检索结果-内蒙古大学图书馆

32nd ACM International Conference on Multimedia, MM 2024

作者： Zeng, Gangyan Zhang, Yuan Wei, Jin Yang, Dongbao Zhang, Peng Gao, Yiwen Qin, Xugong Zhou, Yu School of Cyber Science and Engineering Nanjing University of Science and Technology Nanjing China State Key Laboratory of Media Convergence and Communication Communication University of China Beijing China Lenovo Research Beijing China Institute of Information Engineering Chinese Academy of Sciences Beijing China Laboratory for Advanced Computing and Intelligence Engineering Wuxi China TMCC College of Computer Science Nankai University Tianjin China

ISBN: (纸本)9798400706868

Scene text retrieval aims to find all images containing the query text from an image gallery. Current efforts tend to adopt an Optical Character Recognition (OCR) pipeline, which requires complicated text detection and/or recognition processes, resulting in inefficient and inflexible retrieval. Different from them, in this work we propose to explore the intrinsic potential of Contrastive Language-Image Pre-training (CLIP) for OCR-free scene text retrieval. Through empirical analysis, we observe that the main challenges of CLIP as a text retriever are: 1) limited text perceptual scale, and 2) entangled visual-semantic concepts. To this end, a novel model termed FDP (Focus, Distinguish, and Prompt) is developed. FDP first focuses on scene text via shifting the attention to the text area and probing the hidden text knowledge, and then divides the query text into content word and function word for processing, in which a semantic-aware prompting scheme and a distracted queries assistance module are utilized. Extensive experiments show that FDP significantly enhances the inference speed while achieving better or competitive retrieval accuracy compared to existing methods. Notably, on the IIIT-STR benchmark, FDP surpasses the state-of-the-art model by 4.37% with a 4 times faster speed. Furthermore, additional experiments under phrase-level and attribute-aware scene text retrieval settings validate FDP's particular advantages in handling diverse forms of query text. The source code will be available at https://***/Gyann-z/FDP. © 2024 ACM.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Projected gradient descent algorithm for ab initio crystal structure relaxation under a fixed unit cell volume

arXiv

引用

arXiv 2024年

作者： Hu, Yukuan Yin, Junlei Gao, Xingyu Liu, Xin Song, Haifeng State Key Laboratory of Scientific and Engineering Computing Institute of Computational Mathematics and Scientific/Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences Beijing100190 China School of Mathematical Sciences University of Chinese Academy of Sciences Beijing100049 China State Key Laboratory of Solidification Processing Northwestern Polytechnical University Shaanxi Xi’an710072 China Innovation Center NPU Chongqing Chongqing401135 China Laboratory of Computational Physics Institute of Applied Physics and Computational Mathematics Beijing100088 China

This paper is concerned with ab initio crystal structure relaxation under a fixed unit cell volume, which is a step in calculating the static equations of state and forms the basis of thermodynamic property calculations for materials. The task can be formulated as an energy minimization with a determinant constraint. Widely used line minimization-based methods (e.g., conjugate gradient method) lack both efficiency and convergence guarantees due to the nonconvex nature of the feasible region as well as the significant differences in the curvatures of the potential energy surface with respect to atomic and lattice components. To this end, we propose a projected gradient descent algorithm named PANBB. It is equipped with (i) search direction projections onto the tangent spaces of the nonconvex feasible region for lattice vectors, (ii) distinct curvature-aware initial trial step sizes for atomic and lattice updates, and (iii) a nonrestrictive line minimization criterion as the stopping rule for the inner loop. It can be proved that PANBB favors theoretical convergence to equilibrium states. Across a benchmark set containing 223 structures from various categories, PANBB achieves average speedup factors of approximately 1.41 and 1.45 over the conjugate gradient method and direct inversion in the iterative subspace implemented in off-the-shelf simulation software, respectively. Moreover, it normally converges on all the systems, manifesting its unparalleled robustness. As an application, we calculate the static equations of state for the high-entropy alloy AlCoCrFeNi, which remains elusive owing to 160 atoms representing both chemical and magnetic disorder and the strong local lattice distortion. The results are consistent with the previous calculations and are further validated by experimental thermodynamic data. © 2024, CC BY.

关键词： Conjugate gradient method

来源：评论

学校读者我要写书评

暂无评论

Marangoni-flow-assisted assembly of single-walled carbon nanotube films for human motion sensing

引用

Fundamental Research 2024年第3期4卷 570-574页

作者： Yuguang Chen Yitan Li Lu Han Hao Sun Min Lyu Zeyao Zhang Shigeo Maruyama Yan Li Key Laboratory for the Physics and Chemistry of Nanodevices Beijing National Laboratory of Molecular SciencesState Key Laboratory of Rare Earth Materials Chemistry and ApplicationsCollege of Chemistry and Molecular EngineeringPeking UniversityBeijing 100871China Academy for Advanced Interdisciplinary Studies Peking UniversityBeijing 100871China Bruker(Beijing)Scientific Technology Co.Ltd. Beijing 100192China Department of Mechanical Engineering The University of TokyoTokyo 113-8656Japan Peking University Shenzhen Institute Shenzhen 518057China PKU-HKUST ShenZhen-HongKong Institution Shenzhen 518057China

Single-walled carbon nanotubes(SWCNTs)present excellent electronic and mechanical properties desired in wearable and flexible *** preparation of SWCNT films is the first step for fabricating various *** work developed a scalable and feasible method to assemble SWCNT thin films on water surfaces based on Marangoni flow induced by surface tension *** films possess a large area of 40 cm×30 cm(extensible),a tunable thickness of 15∼150 nm,a high transparency of up to 96%,and a decent *** are ready to be directly transferred to various substrates,including flexible *** strain sensors were fabricated with the films on flexible *** sensors worked with high sensitivity and *** realizing multi-functional human motion sensing,including responding to voices,monitoring artery pulses,and detecting knuckle and muscle actions,the assembled SWCNT films demonstrated the potential for application in smart devices.

关键词： Single-walled carbon nanotubes Marangoni flow Assembly Interfaces Sensors

来源：评论

学校读者我要写书评

暂无评论

FPGA-Based Convolution Accelerator and Memristor IP Core for Cooperative Acceleration of the YOLO Network 4

FPGA-Based Convolution Accelerator and Memristor IP Core for...

引用

4th International Conference on Computer Science, Electronic Information engineering and Intelligent Control Technology, CEI 2024

作者： Tan, Aolong Yang, Yucheng Duan, Shukai Wang, Lidan College of Artificial Intelligence Southwest University Chongqing400715 China State Key Laboratory of Intelligent Vehicle Safety Technology China National Local Joint Engineering Research Center of Intelligent Transmission and Control Technology 400715 China Chongqing Key Laboratory of Brain-inspired Computing and Intelligent Chips Chongqing400715 China Laboratory of Luminescence Analysis and Molecular Sensing Southwest University Ministry of Education Chongqing400715 China

ISBN: (纸本)9798350378542

Object detection technology is widely used in areas such as autonomous driving, drone reconnaissance, smart surveillance and industrial automation, especially in remote or hazardous scenarios where it can improve efficiency and safety. Therefore, the development of a high-precision, low-power acceleration system integrated with low-power hardware is crucial. This paper proposes a collaborative acceleration scheme for the YOLO network based on an FPGA convolution accelerator and simulated memristors, which successfully implements and accelerates the YOLOv4-Tiny model. First, a reconfigurable single-core accelerator operator is designed for various computational operations in the model. Second, taking advantage of the digital device characteristics of memristors and the hardware resources of the FPGA, a memristor simulation IP core is built to exploit its memory characteristics for convolution operations. Through the synergistic effect of the convolution accelerator and the simulated IP core, a hybrid acceleration of the detection model is achieved. The power consumption of the convolution accelerator is only 2.75W, and in testing, the detection time for a single image is 823ms with an average accuracy of over 0.95, achieving high detection performance. This provides an opportunity to explore and research the efficient acceleration of object detection networks with integrated storage and computation. © 2024 IEEE.

关键词： Reconfigurable hardware

来源：评论

学校读者我要写书评

暂无评论

Adversarial Example Soups: Improving Transferability and Stealthiness for Free

arXiv

引用

arXiv 2024年

作者： Yang, Bo Zhang, Hengwei Wang, Jindong Yang, Yulong Lin, Chenhao Shen, Chao Zhao, Zhengyu State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou China School of Cyber Science and Engineering Xi’an Jiaotong University Xi’an China

Transferable adversarial examples cause practical security risks since they can mislead a target model without knowing its internal knowledge. A conventional recipe for maximizing transferability is to keep only the optimal adversarial example from all those obtained in the optimization pipeline. In this paper, for the first time, we question this convention and demonstrate that those discarded, sub-optimal adversarial examples can be reused to boost transferability. Specifically, we propose "Adversarial Example Soups" (AES), with AES-tune for averaging discarded adversarial examples in hyperparameter tuning and AES-rand for stability testing. In addition, our AES is inspired by "model soups", which averages weights of multiple fine-tuned models for improved accuracy without increasing inference time. Extensive experiments validate the global effectiveness of our AES, boosting 10 state-of-the-art transfer attacks and their combinations by up to 13% against 10 diverse (defensive) target models. We also show the possibility of generalizing AES to other types, e.g., directly averaging multiple in-the-wild adversarial examples that yield comparable success. A promising byproduct of AES is the improved stealthiness of adversarial examples since the perturbation variances are naturally reduced. Copyright © 2024, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A New Adaptive Balanced Augmented Lagrangian Method with Application to ISAC Beamforming Design

arXiv

引用

arXiv 2024年

作者： Wu, Jiageng Jiang, Bo Li, Xinxin Liu, Ya-Feng Yuan, Jianhua School of Mathematics Jilin University Changchun130012 China Ministry of Education Key Laboratory for NSLSCS School of Mathematical Sciences Nanjing Normal University Nanjing210023 China State Key Laboratory of Scientific and Engineering Computing Institute of Computational Mathematics and Scientific/Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences Beijing100190 China School of Sciences Beijing University of Posts and Telecommunications Beijing100876 China

In this paper, we consider a class of convex programming problems with linear equality constraints, which finds broad applications in machine learning and signal processing. We propose a new adaptive balanced augmented Lagrangian (ABAL) method for solving these problems. The proposed ABAL method adaptively selects the stepsize parameter and enjoys a low per-iteration complexity, involving only the computation of a proximal mapping of the objective function and the solution of a linear equation. These features make the proposed method wellsuited to large-scale problems. We then custom-apply the ABAL method to solve the ISAC beamforming design problem, which is formulated as a nonlinear semidefinite program in a previous work. This customized application requires careful exploitation of the problem's special structure such as the property that all of its signal-to-interference-and-noise-ratio (SINR) constraints hold with equality at the solution and an efficient computation of the proximal mapping of the objective function. Simulation results demonstrate the efficiency of the proposed ABAL method. © 2024, CC BY.

关键词： Signal to noise ratio

来源：评论

学校读者我要写书评

暂无评论

Stochastic optimization over expectation-formulated generalized Stiefel manifold

arXiv

引用

arXiv 2024年

作者： Jiang, Linshuo Xiao, Nachuan Liu, Xin State Key Laboratory of Scientific and Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences University of Chinese Academy of Sciences China The Institute of Operations Research and Analytics National University of Singapore Singapore

In this paper, we consider a class of stochastic optimization problems over the expectation-formulated generalized Stiefel manifold (SOEGS), where the objective function f is continuously differentiable. We propose a novel constraint dissolving penalty function with a customized penalty term (CDFCP), which maintains the same order of differentiability as f. Our theoretical analysis establishes the global equivalence between CDFCP and SOEGS, in the sense that they share the same first-order and second-order stationary points under mild conditions. These results on equivalence enable the direct implementation of various stochastic optimization approaches to solve SOEGS. In particular, we develop a stochastic gradient algorithm and its accelerated variant by incorporating an adaptive step size strategy. Furthermore, we prove their O(Ε−4) sample complexity for finding an Ε-stationary point of CDFCP. Comprehensive numerical experiments show the efficiency and robustness of our proposed algorithms. Copyright © 2024, The Authors. All rights reserved.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Attention-aligned mean-teacher learning for unsupervised domain adaptive person re-ID

引用

Science China Technological Sciences 2025年

作者： You LYU Zhen ZHANG Guoliang KANG Wei WEI Institute of Artificial Intelligence Beihang University Key Laboratory of Mathematics Informatics Behavioral Semantics Ministry of Education Beihang University State Key Laboratory of Software Development Environment and Advanced Innovation Center for Future Blockchain and Privacy Computing Beihang University School of Automation Science and Electrical Engineering Beihang University AVICAS Generic Technology Co. Ltd Zhongguancun Laboratory School of Mathematics Sciences Beihang University

In recent years, artificial intelligence has fueled the development of numerous applications [1, 2]. Person re-identification (re-ID) is a typical artificial intelligence system designed to automatically retrieve imag...

关键词：

来源：评论

学校读者我要写书评

暂无评论

Decentralized Stochastic Subgradient-based Methods for Nonsmooth Nonconvex Optimization

arXiv

引用

arXiv 2024年

作者： Zhang, Siyuan Xiao, Nachuan Liu, Xin State Key Laboratory of Scientific and Engineering Computing Academy of Mathematics and Systems Science Chinese Academy of Sciences University of Chinese Academy of Sciences China The Institute of Operations Research and Analytics National University of Singapore Singapore

In this paper, we concentrate on decentralized optimization problems with nonconvex and nonsmooth objective functions, especially on the decentralized training of nonsmooth neural networks. We introduce a unified framework to analyze the global convergence of decentralized stochastic subgradient-based methods. We prove the global convergence of our proposed framework under mild conditions, by establishing that the generated sequence asymptotically approximates the trajectories of its associated differential inclusion. Furthermore, we establish that our proposed framework covers a wide range of existing efficient decentralized subgradient-based methods, including decentralized stochastic subgradient descent (DSGD), DSGD with gradient-tracking technique (DSGD-T), and DSGD with momentum (DSGD-M). In addition, we introduce the sign map to regularize the update directions in DSGD-M, and show it is enclosed in our proposed framework. Consequently, our convergence results establish, for the first time, global convergence of these methods when applied to nonsmooth nonconvex objectives. Preliminary numerical experiments demonstrate that our proposed framework yields highly efficient decentralized subgradient-based methods with convergence guarantees in the training of nonsmooth neural networks. Copyright © 2024, The Authors. All rights reserved.

关键词： Lyapunov functions

来源：评论

学校读者我要写书评

暂无评论

An accurate identification method for network devices based on spatial attention mechanism

引用

Security and Safety 2023年第2期2卷 13-29页

作者： Xiuting Wang Ruixiang Li Shaoyong Du Xiangyang Luo Henan Polytechnic Institute Nanyang 473000China Henan Province Key Laboratory of Cyberspace Situation Awareness Zhengzhou 450001China State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou 450001China Institute of Information Engineering State Key Laboratory of Information SecurityBeijing 100093China

With the metaverse being the development direction of the next generation Internet,the popularity of intelligent devices,and the maturity of various emerging technologies,more and more intelligent devices try to connect to the Internet,which poses a major threat to the management and security protection of network *** present,the mainstream method of network equipment identification in the metaverse is to obtain the network traffic data generated in the process of device communication,extract the device features through analysis and processing,and identify the device based on a variety of learning *** methods often require manual participation,and it is difficult to capture the small differences between similar devices,leading to identification ***,we propose a deep learning device recognition method based on a spatial attention ***,we extract the required feature fields from the acquired network traffic ***,we normalize the data and convert it into grayscale *** that,we add a spatial attention mechanism to CNN and MLP respectively to increase the difference between similar network devices and further improve the recognition ***,we identify devices based on the deep learning model.A large number of experiments were carried out on 31 types of network devices such as web cameras,wireless routers,and *** results show that the accuracy of the proposed recognition method based on the spatial attention mechanism is increased by 0.8%and 2.0%,respectively,compared with the recognition method based only on the deep learning model under the CNN and MLP *** method proposed in this paper is significantly superior to the existing method of device-type recognition based only on a deep learning model.

关键词： Metaverse Device identification Deep learning Spatial attention

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：