Searcharxiv⌕ Search

arXiv subjects

Hong Wang

Publications and source records attributed to Hong Wang.

At least 109 records · Page 6Linked to original sources

Privacy-Aware Joint DNN Model Deployment and Partitioning Optimization for Collaborative Edge Inference Services

Edge inference (EI) has emerged as a promising paradigm to address the growing limitations of cloud-based Deep Neural Network (DNN) inference services, such as high response latency, limited scalability, and severe data privacy exposure. However, deploying DNN models on resource-constrained edge devices introduces additional challenges, including limited computation/storage resources, dynamic service demands, and heightened privacy risks. To tackle these issues, this paper presents a novel privacy-aware optimization framework that jointly addresses DNN model deployment, user-server association, and model partitioning, with the goal of minimizing long-term average inference delay under resource and privacy constraints. The problem is formulated as a complex, NP-hard stochastic optimization. To efficiently handle system dynamics and computational complexity, we employ a Lyapunov-based approach to transform the long-term objective into tractable per-slot decisions. Furthermore, we introduce a coalition formation game to enable adaptive user-server association and design a greedy algorithm for model deployment within each coalition. Extensive simulations demonstrate that the proposed algorithm significantly reduces inference delay and consistently satisfies privacy constraints, outperforming state-of-the-art baselines across diverse scenarios.

cs.LG↗

Progressive Language-guided Visual Learning for Multi-Task Visual Grounding

Multi-task visual grounding (MTVG) includes two sub-tasks, i.e., Referring Expression Comprehension (REC) and Referring Expression Segmentation (RES). The existing representative approaches generally follow the research pipeline which mainly consists of three core procedures, including independent feature extraction for visual and linguistic modalities, respectively, cross-modal interaction module, and independent prediction heads for different sub-tasks. Albeit achieving remarkable performance, this research line has two limitations: 1) The linguistic content has not been fully injected into the entire visual backbone for boosting more effective visual feature extraction and it needs an extra cross-modal interaction module; 2) The relationship between REC and RES tasks is not effectively exploited to help the collaborative prediction for more accurate output. To deal with these problems, in this paper, we propose a Progressive Language-guided Visual Learning framework for multi-task visual grounding, called PLVL, which not only finely mine the inherent feature expression of the visual modality itself but also progressively inject the language information to help learn linguistic-related visual features. In this manner, our PLVL does not need additional cross-modal fusion module while fully introducing the language guidance. Furthermore, we analyze that the localization center for REC would help identify the to-be-segmented object region for RES to some extent. Inspired by this investigation, we design a multi-task head to accomplish collaborative predictions for these two sub-tasks. Extensive experiments conducted on several benchmark datasets comprehensively substantiate that our PLVL obviously outperforms the representative methods in both REC and RES tasks. https://github.com/jcwang0602/PLVL

cs.CV↗

The Assouad dimension of Kakeya sets in $\mathbb{R}^3$

This paper studies the structure of Kakeya sets in $\mathbb{R}^3$. We show that for every Kakeya set $K\subset\mathbb{R}^3$, there exist well-separated scales $0<δ<ρ\leq 1$ so that the $δ$ neighborhood of $K$ is almost as large as the $ρ$ neighborhood of $K$. As a consequence, every Kakeya set in $\mathbb{R}^3$ has Assouad dimension 3 and every Ahlfors-David regular Kakeya set in $\mathbb{R}^3$ has Hausdorff dimension 3. We also show that every Kakeya set in $\mathbb{R}^3$ that has "stably equal" Hausdorff and packing dimension (this is a new notion, which is introduced to avoid certain obvious obstructions) must have Hausdorff dimension 3. The above results follow from certain multi-scale structure theorems for arrangements of tubes and rectangular prisms in three dimensions, and a mild generalization of the sticky Kakeya theorem previously proved by the authors.

math.CA↗

NTIRE 2025 Challenge on Day and Night Raindrop Removal for Dual-Focused Images: Methods and Results

This paper reviews the NTIRE 2025 Challenge on Day and Night Raindrop Removal for Dual-Focused Images. This challenge received a wide range of impressive solutions, which are developed and evaluated using our collected real-world Raindrop Clarity dataset. Unlike existing deraining datasets, our Raindrop Clarity dataset is more diverse and challenging in degradation types and contents, which includes day raindrop-focused, day background-focused, night raindrop-focused, and night background-focused degradations. This dataset is divided into three subsets for competition: 14,139 images for training, 240 images for validation, and 731 images for testing. The primary objective of this challenge is to establish a new and powerful benchmark for the task of removing raindrops under varying lighting and focus conditions. There are a total of 361 participants in the competition, and 32 teams submitting valid solutions and fact sheets for the final testing phase. These submissions achieved state-of-the-art (SOTA) performance on the Raindrop Clarity dataset. The project can be found at https://lixinustc.github.io/CVPR-NTIRE2025-RainDrop-Competition.github.io/.

cs.CV↗

A Physics-informed Machine Learning-based Control Method for Nonlinear Dynamic Systems with Highly Noisy Measurements

This study presents a physics-informed machine learning-based control method for nonlinear dynamic systems with highly noisy measurements. Existing data-driven control methods that use machine learning for system identification cannot effectively cope with highly noisy measurements, resulting in unstable control performance. To address this challenge, the present study extends current physics-informed machine learning capabilities for modeling nonlinear dynamics with control and integrates them into a model predictive control framework. To demonstrate the capability of the proposed method we test and validate with two noisy nonlinear dynamic systems: the chaotic Lorenz 3 system, and turning machine tool. Analysis of the results illustrate that the proposed method outperforms state-of-the-art benchmarks as measured by both modeling accuracy and control performance for nonlinear dynamic systems under high-noise conditions.

eess.SY↗

Exploiting Edited Large Language Models as General Scientific Optimizers

Large language models (LLMs) have been widely adopted in mathematical optimization in scientific scenarios for their extensive knowledge and advanced reasoning capabilities. Existing methods mainly focus on utilizing LLMs to solve optimization problems in a prompt-based manner, which takes observational feedback as additional textual descriptions. However, due to LLM's \textbf{high sensitivity to the prompts} and \textbf{tendency to get lost in lengthy prompts}, these methods struggle to effectively utilize the {observational} feedback from each optimization step, which severely hinders the applications for real-world scenarios. To address these challenges, we propose a conceptually simple and general {bi-level} optimization method, namely \textbf{G}eneral \textbf{S}cientific \textbf{O}ptimizers (GSO). Specifically, GSO first utilizes inner-level simulators as experimental platforms to evaluate the current solution and provide observational feedback. Then, LLMs serve as knowledgeable and versatile scientists, generating new solutions by refining potential errors from the feedback as the outer-level optimization. Finally, simulations together with the expert knowledge in LLMs are jointly updated with bi-level interactions via model editing. Extensive experiments show that GSO consistently outperforms existing state-of-the-art methods using \textit{six} different LLM backbones on \textit{seven} different tasks, demonstrating the effectiveness and a wide range of applications.

math.OC↗

Perturbation-Restrained Sequential Model Editing

Model editing is an emerging field that focuses on updating the knowledge embedded within large language models (LLMs) without extensive retraining. However, current model editing methods significantly compromise the general abilities of LLMs as the number of edits increases, and this trade-off poses a substantial challenge to the continual learning of LLMs. In this paper, we first theoretically analyze that the factor affecting the general abilities in sequential model editing lies in the condition number of the edited matrix. The condition number of a matrix represents its numerical sensitivity, and therefore can be used to indicate the extent to which the original knowledge associations stored in LLMs are perturbed after editing. Subsequently, statistical findings demonstrate that the value of this factor becomes larger as the number of edits increases, thereby exacerbating the deterioration of general abilities. To this end, a framework termed Perturbation Restraint on Upper bouNd for Editing (PRUNE) is proposed, which applies the condition number restraints in sequential editing. These restraints can lower the upper bound on perturbation to edited models, thus preserving the general abilities. Systematically, we conduct experiments employing three editing methods on three LLMs across four downstream tasks. The results show that PRUNE can preserve general abilities while maintaining the editing performance effectively in sequential model editing. The code are available at https://github.com/mjy1111/PRUNE.

cs.CL↗

Szemerédi-Trotter bounds for tubes and applications

We prove sharp estimates for incidences involving planar tubes that satisfy packing conditions. We apply them to improve the estimates for the Fourier transform of fractal measures supported on planar curves.

math.CA↗

Universal Machine Learning Interatomic Potentials are Ready for Solid Ion Conductors

With the rapid development of energy storage technology, high-performance solid-state electrolytes (SSEs) have become critical for next-generation lithium-ion batteries. These materials require high ionic conductivity, excellent electrochemical stability, and good mechanical properties to meet the demands of electric vehicles and portable electronics. However, traditional methods like density functional theory (DFT) and empirical force fields face challenges such as high computational costs, poor scalability, and limited accuracy across material systems. Universal machine learning interatomic potentials (uMLIPs) offer a promising solution with their efficiency and near-DFT-level accuracy.This study systematically evaluates six advanced uMLIP models (MatterSim, MACE, SevenNet, CHGNet, M3GNet, and ORBFF) in terms of energy, forces, thermodynamic properties, elastic moduli, and lithium-ion diffusion behavior. The results show that MatterSim outperforms others in nearly all metrics, particularly in complex material systems, demonstrating superior accuracy and physical consistency. Other models exhibit significant deviations due to issues like energy inconsistency or insufficient training data coverage.Further analysis reveals that MatterSim achieves excellent agreement with reference values in lithium-ion diffusivity calculations, especially at room temperature. Studies on Li3YCl6 and Li6PS5Cl uncover how crystal structure, anion disorder levels, and Na/Li arrangements influence ionic conductivity. Appropriate S/Cl disorder levels and optimized Na/Li arrangements enhance diffusion pathway connectivity, improving overall ionic transport performance.

cond-mat.mtrl-sci↗

A Refreshed Similarity-based Upsampler for Direct High-Ratio Feature Upsampling

Feature upsampling is a fundamental and indispensable ingredient of almost all current network structures for dense prediction tasks. Recently, a popular similarity-based feature upsampling pipeline has been proposed, which utilizes a high-resolution feature as guidance to help upsample the low-resolution deep feature based on their local similarity. Albeit achieving promising performance, this pipeline has specific limitations: 1) HR query and LR key features are not well aligned; 2) the similarity between query-key features is computed based on the fixed inner product form; 3) neighbor selection is coarsely operated on LR features, resulting in mosaic artifacts. These shortcomings make the existing methods along this pipeline primarily applicable to hierarchical network architectures with iterative features as guidance and they are not readily extended to a broader range of structures, especially for a direct high-ratio upsampling. Against the issues, we meticulously optimize every methodological design. Specifically, we firstly propose an explicitly controllable query-key feature alignment from both semantic-aware and detail-aware perspectives, and then construct a parameterized paired central difference convolution block for flexibly calculating the similarity between the well-aligned query-key features. Besides, we develop a fine-grained neighbor selection strategy on HR features, which is simple yet effective for alleviating mosaic artifacts. Based on these careful designs, we systematically construct a refreshed similarity-based feature upsampling framework named ReSFU. Extensive experiments substantiate that our proposed ReSFU is finely applicable to various types of architectures in a direct high-ratio upsampling manner, and consistently achieves satisfactory performance on different dense prediction applications, showing superior generality and ease of deployment.

cs.CV↗

Less is More: Simple yet Effective Heuristic Community Detection with Graph Convolution Network

Community detection is crucial in data mining. Traditional methods primarily focus on graph structure, often neglecting the significance of attribute features. In contrast, deep learning-based approaches incorporate attribute features and local structural information through contrastive learning, improving detection performance. However, existing algorithms' complex design and joint optimization make them difficult to train and reduce detection efficiency. Additionally, these methods require the number of communities to be predefined, making the results susceptible to artificial interference. To address these challenges, we propose a simple yet effective community detection algorithm that can adaptively detect communities without relying on data augmentation and contrastive optimization. The proposed algorithm first performs community pre-detection to extract global structural information adaptively. It then utilizes GCN to integrate local structures and attribute features. Subsequently, it combines global, local structures and attribute features in the feature space to discover community affiliations. Finally, a modularity maximization method is employed to optimize the communities based on these three types of information, thereby uncovering the community affiliation of each node. We conduct experimental comparisons across various graph datasets, evaluating the proposed algorithm against traditional methods and state-of-the-art community detection algorithms. The experimental results demonstrate that our algorithm achieves greater efficiency and accuracy in terms of both detection speed and effectiveness. The code is available at https://github.com/wuanghoong/Less-is-More.git.

cs.SI↗

Furstenberg sets estimate in the plane

We fully resolve the Furstenberg set conjecture in $\mathbb{R}^2$, that a $(s, t)$-Furstenberg set has Hausdorff dimension $\ge \min(s+t, \frac{3s+t}{2}, s+1)$. As a result, we obtain an analogue of Elekes' bound for the discretized sum-product problem and resolve an orthogonal projection question of Oberlin.

math.CA↗

DenseGNN: universal and scalable deeper graph neural networks for high-performance property prediction in crystals and molecules

Generative models generate vast numbers of hypothetical materials, necessitating fast, accurate models for property prediction. Graph Neural Networks (GNNs) excel in this domain but face challenges like high training costs, domain adaptation issues, and over-smoothing. We introduce DenseGNN, which employs Dense Connectivity Network (DCN), Hierarchical Node-Edge-Graph Residual Networks (HRN), and Local Structure Order Parameters Embedding (LOPE) to address these challenges. DenseGNN achieves state-of-the-art performance on datasets such as JARVIS-DFT, Materials Project, and QM9, improving the performance of models like GIN, Schnet, and Hamnet on materials datasets. By optimizing atomic embeddings and reducing computational costs, DenseGNN enables deeper architectures and surpasses other GNNs in crystal structure distinction, approaching X-ray diffraction method accuracy. This advances materials discovery and design.

cond-mat.mtrl-sci↗

Time fractional stochastic differential equations driven by pure jump Lévy noise

In this paper we introduce a variable order time fractional differential equation driven by pure jump Lévy noise, which models the motion of a particle exhibiting memory effect. We prove the well-posedness of this equation without assuming any integrability condition on the initial condition and the large jump coefficient, by using a truncation argument. Under some extra conditions, we also derive some $L^p$ moment estimates on the solutions. As an application of moment estimates, we prove the Hölder regularity of the solutions.

math.PR↗

Restriction estimates using decoupling theorems and two-ends Furstenberg inequalities

We propose to study the restriction conjecture using decoupling theorems and two-ends Furstenberg inequalities. Specifically, we pose a two-ends Furstenberg conjecture, which implies the restriction conjecture. As evidence, we prove this conjecture in the plane by using the Furstenberg set estimate. Moreover, we use this planar result to prove a restriction estimate for $p>22/7$ in three dimensions, which implies Wolff's $5/2$-hairbrush bound for Kakeya sets in $\mathbb{R}^3$. Our approach also makes improvements for the restriction conjecture in higher dimensions.

math.CA↗

Unlocking the Potential of Weakly Labeled Data: A Co-Evolutionary Learning Framework for Abnormality Detection and Report Generation

Anatomical abnormality detection and report generation of chest X-ray (CXR) are two essential tasks in clinical practice. The former aims at localizing and characterizing cardiopulmonary radiological findings in CXRs, while the latter summarizes the findings in a detailed report for further diagnosis and treatment. Existing methods often focused on either task separately, ignoring their correlation. This work proposes a co-evolutionary abnormality detection and report generation (CoE-DG) framework. The framework utilizes both fully labeled (with bounding box annotations and clinical reports) and weakly labeled (with reports only) data to achieve mutual promotion between the abnormality detection and report generation tasks. Specifically, we introduce a bi-directional information interaction strategy with generator-guided information propagation (GIP) and detector-guided information propagation (DIP). For semi-supervised abnormality detection, GIP takes the informative feature extracted by the generator as an auxiliary input to the detector and uses the generator's prediction to refine the detector's pseudo labels. We further propose an intra-image-modal self-adaptive non-maximum suppression module (SA-NMS). This module dynamically rectifies pseudo detection labels generated by the teacher detection model with high-confidence predictions by the student.Inversely, for report generation, DIP takes the abnormalities' categories and locations predicted by the detector as input and guidance for the generator to improve the generated reports.

cs.CV↗

Continual Learning for Segment Anything Model Adaptation

Although the current different types of SAM adaptation methods have achieved promising performance for various downstream tasks, such as prompt-based ones and adapter-based ones, most of them belong to the one-step adaptation paradigm. In real-world scenarios, we are generally confronted with the dynamic scenario where the data comes in a streaming manner. Driven by the practical need, in this paper, we first propose a novel Continual SAM adaptation (CoSAM) benchmark with 8 different task domains and carefully analyze the limitations of the existing SAM one-step adaptation methods in the continual segmentation scenario. Then we propose a novel simple-yet-effective Mixture of Domain Adapters (MoDA) algorithm which utilizes the Global Feature Tokens (GFT) and Global Assistant Tokens (GAT) modules to help the SAM encoder extract well-separated features for different task domains, and then provide the accurate task-specific information for continual learning. Extensive experiments demonstrate that our proposed MoDA obviously surpasses the existing classic continual learning methods, as well as prompt-based and adapter-based approaches for continual segmentation. Moreover, after sequential learning on the CoSAM benchmark with diverse data distributions, our MoDA maintains highly competitive results in the natural image domain, approaching the zero-shot performance of the original SAM, demonstrating its superior capability in knowledge preservation. Notably, the proposed MoDA can be seamlessly integrated into various one-step adaptation methods of SAM, which can consistently bring obvious performance gains. Code is available at \url{https://github.com/yangjl1215/CoSAM}

cs.CV↗