$ LLMs and $50+$ MLLMs, SWIFT stands as the open-source framework that provide the most comprehensive support for fine-tuning large models. In particular, it is the first training framework that provides systematic support for MLLMs. In addition to the core functionalities of fine-tuning, SWIFT also integrates post-training processes such as inference, evaluation, and model quantization, to facilitate fast adoptions of large models in various application scenarios. With a systematic integration of various training techniques, SWIFT offers helpful utilities such as benchmark comparisons among different training techniques for large models. For fine-tuning models specialized in agent framework, we show that notable improvements on the ToolBench leader-board can be achieved by training with customized dataset on SWIFT, with an increase of 5.2%-21.8% in the Act.EM metric over various baseline models, a reduction in hallucination by 1.6%-14.1%, and an average performance improvement of 8%-17%.

cs.CL↗

arXiv2025

Searching for True Muonium in Relativistic Heavy Ion Collisions

We investigate the production of the as-yet-undetected true muonium within the quark-gluon plasma formed in relativistic heavy-ion collisions, employing a relativistic Boltzmann transport framework coupled to viscous hydrodynamic simulations. The obtained effective cross sections for central collisions are 1.23~$μb$ in AuAu collisions with $\sqrt{s_{\rm NN}}=200$~GeV and 14.2~$μb$ in PbPb collisions with $\sqrt{s_{\rm NN}}=5.02$~TeV, resulting in a yield of $\mathcal{O}(10^4)$ and $\mathcal{O}(10^5)$ true muonium per billion $AA$ collisions at RHIC and the LHC, respectively. This establishes heavy-ion collisions as a promising process for detecting true muonium.

nucl-th↗

arXiv2025

Evolution and detection of vector superradiant instabilities

Ultralight vectors can extract energy and angular momentum from a Kerr black hole (BH) due to superradiant instability, resulting in the formation of a BH-condensate system. In this work, we carefully investigate the evolution of this system numerically with multiple superradiant modes. Simple formulas are obtained to estimate important timescales, maximum masses of different modes, as well as the BH mass and spin at various times. Due to the coexistence of modes with small frequency differences, the BH-condensate system emits gravitational waves with a unique beat signature, which could be directly observed by current and projected interferometers. Besides, the current BH spin-mass data from the binary BH merger events already exclude the vector mass in the range $5\times 10^{-15}\ \mathrm{eV} <μ< 9\times 10^{-12}\ \mathrm{eV}$.

gr-qc↗

arXiv2025

Constraint Multi-class Positive and Unlabeled Learning for Distantly Supervised Named Entity Recognition

Distantly supervised named entity recognition (DS-NER) has been proposed to exploit the automatically labeled training data by external knowledge bases instead of human annotations. However, it tends to suffer from a high false negative rate due to the inherent incompleteness. To address this issue, we present a novel approach called \textbf{C}onstraint \textbf{M}ulti-class \textbf{P}ositive and \textbf{U}nlabeled Learning (CMPU), which introduces a constraint factor on the risk estimator of multiple positive classes. It suggests that the constraint non-negative risk estimator is more robust against overfitting than previous PU learning methods with limited positive data. Solid theoretical analysis on CMPU is provided to prove the validity of our approach. Extensive experiments on two benchmark datasets that were labeled using diverse external knowledge sources serve to demonstrate the superior performance of CMPU in comparison to existing DS-NER methods.

cs.CL↗

arXiv2025

ReJSHand: Efficient Real-Time Hand Pose Estimation and Mesh Reconstruction Using Refined Joint and Skeleton Features

Accurate hand pose estimation is vital in robotics, advancing dexterous manipulation in human-computer interaction. Toward this goal, this paper presents ReJSHand (which stands for Refined Joint and Skeleton Features), a cutting-edge network formulated for real-time hand pose estimation and mesh reconstruction. The proposed framework is designed to accurately predict 3D hand gestures under real-time constraints, which is essential for systems that demand agile and responsive hand motion tracking. The network's design prioritizes computational efficiency without compromising accuracy, a prerequisite for instantaneous robotic interactions. Specifically, ReJSHand comprises a 2D keypoint generator, a 3D keypoint generator, an expansion block, and a feature interaction block for meticulously reconstructing 3D hand poses from 2D imagery. In addition, the multi-head self-attention mechanism and a coordinate attention layer enhance feature representation, streamlining the creation of hand mesh vertices through sophisticated feature mapping and linear transformation. Regarding performance, comprehensive evaluations on the FreiHand dataset demonstrate ReJSHand's computational prowess. It achieves a frame rate of 72 frames per second while maintaining a PA-MPJPE (Position-Accurate Mean Per Joint Position Error) of 6.3 mm and a PA-MPVPE (Position-Accurate Mean Per Vertex Position Error) of 6.4 mm. Moreover, our model reaches scores of 0.756 for F@05 and 0.984 for F@15, surpassing modern pipelines and solidifying its position at the forefront of robotic hand pose estimators. To facilitate future studies, we provide our source code at ~\url{https://github.com/daishipeng/ReJSHand}.

cs.RO↗

arXiv2025

Mechanism of Electricacupuncture Treating Detrusor Bladder Neck Dyscoordination After Suprasacral Spinal Cord Injury by Proteomics

Objectives This study aimed to elucidate the potential mechanisms of electroacupuncture (EA) in restoring detrusor-bladder neck dyssynergesia (DBND) following suprasacral spinal cord injury. Methods A total of 52 adult female Sprague-Dawley rats were randomly assigned to either a sham group (n=12) or a spinal cord injury model group (n=40). In the model group, DBND was induced in 40 rats through Hassan Shaker spinal cord transection, with 24 rats surviving spinal shock and subsequently randomized into two groups: a model-only group (DBND, n=12) and an EA intervention group (DBND+EA, n=12). DBND+EA was administered at Ciliao (BL32), Zhongji (RN3), and Sanyinjiao (SP6) acupoints, for 20 minutes per session, once daily for 10 consecutive days. On day 29 post-injury, all rats underwent urodynamic assessments, followed by hematoxylin and eosin (HE) staining, tandem mass tag (TMT) proteomics, and Western blot (WB) analysis of the detrusor and bladder neck tissues. Results Urodynamic evaluation demonstrated that EA intervention enhanced bladder function in DBND rats. HE staining indicated reduced fibroplasia in the detrusor muscle and alleviated inflammation in the bladder neck following EA. TMT proteomic analysis revealed 30 differentially expressed proteins (DEPs) in the detrusor and 59 DEPs in the bladder neck post-EA treatment. WB results corroborated these TMT findings. Conclusion EA effectively promotes synergy between the detrusor muscle and bladder neck in DBND, likely by enhancing detrusor contractility and facilitating bladder neck relaxation during urination. This study provides mechanistic insights into the therapeutic role of EA in managing DBND.

q-bio.BM↗

arXiv2025

Chemical Continuous Time Random Walks under Anomalous Diffusion

Chemical master equation plays an important role to describe the time evolution of homogeneous chemical system. In addition to the reaction process, it is also accompanied by physical diffusion of the reactants in complex system that is generally not homogeneous, which will result in non-exponential waiting times for particle reactions and diffusion. In this paper we shall introduce a chemical continuous time random walk under anomalous diffusion model based on renewal process to describe the general reaction-diffusion process in the heterogeneous system, where the waiting times are arbitrary distributed. According to this model, we will develop the systematic stochastic theory including generalizing the chemical diffusion master equation, deriving the corresponding mass action law, and extending the Gillespie algorithm. As an example, we analyze the monomolecular $A\leftrightarrow B$ reaction-diffusion system for exponential and power-law waiting times respectively, and show the strong fractional memory effect of the concentration of the reactants on the history of the concentration in power-law case.

physics.chem-ph↗

arXiv2025

FlowPlan: Zero-Shot Task Planning with LLM Flow Engineering for Robotic Instruction Following

Robotic instruction following tasks require seamless integration of visual perception, task planning, target localization, and motion execution. However, existing task planning methods for instruction following are either data-driven or underperform in zero-shot scenarios due to difficulties in grounding lengthy instructions into actionable plans under operational constraints. To address this, we propose FlowPlan, a structured multi-stage LLM workflow that elevates zero-shot pipeline and bridges the performance gap between zero-shot and data-driven in-context learning methods. By decomposing the planning process into modular stages--task information retrieval, language-level reasoning, symbolic-level planning, and logical evaluation--FlowPlan generates logically coherent action sequences while adhering to operational constraints and further extracts contextual guidance for precise instance-level target localization. Benchmarked on the ALFRED and validated in real-world applications, our method achieves competitive performance relative to data-driven in-context learning methods and demonstrates adaptability across diverse environments. This work advances zero-shot task planning in robotic systems without reliance on labeled data. Project website: https://instruction-following-project.github.io/.

cs.RO↗

arXiv2025

Next-to-leading order corrections to scalar perturbations of Kerr-anti-de Sitter black holes

The small Kerr-anti-de Sitter black hole demonstrates instability due to the superradiance of either a massive or massless scalar field. Previous leading-order approximations of the spectrum are inefficient. In particular, the leading-order real part of the eigenfrequency is insensitive to the spin of the black hole. In this work, we improve the analysis by including the next-to-leading-order contribution. Compared to the numerical calculation, the new spin-dependent real part presents significantly better agreement, and the error in the imaginary part is also reduced to less than 60% for most black hole spins.

gr-qc↗

arXiv2025

SVDC: Consistent Direct Time-of-Flight Video Depth Completion with Frequency Selective Fusion

Lightweight direct Time-of-Flight (dToF) sensors are ideal for 3D sensing on mobile devices. However, due to the manufacturing constraints of compact devices and the inherent physical principles of imaging, dToF depth maps are sparse and noisy. In this paper, we propose a novel video depth completion method, called SVDC, by fusing the sparse dToF data with the corresponding RGB guidance. Our method employs a multi-frame fusion scheme to mitigate the spatial ambiguity resulting from the sparse dToF imaging. Misalignment between consecutive frames during multi-frame fusion could cause blending between object edges and the background, which results in a loss of detail. To address this, we introduce an adaptive frequency selective fusion (AFSF) module, which automatically selects convolution kernel sizes to fuse multi-frame features. Our AFSF utilizes a channel-spatial enhancement attention (CSEA) module to enhance features and generates an attention map as fusion weights. The AFSF ensures edge detail recovery while suppressing high-frequency noise in smooth regions. To further enhance temporal consistency, We propose a cross-window consistency loss to ensure consistent predictions across different windows, effectively reducing flickering. Our proposed SVDC achieves optimal accuracy and consistency on the TartanAir and Dynamic Replica datasets. Code is available at https://github.com/Lan1eve/SVDC.

cs.CV↗

arXiv2025

HGDiffuser: Efficient Task-Oriented Grasp Generation via Human-Guided Grasp Diffusion Models

Task-oriented grasping (TOG) is essential for robots to perform manipulation tasks, requiring grasps that are both stable and compliant with task-specific constraints. Humans naturally grasp objects in a task-oriented manner to facilitate subsequent manipulation tasks. By leveraging human grasp demonstrations, current methods can generate high-quality robotic parallel-jaw task-oriented grasps for diverse objects and tasks. However, they still encounter challenges in maintaining grasp stability and sampling efficiency. These methods typically rely on a two-stage process: first performing exhaustive task-agnostic grasp sampling in the 6-DoF space, then applying demonstration-induced constraints (e.g., contact regions and wrist orientations) to filter candidates. This leads to inefficiency and potential failure due to the vast sampling space. To address this, we propose the Human-guided Grasp Diffuser (HGDiffuser), a diffusion-based framework that integrates these constraints into a guided sampling process. Through this approach, HGDiffuser directly generates 6-DoF task-oriented grasps in a single stage, eliminating exhaustive task-agnostic sampling. Furthermore, by incorporating Diffusion Transformer (DiT) blocks as the feature backbone, HGDiffuser improves grasp generation quality compared to MLP-based methods. Experimental results demonstrate that our approach significantly improves the efficiency of task-oriented grasp generation, enabling more effective transfer of human grasping strategies to robotic systems. To access the source code and supplementary videos, visit https://sites.google.com/view/hgdiffuser.

cs.RO↗

arXiv2025

A unified convergence analysis framework of the energy-stable ETDRK3 schemes for the No-slope-selection thin film model

This paper establishes a unified framework for the space-time convergence analysis of the energy-stable third-order accurate exponential time differencing Runge-Kutta schemes. By employing Fourier pseudo-spectral discretization in space and the inner product technique, we derive a rigorous Fourier eigenvalue analysis, which provides a detailed optimal convergence rate and error estimate. The primary challenge is addressing the complex nonlinear terms in the NSS equation. Fortunately, this challenge could be resolved through careful eigenvalue bound estimates for various operators.

math.NA↗

arXiv2025

FUNCTO: Function-Centric One-Shot Imitation Learning for Tool Manipulation

Learning tool use from a single human demonstration video offers a highly intuitive and efficient approach to robot teaching. While humans can effortlessly generalize a demonstrated tool manipulation skill to diverse tools that support the same function (e.g., pouring with a mug versus a teapot), current one-shot imitation learning (OSIL) methods struggle to achieve this. A key challenge lies in establishing functional correspondences between demonstration and test tools, considering significant geometric variations among tools with the same function (i.e., intra-function variations). To address this challenge, we propose FUNCTO (Function-Centric OSIL for Tool Manipulation), an OSIL method that establishes function-centric correspondences with a 3D functional keypoint representation, enabling robots to generalize tool manipulation skills from a single human demonstration video to novel tools with the same function despite significant intra-function variations. With this formulation, we factorize FUNCTO into three stages: (1) functional keypoint extraction, (2) function-centric correspondence establishment, and (3) functional keypoint-based action planning. We evaluate FUNCTO against exiting modular OSIL methods and end-to-end behavioral cloning methods through real-robot experiments on diverse tool manipulation tasks. The results demonstrate the superiority of FUNCTO when generalizing to novel tools with intra-function geometric variations. More details are available at https://sites.google.com/view/functo.

cs.RO↗

arXiv2025

Nondestructive quantitative estimation of cross-sectional corrosion degree of rebar using self-magnetic flux leakage field variation

To accurately assess the structural performance of corroded reinforced concrete structures, obtaining precise information on the corrosion range and corrosion degree of rebar is crucial. In this study, based on a comprehensive analysis of extensive high-precision magnetic field and three-dimensional structural light scanning data of 21 corroded rebars, it was found that the self-magnetic flux leakage can accurately identify the corrosion range, with an error not exceeding 3%. A proposed quantitative index NHx of the self-magnetic flux leakage amplitude exhibits a linear correlation with the cross-sectional corrosion degree of rebar, whose probability density distribution can be accurately described using the Weibull distribution function. Utilizing the Weibull distribution function of NHx and a Bayesian model, automatically rapid quantification of the rebar's cross-sectional corrosion degree based on the non-destructive testing-derived NHx values can be conveniently realized. This self-magnetic flux leakage-based novel method for quantifying rebar's cross-sectional corrosion degree is accurate, efficient, and well-suited for practical engineering applications, providing robust support for a precise assessment of the structural performance of corroded reinforced concrete structures.

physics.app-ph↗