检索结果-内蒙古大学图书馆

arXiv 2024年

作者： Ju, Wei Gu, Yiyang Mao, Zhengyang Qiao, Ziyue Qin, Yifang Luo, Xiao Xiong, Hui Zhang, Ming School of Computer Science National Key Laboratory for Multimedia Information Processing Peking University Beijing100871 China The Hong Kong University of Science and Technology Guangzhou511453 China Department of Computer Science University of California Los Angeles90095 United States

Self-supervised graph representation learning has recently shown considerable promise in a range of fields, including bioinformatics and social networks. A large number of graph contrastive learning approaches have shown promising performance for representation learning on graphs, which train models by maximizing agreement between original graphs and their augmented views (i.e., positive views). Unfortunately, these methods usually involve pre-defined augmentation strategies based on the knowledge of human experts. Moreover, these strategies may fail to generate challenging positive views to provide sufficient supervision signals. In this paper, we present a novel approach named Graph Pooling ContraSt (GPS) to address these issues. Motivated by the fact that graph pooling can adaptively coarsen the graph with the removal of redundancy, we rethink graph pooling and leverage it to automatically generate multi-scale positive views with varying emphasis on providing challenging positives and preserving semantics, i.e., strongly-augmented view and weakly-augmented view. Then, we incorporate both views into a joint contrastive learning framework with similarity learning and consistency learning, where our pooling module is adversarially trained with respect to the encoder for adversarial robustness. Experiments on twelve datasets on both graph classification and transfer learning tasks verify the superiority of the proposed method over its counterparts. Copyright © 2024, The Authors. All rights reserved.

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Efficient Routing Protection Algorithm Based on Optimized Network Topology

引用

computers, Materials & Continua 2022年第9期72卷 4525-4540页

作者： Haijun Geng Zikun Jin Jiangyuan Yao Han Zhang Zhiguo Hu Bo Yang Yingije Guo Wei Wang Qidong Zhang Guoao Duan School of Computer and Information Technology Shanxi UniversityShanxi030006China School of Automation and Software Engineering Shanxi UniversityShanxi030006China School of Computer Science and Technology Hainan UniversityHaikou570228China Department of Computer Science&Technology Tsinghua UniversityBeijing100084China Department of Urban&Regional Planning San JoséState University San JoséCA95192USA Key Laboratory of Embedded System and Service Computing(Tongji University) Ministry of EducationShanghai201804China Institute of Fundamental and Frontier Sciences University of Electronic Science and Technology of ChinaChengdu610054China Beijinng University of Technology Beijing-Dublin International CollegeBeijing University of TechnologyBeijing100124China

Network failures are unavoidable and occur *** the network fails,intra-domain routing protocols deploying on the Internet need to undergo a long convergence *** this period,a large number of messages are discarded,which results in a decline in the user experience and severely affects the quality of service of Internet Service Providers(ISP).Therefore,improving the availability of intra-domain routing is a trending research question to be *** usually employs routing protection algorithms to improve intra-domain routing ***,existing routing protection schemes compute as many backup paths as possible to reduce message loss due to network failures,which increases the cost of the network and impedes the methods deployed in *** address the issues,this study proposes an efficient routing protection algorithm based on optimized network topology(ERPBONT).ERPBONT adopts the optimized network topology to calculate a backup path with the minimum path coincidence degree with the shortest path for all source ***,the backup path with the minimum path coincidence with the shortest path is described as an integer programming *** the simulated annealing algorithm ERPBONT is used to find the optimal ***,the algorithm is tested on the simulated topology and the real *** experimental results show that ERPBONT effectively reduces the path coincidence between the shortest path and the backup path,and significantly improves the routing availability.

关键词： Routing protection algorithm network availability network failure shortest path backup path

来源：评论

学校读者我要写书评

暂无评论

Tackling Noisy Labels with Network Parameter Additive Decomposition

arXiv

引用

arXiv 2024年

作者： Wang, Jingyi Xia, Xiaobo Lan, Long Wu, Xinghao Yu, Jun Yang, Wenjing Han, Bo Liu, Tongliang The Department of Intelligent Data Science College of Computer Science and Technology National University of Defense Technology Changsha410073 China The Trustworthy Machine Learning Lab School of Computer Science Faculty of Engineering University of Sydney DarlingtonNSW2008 Australia The State Key Laboratory of Virtual Reality Technology and Systems School of Computer Science and Engineering Beihang University Beijing100191 China The Department of Automation University of Science and Technology of China Hefei230026 China The Department of Computer Science Hong Kong Baptist University Hong Kong

Given data with noisy labels, over-parameterized deep networks suffer overfitting mislabeled data, resulting in poor generalization. The memorization effect of deep networks shows that although the networks have the ability to memorize all noisy data, they would first memorize clean training data, and then gradually memorize mislabeled training data. A simple and effective method that exploits the memorization effect to combat noisy labels is early stopping. However, early stopping cannot distinguish the memorization of clean data and mislabeled data, resulting in the network still inevitably overfitting mislabeled data in the early training stage. In this paper, to decouple the memorization of clean data and mislabeled data, and further reduce the side effect of mislabeled data, we perform additive decomposition on network parameters. Namely, all parameters are additively decomposed into two groups, i.e. , parameters w are decomposed as w = σ + γ. Afterward, the parameters σ are considered to memorize clean data, while the parameters γ are considered to memorize mislabeled data. Benefiting from the memorization effect, the updates of the parameters σ are encouraged to fully memorize clean data in early training, and then discouraged with the increase of training epochs to reduce interference of mislabeled data. The updates of the parameters γ are the opposite. In testing, only the parameters σ are employed to enhance generalization. Extensive experiments on both simulated and real-world benchmarks confirm the superior performance of our method. The implementation is available at: https://***/wangjingyi9924/TNLPAD. Copyright © 2024, The Authors. All rights reserved.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

DifAttack++: Query-Efficient Black-Box Adversarial Attack via Hierarchical Disentangled Feature Space in Cross Domain

arXiv

引用

arXiv 2024年

作者： Liu, Jun Zhou, Jiantao Zeng, Jiandian Tian, Jinyu Li, Zheng State Key Laboratory of Internet of Things for Smart City Department of Computer and Information Science University of Macau China Institute of Artificial Intelligence and Future Networks Beijing Normal University China School of Computer Science and Engineering Macau University of Science and Technology China

This work investigates efficient score-based black-box adversarial attacks with a high Attack Success Rate (ASR) and good generalizability. We design a novel attack method based on a hierarchical DIsentangled Feature space, called DifAttack++, which differs significantly from the existing ones operating over the entire feature space. Specifically, DifAttack++ firstly disentangles an image’s latent feature into an Adversarial Feature (AF) and a Visual Feature (VF) via an autoencoder equipped with our specially designed Hierarchical Decouple-Fusion (HDF) module, where the AF dominates the adversarial capability of an image, while the VF largely determines its visual appearance. We train such two autoencoders for the clean and adversarial image domains (i.e., cross-domain) respectively to achieve image reconstructions and feature disentanglement, by using pairs of clean images and their Adversarial Examples (AEs) generated from available surrogate models via white-box attack methods. Eventually, in the black-box attack stage, DifAttack++ iteratively optimizes the AF according to the query feedback from the victim model until a successful AE is generated, while keeping the VF unaltered. Extensive experimental results demonstrate that our DifAttack++ leads to superior ASR and query efficiency than state-of-the-art methods, meanwhile exhibiting much better visual quality of AEs. The code is available at https://***/csjunjun/***. © 2024, CC BY-NC-ND.

关键词： Codes (symbols)

来源：评论

学校读者我要写书评

暂无评论

Approximating parameters of photovoltaic models using an amended reptile search algorithm

引用

Journal of Ambient Intelligence and Humanized Computing 2023年第7期14卷 9073-9088页

作者： Chauhan, Sumika Vashishtha, Govind Kumar, Anil Department of Electrical and Instrumentation Engineering Sant Longowal Institute of Engineering and Technology Longowal148 106 India Precision Metrology Laboratory Department of Mechanical Engineering Sant Longowal Institute of Engineering and Technology Longowal148 106 India College of Mechanical and Electrical Engineering Wenzhou University Wenzhou325 035 China Department of Computer Science and Engineering National Institute of Technology Delhi New Delhi110036 India

The appropriate selection of parameters of photovoltaic models is necessary for an accurate evaluation, control, and optimization of photovoltaic systems. Even though various strategies have been developed to address this issue, but, a precise and reliable scheme for identifying the model parameters remains a challenge. To improve parameter identification of different photovoltaic models, an opposition-based learning reptile search algorithm with Cauchy mutation strategy (OBL-RSACM) is introduced in this research. In OBL-RSACM, the individuals in search space get doubled by generating their opposite guess of the solution which overcomes the issue of strucking of solution in local minima and also enhances the convergence speed. Cauchy mutation strategy is also incorporated in the basic reptile search algorithm (RSA) which enhances the search mechanism, modifies the control parameter, mutation-driven scheme, and greedy approach of selection during the search process of the RSA. Thus, improves the exploration process and maintains the proper balance between exploration and exploitation. The proposed OBL-RSACM is applied to estimate the parameters of different photovoltaic models, i.e., single diode, double diode, and photovoltaic module. A comprehensive comparison of experimental results and analysis demonstrated that OBL-RSACM outperformed other state-of-the-art algorithms in terms of accuracy, reliability, and computational efficacy. © 2022, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.

关键词： Learning algorithms

来源：评论

学校读者我要写书评

暂无评论

Toward Enhancing Cross-Lingual Domain Knowledge Sharing and Transferring for Multilingual Domain Adaptation in NMT

IEEE Transactions on Audio, Speech and Language Processing

引用

IEEE Transactions on Audio, Speech and Language Processing 2025年 33卷 2125-2138页

作者： Zhengtao Yu Jinlei Xu Xiang Huang Zhongfen Deng Yonghua Wen Li Sun Yuxin Huang Hao Peng Philip S. Yu Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming China Yunnan Province Key Laboratory of Artificial Intelligence Kunming University of Science and Technology Kunming China School of Cyber Science and Technology Beihang University Beijing China Department of Computer Science University of Illinois Chicago IL USA School of Control and Computer Engineering North China Electric Power University Beijing China

Multilingual Neural Machine Translation (NMT) excels in sharing knowledge across languages and transferring insights from high-resource languages to improve performance in low-resource languages. However, its performance lags in specific domains such as legal and medical. Previous works have focused on adding language-specific and domain-specific adapters to achieve domain adaptation. Although effective, these adapter-based methods only use domain data to train additional parameters, limiting the performance of multilingual NMT. In this paper, we propose CDSTX, a novel approach that achieves robust multilingual domain adaptation without relying on adapters, solely leveraging multilingual models. Specifically, we utilize the pretrained model XLM-R and frozen embeddings to preserve its robust multilingual capabilities and design a two-stage training strategy for domain adaptation. This includes separate training for the decoder and fine-tuning of the entire model, ensuring that the model effectively acquires domain knowledge and correctly represents domain-specific text. Moreover, to better utilize the domain information conveyed implicitly by the training data, we devise special domain tokens at the beginning of the source and target sentences called Source&Target Domain Tags. In addition, back translation is employed to enhance the cross-lingual transfer ability of our approach. Our proposed method is evaluated on two datasets across three translation tasks: Single-domain Multilingual NMT, Multi-domain bilingual NMT, and Multilingual Multi-domain NMT. Notably, in the single-domain multilingual NMT task, CDSTX significantly enhances zero-shot domain translation performance, achieving improvements of up to +30 BLEU points through the utilization of back-translation techniques. Even in the bilingual multi-domain NMT task where specific domain data for the target translation direction is unavailable, our method consistently outperforms all SOTA methods. Moreover, in the

关键词： Training Adaptation models Translation Limiting Law Large language models Training data Neural machine translation Multilingual Speech processing

来源：评论

学校读者我要写书评

暂无评论

Radiologist-in-the-Loop Self-Training for Generalizable CT Metal Artifact Reduction

引用

IEEE Transactions on Medical Imaging 2025年第6期44卷 2504-2514页

作者： Ma, Chenglong Li, Zilong Li, Yuanlin Han, Jing Zhang, Junping Zhang, Yi Liu, Jiannan Shan, Hongming Institute of Science and Technology for Brain-Inspired Intelligence MOE Frontiers Center for Brain Science Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence Fudan University Shanghai200433 China Fudan University Shanghai Key Laboratory of Intelligent Information Processing School of Computer Science Shanghai200433 China Shanghai Jiao Tong University School of Medicine Department of Oral Maxillofacial Head and Neck Oncology Shanghai Ninth People’s Hospital Shanghai200025 China Shanghai Jiao Tong University College of Stomatology Shanghai200240 China National Center for Stomatology Shanghai200011 China Shanghai Research Institute of Stomatology National Clinical Research Center for Oral Diseases Shanghai Key Laboratory of Stomatology Shanghai200011 China Sichuan University School of Cyber Science and Engineering Sichuan Chengdu610065 China

Metal artifacts in computed tomography (CT) images can significantly degrade image quality and impede accurate diagnosis. Supervised metal artifact reduction (MAR) methods, trained using simulated datasets, often struggle to perform well on real clinical CT images due to a substantial domain gap. Although state-of-the-art semi-supervised methods use pseudo ground-truths generated by a prior network to mitigate this issue, their reliance on a fixed prior limits both the quality and quantity of these pseudo ground-truths, introducing confirmation bias and reducing clinical applicability. To address these limitations, we propose a novel radiologist-in-the-loop self-training framework for MAR, termed RISE-MAR, which can integrate radiologists’ feedback into the semi-supervised learning process, progressively improving the quality and quantity of pseudo ground-truths for enhanced generalization on real clinical CT images. For quality assurance, we introduce a clinical quality assessor model that emulates radiologist evaluations, effectively selecting high-quality pseudo ground-truths for semi-supervised training. For quantity assurance, our self-training framework iteratively generates additional high-quality pseudo ground-truths, expanding the clinical dataset and further improving model generalization. Extensive experimental results on multiple clinical datasets demonstrate the superior generalization performance of our RISE-MAR over state-of-the-art methods, advancing the development of MAR models for practical application. © 1982-2012 IEEE.

关键词： computerized tomography

来源：评论

学校读者我要写书评

暂无评论

Privacy Preserving Quantum Search Mechanism Using Grover's Algorithm

Privacy Preserving Quantum Search Mechanism Using Grover's A...

引用

Quantum Communications, Networking, and Computing (QCNC), International Conference on

作者： Keyi Ju Xiaoqi Qin Hui Zhong Xinyue Zhang Miao Pan Baoling Liu State Key Laboratory of Networking and Switching Technology Beijing University of Posts and Telecommunications Beijing China Department of Electrical and Computer Engineering University of Houston Houston TX Department of Computer Science Kennesaw State University Marietta GA

ISBN: (数字)9798350366778

ISBN: (纸本)9798350366785

Quantum computing changes the way of solving complex problems and handling vast datasets. The introduction of quantum search algorithms, notably Grover's algorithm, addresses the inefficiencies of classical search methods in handling vast datasets by offering substantial speed improvements. However, an inherent aspect of quantum searches is their failure probability. This paper aims to explore the relationship between the failure probability in Grover's algorithm and the privacy guarantees of differential privacy (DP). We further investigate the impact of several quantum circuit parameters, including the number of shots, the number of solutions within the dataset, and circuit noise levels, on this failure probability. Our analysis demonstrates the indirect connection between these quantum circuit elements and the privacy guarantees of DP. Through Monte Carlo simulations, we offer an in-depth analysis that paves the way for secure and faster data querying methods.

关键词： Privacy Differential privacy Monte Carlo methods Search methods Circuit noise Quantum circuit

来源：评论

学校读者我要写书评

暂无评论

Weakly Supervised Salient Object Detection With Oversize Bounding Box

引用

International Journal of computer Vision 2025年 1-20页

作者： Wu, Zhihao Xu, Yong Yang, Jian Zhang, David School of Artificial Intelligence Shenzhen University Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China School of Computer Science and Technology Harbin Institute of Technology Shenzhen China Department of Computer Science and Engineering Nanjing University of Science and Technology Nanjing China School of Data Science The Chinese University of Hong Kong (Shenzhen) Shenzhen China

Compared to laborious pixel-level annotations, scribbles, bounding boxes, and points are much more efficient in salient object detection (SOD). However, the annotation cost of these forms linearly increases with the number of salient objects in the image, which is not ideal for real-world scenarios. To address this issue, we propose a novel annotation form called oversize bounding box (OBB), i.e., a box that encompasses all salient objects without the need for tight enclosure. It has two characteristics: (1) All pixels outside the box are from the background. (2) A subset of the pixels inside the box belongs to salient objects. Therefore, the core issue is how to highlight integral and accurate object regions as pseudo-labels. Inspired by the powerful visual understanding and vision-language correlations demonstrated by large multimodal models, we devise the first multimodal model-based pseudo-label generation method for SOD. It utilizes MiniGPT-4 to generate descriptions of salient objects as text prompts for CLIP, combined with the CAM-based technique and proposed refinement algorithm based on the multi-head self-attention and superpixel to activate the object regions. Then we use OBB to further correct the activation regions based on its two characteristics. Given the potential errors of pseudo-labels within the box, we propose a center-pixel-based cross entropy loss, as activated pixels closer to the center are generally more reliable. Moreover, we establish an OBB-supervised training dataset by relabeling the DUTS dataset. Extensive experiments on six benchmarks demonstrate that our method achieves the state-of-the-art performance.

关键词： Weak supervision Salient object detection Multimodal large model Activation map

来源：评论

学校读者我要写书评

暂无评论

SIMAC: A Semantic-Driven Integrated Multimodal Sensing And Communication Framework

arXiv

引用

arXiv 2025年

作者： Peng, Yubo Xiang, Luping Yang, Kun Jiang, Feibo Wang, Kezhi Wu, Dapeng Oliver State Key Laboratory of Novel Software Technology Nanjing University Nanjing China Suzhou China School of Information Science and Engineering Hunan Normal University Changsha China Department of Computer Science Brunel University London United Kingdom Department of Computer Science City University of Hong Kong Hong Kong

Traditional single-modality sensing faces limitations in accuracy and capability, and its decoupled implementation with communication systems increases latency in bandwidth-constrained environments. Additionally, single-task-oriented sensing systems fail to address users’ diverse demands. To overcome these challenges, we propose a semantic-driven integrated multimodal sensing and communication (SIMAC) framework. This framework leverages a joint source-channel coding architecture to achieve simultaneous sensing decoding and transmission of sensing results. Specifically, SIMAC first introduces a multimodal semantic fusion (MSF) network, which employs two extractors to extract semantic information from radar signals and images, respectively. MSF then applies cross-attention mechanisms to fuse these unimodal features and generate multimodal semantic representations. Secondly, we present a large language model (LLM)-based semantic encoder (LSE), where relevant communication parameters and multimodal semantics are mapped into a unified latent space and input to the LLM, enabling channel-adaptive semantic encoding. Thirdly, a task-oriented sensing semantic decoder (SSD) is proposed, in which different decoded heads are designed according to the specific needs of tasks. Simultaneously, a multi-task learning strategy is introduced to train the SIMAC framework, achieving diverse sensing services. Finally, experimental simulations demonstrate that the proposed framework achieves diverse sensing services and higher accuracy. Copyright © 2025, The Authors. All rights reserved.

关键词： Decoding

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：