检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Zhang, Wei Ma, Mengting Jiang, Yizhen Lian, Rongrong Wu, Zhenkai Cui, Kangning Ma, Xiaowen School of Software Technology Zhejiang University Hangzhou310027 China Innovation Center of Yangtze River Delta Zhejiang University Zhejiang Jiaxing314103 China School of Computer Science and Technology Zhejiang University Hangzhou310027 China Department of Mathematics City University of Hong Kong Hong Kong Noah’s Ark Lab Huawei Shanghai201206 China

Compared with natural images, remote sensing images (RSIs) have the unique characteristic. i.e., larger intraclass variance, which makes semantic segmentation for remote sensing images more challenging. Moreover, existing semantic segmentation models for remote sensing images usually employ a vanilla softmax classifier, which has three drawbacks: (1) non-direct supervision for the pixel representations during training;(2) inadequate modeling ability of parametric softmax classifiers under large intraclass variance;and (3) opaque process of classification decision. In this paper, we propose a novel classifier (called CenterSeg) customized for RSI semantic segmentation, which solves the abovementioned problems with multiple prototypes, direct supervision under Grassmann manifold, and interpretability strategy. Specifically, for each class, our CenterSeg obtains local class centers by aggregating corresponding pixel features based on ground-truth masks, and generates multiple prototypes through hard attention assignment and momentum updating. In addition, we introduce the Grassmann manifold and constrain the joint embedding space of pixel features and prototypes based on two additional regularization terms. Especially, during the inference, CenterSeg can further provide interpretability to the model by restricting the prototype as a sample of the training set. Experimental results on three remote sensing segmentation datasets validate the effectiveness of the model. Besides the superior performance, CenterSeg has the advantages of simplicity, lightweight, compatibility, and interpretability. Code is available at https://***/xwmaxwma/rssegmentation. © 2025, CC BY.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

MLink: Linking Black-Box Models from Multiple Domains for Collaborative Inference

arXiv

引用

arXiv 2022年

作者： Yuan, Mu Zhang, Lan Zheng, Zimu Zhang, Yi-Nan Li, Xiang-Yang The School of Computer Science and Technology University of Science and Technology of China Hefei China Institute of Artificial Intelligence Hefei Comprehensive National Science Center Hefei China The Edge Cloud Innovation Lab Huawei Cloud China

The cost efficiency of model inference is critical to real-world machine learning (ML) applications, especially for delay-sensitive tasks and resource-limited devices. A typical dilemma is: in order to provide complex intelligent services (e.g. smart city), we need inference results of multiple ML models, but the cost budget (e.g. GPU memory) is not enough to run all of them. In this work, we study underlying relationships among black-box ML models and propose a novel learning task: model linking, which aims to bridge the knowledge of different black-box models by learning mappings (dubbed model links) between their output spaces. We propose the design of model links which supports linking heterogeneous black-box ML models. Also, in order to address the distribution discrepancy challenge, we present adaptation and aggregation methods of model links. Based on our proposed model links, we developed a scheduling algorithm, named MLink. Through collaborative multi-model inference enabled by model links, MLink can improve the accuracy of obtained inference results under the cost budget. We evaluated MLink on a multi-modal dataset with seven different ML models and two real-world video analytics systems with six ML models and 3,264 hours of video. Experimental results show that our proposed model links can be effectively built among various black-box models. Under the budget of GPU memory, MLink can save 66.7% inference computations while preserving 94% inference accuracy, which outperforms multi-task learning, deep reinforcement learning-based scheduler and frame filtering baselines. Copyright © 2022, The Authors. All rights reserved.

关键词： Scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

3Deformer: A Common Framework for Image-Guided Mesh Deformation

arXiv

引用

arXiv 2023年

作者： Su, Hao Liu, Xuefeng Niu, Jianwei Wan, Ji Wu, Xinghao State Key Lab of VR Technology and System School of Computer Science and Engineering Beihang University Beijing100000 China Industrial Technology Research Institute School of Information Engineering Zhengzhou University Henan450000 China Hangzhou Innovation Institute Beihang University Zhejiang 310000 China

We propose 3Deformer, a general-purpose framework for interactive 3D shape editing. Given a source 3D mesh with semantic materials, and a user-specified semantic image, 3Deformer can accurately edit the source mesh following the shape guidance of the semantic image, while preserving the source topology as rigid as possible. Recent studies of 3D shape editing mostly focus on learning neural networks to predict 3D shapes, which requires high-cost 3D training datasets and is limited to handling objects involved in the datasets. Unlike these studies, our 3Deformer is a non-training and common framework, which only requires supervision of readily-available semantic images, and is compatible with editing various objects unlimited by datasets. In 3Deformer, the source mesh is deformed utilizing the differentiable renderer technique, according to the correspondences between semantic images and mesh materials. However, guiding complex 3D shapes with a simple 2D image incurs extra challenges, that is, the deform accuracy, surface smoothness, geometric rigidity, and global synchronization of the edited mesh should be guaranteed. To address these challenges, we propose a hierarchical optimization architecture to balance the global and local shape features, and propose further various strategies and losses to improve properties of accuracy, smoothness, rigidity, and so on. Extensive experiments show that our 3Deformer is able to produce impressive results and reaches the state-of-the-art level. Copyright © 2023, The Authors. All rights reserved.

关键词： Rigidity

来源：评论

学校读者我要写书评

暂无评论

Traceable and Authenticable Image Tagging for Fake News Detection

arXiv

引用

arXiv 2022年

作者： Meng, Ruohan Zhou, Zhili Cui, Qi Lam, Kwok-Yan Kot, Alex School of Computer and Software Nanjing University of Information Science and Technology China Institute of Artificial Intelligence and Blockchain Guangzhou University China Computer Science and Engineering Nanyang Technological University Singapore Lab Nanyang Technological University Singapore Centre for Computer Vision and Deep Learning University of Windsor Canada

To prevent fake news images from misleading the public, it is desirable not only to verify the authenticity of news images but also to trace the source of fake news, so as to provide a complete forensic chain for reliable fake news detection. To simultaneously achieve the goals of authenticity verification and source tracing, we propose a traceable and authenticable image tagging approach that is based on a design of Decoupled Invertible Neural Network (DINN). The designed DINN can simultaneously embed the dual-tags, i.e., authenticable tag and traceable tag, into each news image before publishing, and then separately extract them for authenticity verification and source tracing. Moreover, to improve the accuracy of dual-tags extraction, we design a parallel Feature Aware Projection Model (FAPM) to help the DINN preserve essential tag information. In addition, we define a Distance Metric-Guided Module (DMGM) that learns asymmetric one-class representations to enable the dual-tags to achieve different robustness performances under malicious manipulations. Extensive experiments, on diverse datasets and unseen manipulations, demonstrate that the proposed tagging approach achieves excellent performance in the aspects of both authenticity verification and source tracing for reliable fake news detection and outperforms the prior works. Copyright © 2022, The Authors. All rights reserved.

关键词： Image annotation

来源：评论

学校读者我要写书评

暂无评论

Scale Optimization Using Evolutionary Reinforcement Learning for Object Detection on Drone Imagery

arXiv

引用

arXiv 2023年

作者： Zhang, Jialu Yang, Xiaoying He, Wentao Ren, Jianfeng Zhang, Qian Zhao, Yitian Bai, Ruibin He, Xiangjian Liu, Jiang The Digital Port Technologies Lab School of Computer Science University of Nottingham Ningbo China Nottingham Ningbo China Beacons of Excellence Research and Innovation Institute University of Nottingham Ningbo China Cixi Institute of Biomedical Engineering Chinese Academy of Sciences China Department of Computer Science and Engineering Southern University of Science and Technology China

Object detection in aerial imagery presents a significant challenge due to large scale variations among objects. This paper proposes an evolutionary reinforcement learning agent, integrated within a coarse-to-fine object detection framework, to optimize the scale for more effective detection of objects in such images. Specifically, a set of patches potentially containing objects are first generated. A set of rewards measuring the localization accuracy, the accuracy of predicted labels, and the scale consistency among nearby patches are designed in the agent to guide the scale optimization. The proposed scale-consistency reward ensures similar scales for neighboring objects of the same category. Furthermore, a spatial-semantic attention mechanism is designed to exploit the spatial semantic relations between patches. The agent employs the proximal policy optimization strategy in conjunction with the evolutionary strategy, effectively utilizing both the current patch status and historical experience embedded in the agent. The proposed model is compared with state-of-the-art methods on two benchmark datasets for object detection on drone imagery. It significantly outperforms all the compared methods. Copyright © 2023, The Authors. All rights reserved.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Using Problem Frames Approach for Key Information Extraction from Natural Language Requirements

Using Problem Frames Approach for Key Information Extraction...

引用

IEEE International Conference on Software Quality, Reliability and Security Companion (QRS-C)

作者： Shangzhi Tang Xuan Chen Hongbin Xiao Jiahao Wei Zhi Li School of Computer Science and Engineering Guangxi Normal University Guilin China Key Lab of Education Blockchain & Intelligent Technology Ministry of Education Guangxi Normal University Guilin China Guangxi Key Laboratory of Multi-source Information Mining and Security Guangxi Normal University Guilin China

The correct capturing and understanding of requirements play a crucial role in the software development process. However, the ambiguity in requirements described using natural language (NL) poses a challenge. Existing methods often require the use of specific requirements templates to rewrite system requirements, which increases efforts in analyzing requirements. Additionally, frequent requirement changes during the development process can lead to extra costs. In this paper, we propose a visualization method that allows user to extract key information from NL requirements, aiming to mitigate the difficulty and efforts in analyzing NL requirements. We also argue that more attention should be paid to the problems that need to be solved in the development process, especially the security problems. Application of our method to practical requirements sentences demonstrates its effectiveness in extracting the problem concerns from the requirements, leading to the stepwise derivation of solutions. We envision that our approach may have the potential to support iterative updates and analysis of NL requirements automatically.

关键词：

来源：评论

学校读者我要写书评

暂无评论

HL-DPoS: An Enhanced Anti-Long-Range Attack DPoS Algorithm

arXiv

引用

arXiv 2023年

作者： Li, Yang Xia, Chunhe Li, Chunyan Zhao, Yuan Chen, Chen Wang, Tianbo School of Computer Science and Engineering Beihang University Beijing China Beijing Key Lab. of Network Technology Beihang University Beijing China Guangxi Collaborative Innovation Center of Multi-source Information Integration and Intelligent Processing Guangxi Normal University Guilin China School of Cyber Science and Technology Beihang University Beijing China

The consensus algorithm is crucial in blockchain for ensuring the validity and security of transactions across the decentralized network. However, achieving consensus among nodes and packaging blocks in blockchain networks is a complex task that requires efficient and secure consensus algorithms. The DPoS consensus algorithm has emerged as a popular choice due to its fast transaction processing and high throughput. Despite these advantages, the algorithm still suffers from weaknesses such as centralization and vulnerability to long-range attacks, which can compromise the integrity of the blockchain network. To combat these problems, we developed an Enhanced Anti-Long-Range Attack DPoS algorithm (HL-DPoS). First, we split nodes into pieces to reduce centralization issues while giving witness nodes the power to report and benefit from malicious node's reports, maintaining high efficiency and high security. Second, we propose a validation method in HL-DPoS that compares consensuses transactions with the longest chain to detect long-range attacks. Algorithm analysis and simulation experiment results demonstrate that our HL-DPoS consensus algorithm improves security while achieving better consensus performance. Copyright © 2023, The Authors. All rights reserved.

关键词： blockchain

来源：评论

学校读者我要写书评

暂无评论

A Cross-chain Model Based on Credit Hierarchical Notary Group for IoT Roaming Settlement 6th

A Cross-chain Model Based on Credit Hierarchical Notary Gro...

引用

6th International Conference on blockchain and Trustworthy Systems, BlockSys 2024

作者： Lu, Shuhao Liang, Hai Wang, Yujue Ding, Yong Yang, Changsong Guo, Zhenwei Guangxi Key Laboratory of Cryptography and Information Security School of Computer Science and Information Security Guilin University of Electronic Technology Guilin541004 China Guangxi Engineering Research Center of Industrial Internet Security and Blockchain Guilin University of Electronic Technology Guilin China Hangzhou Innovation Institute of Beihang University Hangzhou China Guangxi Key Laboratory of Digital Infrastructure Guangxi Zhuang Autonomous Region Information Center Nanning China Institute of Cyberspace Technology HKCT Institute for Higher Education Hong Kong

ISBN: (纸本)9789819614134

In IoT with blockchain, an increasing number of mobile IoT devices are starting to work across various locations, requiring operators connected to these IoT devices to provide roaming services. Due to the differences in the blockchains used by operators, a safe and efficient cross-chain model for IoT device roaming settlement is significant for operators. In this paper, a cross-chain model based on credit hierarchical multi-notary is proposed to connect different blockchains in IoT devices roaming settlement. Firstly, a cross-chain election method by sub-notary group is proposed to solve the problem of "Richer Get Richer" in blockchain with credit scheme and avoid risk of malicious notary nodes. Secondly, a re-election mechanism is used to make the process of cross-chain running properly with the number of IoT devices is increasing rapidly and transaction concurrency. Moreover, a weighted voting system is designed for sub-notary group cross-chain model to reduce the influence of malicious notary nodes. Experimental results demonstrate that the proposed cross-chain model performs better on credit ecology and defending malicious nodes’ influence than the model without sub-notary mechanism. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： blockchain

来源：评论

学校读者我要写书评

暂无评论

Natural Language Processing-Based Requirements Modeling: A Case Study on Problem Frames

Natural Language Processing-Based Requirements Modeling: A C...

引用

Asia-Pacific Conference on Software Engineering

作者： Jiahao Wei Xuan Chen Hongbin Xiao Shangzhi Tang Xiaolan Xie Zhi Li School of Computer Science and Engineering Guangxi Normal University Guilin China College of Information Science and Engineering Guilin University of Technology Guilin China Key Lab of Education Blockchain & Intelligent Technology Ministry of Education Guangxi Normal University Guilin China Guangxi Key Laboratory of Multi-source Information Mining and Security Guangxi Normal University Guilin China

Natural Language Processing (NLP) aims to study various theories and methods that enable effective communication between humans and computers in natural language. One specific technique, known as Keyphrase Extraction (KPE), has achieved significant success in recent years through pre-trained Language Models (LM), particularly BERT and ELMo. Currently, researchers have presented NLP4RE at Requirements Engineering (RE) conferences, contemplating how to leverage the cutting-edge advancements in NLP to achieve the integration of closely related research domains. In the practice of requirements engineering, it is not always assumed that the initial requirements description is complete, which can lead to requirements missing or changes. To address this issue, this paper proposes an unsupervised keyword extraction modeling method. Specifically, using the problem frames model as a case study, this method is integrated into the team's development of an iOS-based Problem Frames (PF) modeling tool in the form of an assisting dropdown list. It is linked with external knowledge bases to predict new keywords. We compare five unsupervised keyword extraction techniques with different principles and evaluate them using samples from the requirements engineering domain. In addition, an eye movement experiment is conducted to further assess the proposed method.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Synthesized Acoustic Vortex-Frequency Comb via Rotational Doppler Effect

引用

Physical Review Letters 2025年第13期134卷 137001-137001页

作者： Yurou Jia Xuan Zhang Suying Zhang Guoqiang Xu Taoyu Chen Hongtao Zhou Yechao Bai Ying Cheng Dajian Wu Department of Physics MOE Key Laboratory of Modern Acoustics Collaborative Innovation Center of Advanced Microstructures Jiangsu Physical Science Research Center Nanjing University Nanjing 210093 China Department of Electrical and Computer Engineering National University of Singapore Singapore Department of Mechanics School of Mechanical Engineering Tianjin University Tianjin 300350 China School of Electronic Science and Engineering Nanjing University Nanjing 210023 China Jiangsu Key Lab on Opto-Electronic Technology School of Physics and Technology Nanjing Normal University Nanjing 210023 China

The synergistic control of light’s frequency and orbital angular momentum (OAM) via integrated nonlinear ring microresonators is crucial for creating spatiotemporal optical waveforms and advancing optical metrology. A direct analog of this capability in acoustics is challenging due to the nonlinearity-based mechanisms’ dependence on external driving and limited frequency conversion efficiency. Here, we present an acoustic synthesized comb structure derived from spatiotemporal metasurfaces, concurrently controlling frequency and OAM characteristics with high conversion efficiency and removing the constraints of critical driving frequency and power level in nonlinear mechanisms. By optimizing comb structures, energy distributions among OAM modes are adjusted to different frequency lines. We finally propose spatiotemporal metalens devices based on vortex-frequency combs, enabling synchronously multiple-order spatial differential operations for parallel data processing. This Letter is groundbreaking in integrating spectral- and spatial-domain acoustic waves, promising multifunctional imaging and advantages in spatiotemporal beams.

关键词： Vortex flow

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：