arXiv subjects

Yan Wang

Publications and source records attributed to Yan Wang.

At least 415 records · Page 23Linked to original sources

arXiv2025

Block-Attention for Efficient Prefilling

We introduce Block-attention, an attention mechanism designed to address the increased inference latency and cost in Retrieval-Augmented Generation (RAG) scenarios. Traditional approaches often encode the entire context in an auto-regressive manner. Instead, Block-attention divides retrieved documents into discrete blocks, with each block independently calculating key-value (KV) states except for the final block. In RAG scenarios, by defining each passage as a block, Block-attention enables us to reuse the KV states of passages that have been seen before, thereby significantly reducing the latency and the computation overhead during inference. The implementation of Block-attention involves block segmentation, position re-encoding, and fine-tuning the LLM to adapt to the Block-attention mechanism. Experiments on 11 diverse benchmarks, including RAG, ICL, and general domains, demonstrate that after block fine-tuning, the Block-attention model not only achieves performance comparable to that of full-attention models, but can also seamlessly switch between the block and full attention modes without any performance loss. Notably, Block-attention significantly reduces the time to first token (TTFT) and floating point operations (FLOPs) to a very low level. It only takes 45 ms to output the first token for an input sequence with a total length of 32K. Compared to the full-attention models, the TTFT and corresponding FLOPs are reduced by 98.7% and 99.8%, respectively. Additionally, in Appendix A, we elaborate on how Block-attention is applied in Game AI scenario and the substantial potential benefits it entails. We strongly suggest researchers in the gaming field not to overlook this section.

cs.LG↗

arXiv2025

Safe Data-Driven Predictive Control

In the realm of control systems, model predictive control (MPC) has exhibited remarkable potential; however, its reliance on accurate models and substantial computational resources has hindered its broader application, especially within real-time nonlinear systems. This study presents an innovative control framework to enhance the practical viability of the MPC. The developed safe data-driven predictive control aims to eliminate the requirement for precise models and alleviate computational burdens in the nonlinear MPC (NMPC). This is achieved by learning both the system dynamics and the control policy, enabling efficient data-driven predictive control while ensuring system safety. The methodology involves a spatial temporal filter (STF)-based concurrent learning for system identification, a robust control barrier function (RCBF) to ensure the system safety amid model uncertainties, and a RCBF-based NMPC policy approximation. An online policy correction mechanism is also introduced to counteract performance degradation caused by the existing model uncertainties. Demonstrated through simulations on two applications, the proposed approach offers comparable performance to existing benchmarks with significantly reduced computational costs.

eess.SY↗

arXiv2025

Significant ballistic thermal transport across graphene layers: effect of nanoholes and lithium intercalation

Porous graphene and graphite are increasingly utilized in electrochemical energy storage and solar-thermal applications due to their unique structural and thermal properties. In this study, we conduct a comprehensive analysis of the lattice thermal transport and spectral phonon characteristics of holey graphite and multilayer graphene. Our results reveal that phonon modes propagating obliquely with respect to the graphene basal planes are the primary contributors to cross-plane thermal transport. These modes exhibit a predominantly ballistic nature, resulting in an almost linear increase in cross-plane thermal conductivity with the number of layers. The presence of nanoholes in graphene induces a broadband suppression of cross-plane phonon transport, whereas lithium ion intercalation shows potential to enhance it. These findings provide critical insights into the mechanisms governing cross-plane heat conduction in key graphene-based structures, offering valuable guidance for thermal management and engineering of van der Waals materials.

cond-mat.mes-hall↗

arXiv2025

AVP-AP: Self-supervised Automatic View Positioning in 3D cardiac CT via Atlas Prompting

Automatic view positioning is crucial for cardiac computed tomography (CT) examinations, including disease diagnosis and surgical planning. However, it is highly challenging due to individual variability and large 3D search space. Existing work needs labor-intensive and time-consuming manual annotations to train view-specific models, which are limited to predicting only a fixed set of planes. However, in real clinical scenarios, the challenge of positioning semantic 2D slices with any orientation into varying coordinate space in arbitrary 3D volume remains unsolved. We thus introduce a novel framework, AVP-AP, the first to use Atlas Prompting for self-supervised Automatic View Positioning in the 3D CT volume. Specifically, this paper first proposes an atlas prompting method, which generates a 3D canonical atlas and trains a network to map slices into their corresponding positions in the atlas space via a self-supervised manner. Then, guided by atlas prompts corresponding to the given query images in a reference CT, we identify the coarse positions of slices in the target CT volume using rigid transformation between the 3D atlas and target CT volume, effectively reducing the search space. Finally, we refine the coarse positions by maximizing the similarity between the predicted slices and the query images in the feature space of a given foundation model. Our framework is flexible and efficient compared to other methods, outperforming other methods by 19.8% average structural similarity (SSIM) in arbitrary view positioning and achieving 9% SSIM in two-chamber view compared to four radiologists. Meanwhile, experiments on a public dataset validate our framework's generalizability.

eess.IV↗

arXiv2025

SELC: Self-Supervised Efficient Local Correspondence Learning for Low Quality Images

Accurate and stable feature matching is critical for computer vision tasks, particularly in applications such as Simultaneous Localization and Mapping (SLAM). While recent learning-based feature matching methods have demonstrated promising performance in challenging spatiotemporal scenarios, they still face inherent trade-offs between accuracy and computational efficiency in specific settings. In this paper, we propose a lightweight feature matching network designed to establish sparse, stable, and consistent correspondence between multiple frames. The proposed method eliminates the dependency on manual annotations during training and mitigates feature drift through a hybrid self-supervised paradigm. Extensive experiments validate three key advantages: (1) Our method operates without dependency on external prior knowledge and seamlessly incorporates its hybrid training mechanism into original datasets. (2) Benchmarked against state-of-the-art deep learning-based methods, our approach maintains equivalent computational efficiency at low-resolution scales while achieving a 2-10x improvement in computational efficiency for high-resolution inputs. (3) Comparative evaluations demonstrate that the proposed hybrid self-supervised scheme effectively mitigates feature drift in long-term tracking while maintaining consistent representation across image sequences.

cs.RO↗

arXiv2025

A Survey on Video Analytics in Cloud-Edge-Terminal Collaborative Systems

The explosive growth of video data has driven the development of distributed video analytics in cloud-edge-terminal collaborative (CETC) systems, enabling efficient video processing, real-time inference, and privacy-preserving analysis. Among multiple advantages, CETC systems can distribute video processing tasks and enable adaptive analytics across cloud, edge, and terminal devices, leading to breakthroughs in video surveillance, autonomous driving, and smart cities. In this survey, we first analyze fundamental architectural components, including hierarchical, distributed, and hybrid frameworks, alongside edge computing platforms and resource management mechanisms. Building upon these foundations, edge-centric approaches emphasize on-device processing, edge-assisted offloading, and edge intelligence, while cloud-centric methods leverage powerful computational capabilities for complex video understanding and model training. Our investigation also covers hybrid video analytics incorporating adaptive task offloading and resource-aware scheduling techniques that optimize performance across the entire system. Beyond conventional approaches, recent advances in large language models and multimodal integration reveal both opportunities and challenges in platform scalability, data protection, and system reliability. Future directions also encompass explainable systems, efficient processing mechanisms, and advanced video analytics, offering valuable insights for researchers and practitioners in this dynamic field.

cs.NI↗

arXiv2025

Unveiling the Mist over 3D Vision-Language Understanding: Object-centric Evaluation with Chain-of-Analysis

Existing 3D vision-language (3D-VL) benchmarks fall short in evaluating 3D-VL models, creating a "mist" that obscures rigorous insights into model capabilities and 3D-VL tasks. This mist persists due to three key limitations. First, flawed test data, like ambiguous referential text in the grounding task, can yield incorrect and unreliable test results. Second, oversimplified metrics such as simply averaging accuracy per question answering (QA) pair, cannot reveal true model capability due to their vulnerability to language variations. Third, existing benchmarks isolate the grounding and QA tasks, disregarding the underlying coherence that QA should be based on solid grounding capabilities. To unveil the "mist", we propose Beacon3D, a benchmark for 3D-VL grounding and QA tasks, delivering a perspective shift in the evaluation of 3D-VL understanding. Beacon3D features (i) high-quality test data with precise and natural language, (ii) object-centric evaluation with multiple tests per object to ensure robustness, and (iii) a novel chain-of-analysis paradigm to address language robustness and model performance coherence across grounding and QA. Our evaluation of state-of-the-art 3D-VL models on Beacon3D reveals that (i) object-centric evaluation elicits true model performance and particularly weak generalization in QA; (ii) grounding-QA coherence remains fragile in current 3D-VL models, and (iii) incorporating large language models (LLMs) to 3D-VL models, though as a prevalent practice, hinders grounding capabilities and has yet to elevate QA capabilities. We hope Beacon3D and our comprehensive analysis could benefit the 3D-VL community towards faithful developments.

cs.CV↗

arXiv2025

Detection and parameter estimation of supermassive black hole ringdown signals using a pulsar timing array

Gravitational wave (GW) searches using pulsar timing arrays (PTAs) are commonly assumed to be limited to a GW frequency of $\lesssim 4\times 10^{-7}$Hz given by the Nyquist rate associated with the average observational cadence of $2$ weeks for a single pulsar. However, by taking advantage of asynchronous observations of multiple pulsars, a PTA can detect GW signals at higher frequencies. This allows a sufficiently large PTA to detect and characterize the ringdown signals emitted following the merger of supermassive binary black holes (SMBBHs), leading to stringent tests of the no-hair theorem in the mass range of such systems. Such large-scale PTAs are imminent with the advent of the FAST telescope and the upcoming era of the Square Kilometer Array (SKA). To scope out the data analysis challenges involved in such a search, we propose a likelihood-based method coupled with Particle Swarm Optimization and apply it to a simulated large-scale PTA comprised of $100$ pulsars, each having a timing residual noise standard deviation of $100$~nsec, with randomized observation times. Focusing on the dominant $(2,2)$ mode of the ringdown signal, we show that it is possible to achieve a $99\%$ detection probability with a false alarm probability below $0.2\%$ for an optimal signal-to-noise ratio (SNR) $>10$. This corresponds, for example, to an equal-mass non-spinning SMBBH with an observer frame chirp mass $M_c = 9.52\times10^{9}M_{\odot}$ at a luminosity distance of $D_L = 420$ Mpc.

astro-ph.HE↗

arXiv2025

Fundamental Limit of Angular Resolution in Partly Calibrated Arrays with Position Errors

We consider high angular resolution detection using distributed mobile platforms implemented with so-called partly calibrated arrays, where position errors between subarrays exist and the counterparts within each subarray are ideally calibrated. Since position errors between antenna arrays affect the coherent processing of measurements from these arrays, it is commonly believed that its angular resolution is influenced. A key question is whether and how much the angular resolution of partly calibrated arrays is affected by the position errors, in comparison with ideally calibrated arrays. To address this fundamental problem, we theoretically illustrate that partly calibrated arrays approximately achieve high angular resolution. Our analysis uses a special characteristic of Cramer-Rao lower bound (CRB) w.r.t. the source separation: When the source separation increases, the CRB first declines rapidly, then plateaus out, and the turning point is close to the angular resolution limit. This means that the turning point of CRB can be used to indicate angular resolution. We then theoretically analyze the declining and plateau phases of CRB, and explain that the turning point of CRB in partly calibrated arrays is close to the angular resolution limit of distributed arrays without errors, demonstrating high resolution ability. This work thus provides a theoretical guarantee for the high-resolution performance of distributed antenna arrays in mobile platforms.

eess.SP↗

arXiv2025

SIT-FER: Integration of Semantic-, Instance-, Text-level Information for Semi-supervised Facial Expression Recognition

Semi-supervised deep facial expression recognition (SS-DFER) has gained increasingly research interest due to the difficulty in accessing sufficient labeled data in practical settings. However, existing SS-DFER methods mainly utilize generated semantic-level pseudo-labels for supervised learning, the unreliability of which compromises their performance and undermines the practical utility. In this paper, we propose a novel SS-DFER framework that simultaneously incorporates semantic, instance, and text-level information to generate high-quality pseudo-labels. Specifically, for the unlabeled data, considering the comprehensive knowledge within the textual descriptions and instance representations, we respectively calculate the similarities between the facial vision features and the corresponding textual and instance features to obtain the probabilities at the text- and instance-level. Combining with the semantic-level probability, these three-level probabilities are elaborately aggregated to gain the final pseudo-labels. Furthermore, to enhance the utilization of one-hot labels for the labeled data, we also incorporate text embeddings excavated from textual descriptions to co-supervise model training, enabling facial visual features to exhibit semantic correlations in the text space. Experiments on three datasets demonstrate that our method significantly outperforms current state-of-the-art SS-DFER methods and even exceeds fully supervised baselines. The code will be available at https://github.com/PatrickStarL/SIT-FER.

cs.CV↗

arXiv2025

Linear hypermaps--modelling linear hypergraphs on surfaces

A hypergraph is linear if each pair of distinct vertices appears in at most one common edge. We say $\varGamma=(V,E)$ is an associated graph of a linear hypergraph $\mathcal{H}=(V, X)$ if for any $x\in X$, the induced subgraph $\varGamma[x]$ is a cycle, and for any $e\in E$, there exists a unique edge $y\in X$ such that $e\subseteq y$. A linear hypermap $\mathcal{M}$ is a $2$-cell embedding of a connected linear hypergraph $\mathcal{H}$'s associated graph $\varGamma$ on a compact connected surface, such that for any edge $x\in E(\mathcal{H})$, $\varGamma[x]$ is the boundary of a $2$-cell and for any $e\in E(\varGamma)$, $e$ is incident with two distinct $2$-cells. In this paper, we introduce linear hypermaps to model linear hypergraphs on surfaces and regular linear hypermaps modelling configurations on the surfaces. As an application, we classify regular linear hypermaps on the sphere and determine the total number of proper regular linear hypermaps of genus 2 to 101.

math.CO↗

arXiv2025

A Survey on RGB, 3D, and Multimodal Approaches for Unsupervised Industrial Image Anomaly Detection

In the advancement of industrial informatization, unsupervised anomaly detection technology effectively overcomes the scarcity of abnormal samples and significantly enhances the automation and reliability of smart manufacturing. As an important branch, industrial image anomaly detection focuses on automatically identifying visual anomalies in industrial scenarios (such as product surface defects, assembly errors, and equipment appearance anomalies) through computer vision techniques. With the rapid development of Unsupervised industrial Image Anomaly Detection (UIAD), excellent detection performance has been achieved not only in RGB setting but also in 3D and multimodal (RGB and 3D) settings. However, existing surveys primarily focus on UIAD tasks in RGB setting, with little discussion in 3D and multimodal settings. To address this gap, this artical provides a comprehensive review of UIAD tasks in the three modal settings. Specifically, we first introduce the task concept and process of UIAD. We then overview the research on UIAD in three modal settings (RGB, 3D, and multimodal), including datasets and methods, and review multimodal feature fusion strategies in multimodal setting. Finally, we summarize the main challenges faced by UIAD tasks in the three modal settings, and offer insights into future development directions, aiming to provide researchers with a comprehensive reference and offer new perspectives for the advancement of industrial informatization. Corresponding resources are available at https://github.com/Sunny5250/Awesome-Multi-Setting-UIAD.

cs.CV↗

arXiv2025

AnimatePainter: A Self-Supervised Rendering Framework for Reconstructing Painting Process

Humans can intuitively decompose an image into a sequence of strokes to create a painting, yet existing methods for generating drawing processes are limited to specific data types and often rely on expensive human-annotated datasets. We propose a novel self-supervised framework for generating drawing processes from any type of image, treating the task as a video generation problem. Our approach reverses the drawing process by progressively removing strokes from a reference image, simulating a human-like creation sequence. Crucially, our method does not require costly datasets of real human drawing processes; instead, we leverage depth estimation and stroke rendering to construct a self-supervised dataset. We model human drawings as "refinement" and "layering" processes and introduce depth fusion layers to enable video generation models to learn and replicate human drawing behavior. Extensive experiments validate the effectiveness of our approach, demonstrating its ability to generate realistic drawings without the need for real drawing process data.

cs.CV↗

arXiv2025

Vertex degree sums for rainbow matchings in 3-uniform hypergraphs

Let $n \in 3\mathbb{Z}$ be sufficiently large. Zhang, Zhao and Lu proved that if $H$ is a 3-uniform hypergraph with $n$ vertices and no isolated vertices, and if $deg(u)+deg(v) > \frac{2}{3}n^2 - \frac{8}{3}n + 2$ for any two vertices $u$ and $v$ that are contained in some edge of $H$, then $ H $ admits a perfect matching. In this paper, we prove that the rainbow version of Zhang, Zhao and Lu's result is asymptotically true. More specifically, let $δ> 0$ and $ F_1, F_2, \dots, F_{n/3} $ be 3-uniform hypergraphs on a common set of $n$ vertices. For each $ i \in [n/3] $, suppose that $F_i$ has no isolated vertices and $deg_{F_i}(u)+deg_{F_i}(v) > \left( \frac{2}{3} + δ\right)n^2$ holds for any two vertices $u$ and $v$ that are contained in some edge of $F_i$. Then $ \{ F_1, F_2, \dots, F_{n/3} \} $ admits a rainbow matching. Note that this result is asymptotically tight.

math.CO↗

arXiv2025

Preference Construction: A Bayesian Interactive Preference Elicitation Framework Based on Monte Carlo Tree Search

We present a novel preference learning framework to capture participant preferences efficiently within limited interaction rounds. It involves three main contributions. First, we develop a variational Bayesian approach to infer the participant's preference model by estimating posterior distributions and managing uncertainty from limited information. Second, we propose an adaptive questioning policy that maximizes cumulative uncertainty reduction, formulating questioning as a finite Markov decision process and using Monte Carlo Tree Search to prioritize promising question trajectories. By considering long-term effects and leveraging the efficiency of the Bayesian approach, the policy avoids shortsightedness. Third, we apply the framework to Multiple Criteria Decision Aiding, with pairwise comparison as the preference information and an additive value function as the preference model. We integrate the reparameterization trick to address high-variance issues, enhancing robustness and efficiency. Computational studies on real-world and synthetic datasets demonstrate the framework's practical usability, outperforming baselines in capturing preferences and achieving superior uncertainty reduction within limited interactions.

cs.LG↗

arXiv2025

Extrapolated Urban View Synthesis Benchmark

Photorealistic simulators are essential for the training and evaluation of vision-centric autonomous vehicles (AVs). At their core is Novel View Synthesis (NVS), a crucial capability that generates diverse unseen viewpoints to accommodate the broad and continuous pose distribution of AVs. Recent advances in radiance fields, such as 3D Gaussian Splatting, achieve photorealistic rendering at real-time speeds and have been widely used in modeling large-scale driving scenes. However, their performance is commonly evaluated using an interpolated setup with highly correlated training and test views. In contrast, extrapolation, where test views largely deviate from training views, remains underexplored, limiting progress in generalizable simulation technology. To address this gap, we leverage publicly available AV datasets with multiple traversals, multiple vehicles, and multiple cameras to build the first Extrapolated Urban View Synthesis (EUVS) benchmark. Meanwhile, we conduct both quantitative and qualitative evaluations of state-of-the-art NVS methods across different evaluation settings. Our results show that current NVS methods are prone to overfitting to training views. Besides, incorporating diffusion priors and improving geometry cannot fundamentally improve NVS under large view changes, highlighting the need for more robust approaches and large-scale training. We will release the data to help advance self-driving and urban robotics simulation technology.

cs.CV↗

arXiv2025

TR-DQ: Time-Rotation Diffusion Quantization

Diffusion models have been widely adopted in image and video generation. However, their complex network architecture leads to high inference overhead for its generation process. Existing diffusion quantization methods primarily focus on the quantization of the model structure while ignoring the impact of time-steps variation during sampling. At the same time, most current approaches fail to account for significant activations that cannot be eliminated, resulting in substantial performance degradation after quantization. To address these issues, we propose Time-Rotation Diffusion Quantization (TR-DQ), a novel quantization method incorporating time-step and rotation-based optimization. TR-DQ first divides the sampling process based on time-steps and applies a rotation matrix to smooth activations and weights dynamically. For different time-steps, a dedicated hyperparameter is introduced for adaptive timing modeling, which enables dynamic quantization across different time steps. Additionally, we also explore the compression potential of Classifier-Free Guidance (CFG-wise) to establish a foundation for subsequent work. TR-DQ achieves state-of-the-art (SOTA) performance on image generation and video generation tasks and a 1.38-1.89x speedup and 1.97-2.58x memory reduction in inference compared to existing quantization methods.

cs.CV↗

arXiv2025

Electromagnetic Information Theory: Fundamentals, Paradigm Shifts, and Applications

This paper explores the emerging research direction of electromagnetic information theory (EIT), which aims to integrate traditional Shannon-based methodologies with physical consistency, particularly the electromagnetic properties of communication channels. We propose an EIT-based multiple-input multiple-output (MIMO) paradigm that enhances conventional spatially-discrete MIMO models by incorporating the concepts of electromagnetic (EM) precoding and EM combining. This approach aims to improve the modeling of next-generation systems while remaining consistent with Shannon's theoretical foundations. We explore typical EIT applications, such as densely spaced MIMO, near-field communications, and tri-polarized antennas, and analyze their channel characteristics through theoretical simulations and measured datasets. The paper also discusses critical research challenges and opportunities for EIT applications from an industrial perspective, emphasizing the field's potential for practical applications.

cs.IT↗