检索结果-内蒙古大学图书馆

Re-quantization based binary graph neural networks

science China(Information sciences) 2024年第7期67卷 160-171页

作者： Kai-Lang YAO Wu-Jun LI National Key Laboratory for Novel Software Technology Department of Computer Science and TechnologyNanjing University

Binary neural networks have become a promising research topic due to their advantages of fast inference speed and low energy consumption. However, most existing studies focus on binary convolutional neural networks, while less attention has been paid to binary graph neural networks. A common drawback of existing studies on binary graph neural networks is that they still include lots of inefficient full-precision operations in multiplying three matrices and are therefore not efficient enough. In this paper, we propose a novel method, called re-quantization-based binary graph neural networks(RQBGN), for binarizing graph neural networks. Specifically, re-quantization, a necessary procedure contributing to the further reduction of superfluous inefficient full-precision operations, quantizes the results of multiplication between any two matrices during the process of multiplying three matrices. To address the challenges introduced by requantization, in RQBGN we first study the impact of different computation orders to find an effective one and then introduce a mixture of experts to increase the model capacity. Experiments on five benchmark datasets show that performing re-quantization in different computation orders significantly impacts the performance of binary graph neural network models, and RQBGN can outperform other baselines to achieve state-of-the-art performance.

关键词： graph neural networks binary neural networks mixture of experts computation-efficient algorithms

来源：评论

学校读者我要写书评

暂无评论

Stochastic normalized gradient descent with momentum for large-batch training

引用

science China(Information sciences) 2024年第11期67卷 77-91页

作者： Shen-Yi ZHAO Chang-Wei SHI Yin-Peng XIE Wu-Jun LI National Key Laboratory for Novel Software Technology Department of Computer Science and TechnologyNanjing University

Stochastic gradient descent(SGD) and its variants have been the dominating optimization methods in machine learning. Compared with SGD with small-batch training, SGD with large-batch training can better utilize the computational power of current multi-core systems such as graphics processing units(GPUs)and can reduce the number of communication rounds in distributed training settings. Thus, SGD with large-batch training has attracted considerable attention. However, existing empirical results showed that large-batch training typically leads to a drop in generalization accuracy. Hence, how to guarantee the generalization ability in large-batch training becomes a challenging task. In this paper, we propose a simple yet effective method, called stochastic normalized gradient descent with momentum(SNGM), for large-batch training. We prove that with the same number of gradient computations, SNGM can adopt a larger batch size than momentum SGD(MSGD), which is one of the most widely used variants of SGD, to converge to an?-stationary point. Empirical results on deep learning verify that when adopting the same large batch size,SNGM can achieve better test accuracy than MSGD and other state-of-the-art large-batch training methods.

关键词： non-convex problems large-batch training stochastic normalized gradient descent momentum

来源：评论

学校读者我要写书评

暂无评论

CSNet:A Count-Supervised Network via Multiscale MLP-Mixer for Wheat Ear Counting

引用

植物表型组学（英文） 2024年第4期6卷 995-1009页

作者： Yaoxi Li Xingcai Wu Qi Wang Zhixun Pei Kejun Zhao Panfeng Chen Gefei Hao State Key Laboratory of Public Big Data College of Computer Science and TechnologyGuizhou UniversityGuiyang 550025China State Key Laboratory of Public Big Data College of Computer Science and TechnologyGuizhou UniversityGuiyang 550025China Department of Computer Science and Technology Tsinghua UniversityBeijing 100084China State Key Laboratory of Public Big Data College of Computer Science and TechnologyGuizhou UniversityGuiyang 550025China National Key Laboratory of Green Pesticide Key Laboratory of Green Pesticide and Agricultural BioengineeringMinistry of EducationGuiyang 550025China

Wheat is the most widely grown crop in the world,and its yield is closely related to global food *** number of ears is important for wheat breeding and yield ***,automated wheat ear counting techniques are essential for breeding high-yield varieties and increasing grain ***,all existing methods require position-level annotation for training,implying that a large amount of labor is required for annotation,limiting the application and development of deep learning technology in the agricultural *** address this problem,we propose a count-supervised multiscale perceptive wheat counting network(CSNet,count-supervised network),which aims to achieve accurate counting of wheat ears using quantity *** particular,in the absence of location information,CSNet adopts MLP-Mixer to construct a multiscale perception module with a global receptive field that implements the learning of small target attention maps between wheat ear *** conduct comparative experiments on a publicly available global wheat head detection dataset,showing that the proposed count-supervised strategy outperforms existing position-supervised methods in terms of mean absolute error(MAE)and root mean square error(RMSE).This superior performance indicates that the proposed approach has a positive impact on improving ear counts and reducing labeling costs,demonstrating its great potential for agricultural counting *** code is available at .

关键词： network counting csnet mlp-mixer multiscale supervised wheat

来源：评论

学校读者我要写书评

暂无评论

Enhancing Entity Relationship Extraction in Dialogue Texts Using Hypergraph and Heterogeneous Graph

引用

Chinese Journal of Electronics 2025年第1期34卷 295-308页

作者： Shunmiao Zhang Siyuan Zheng Degen Huang Dan Li School of Computer Science and Mathematics Fujian University of Technology Fujian Provincial Key Laboratory of Big Data Mining and Applications Fujian University of Technology School of Computer Science and Technology Dalian University of Technology Elsevier

Dialogue-based relation extraction(DialogRE) aims to predict relationships between two entities in dialogue. Current approaches to dialogue relationship extraction grapple with long-distance entity relationships in dialogue data as well as complex entity relationships, such as a single entity with multiple types of connections. To address these issues, this paper presents a novel approach for dialogue relationship extraction termed the hypergraphs and heterogeneous graphs model(HG2G). This model introduces a two-tiered structure, comprising dialogue hypergraphs and dialogue heterogeneous graphs, to address the shortcomings of existing methods. The dialogue hypergraph establishes connections between similar nodes using hyper-edges and utilizes hypergraph convolution to capture multi-level features. Simultaneously, the dialogue heterogeneous graph connects nodes and edges of different types, employing heterogeneous graph convolution to aggregate cross-sentence information. Ultimately, the integrated nodes from both graphs capture the semantic nuances inherent in dialogue. Experimental results on the DialogRE dataset demonstrate that the HG2G model outperforms existing state-of-the-art methods.

关键词： Convolution Aggregates Semantics Feature extraction Data mining

来源：评论

学校读者我要写书评

暂无评论

YOLO-Drone:A Scale-Aware Detector for Drone Vision

引用

Chinese Journal of Electronics 2024年第4期33卷 1034-1045页

作者： Yutong LI Miao MA Shichang LIU Chao YAO Longjiang GUO School of Computer Science Shaanxi Normal University Key Laboratory of Modern Teaching Technology Ministry of Education College of Computer Science Sichuan University

Object detection is an important task in drone vision. Since the number of objects and their scales always vary greatly in the drone-captured video, small object-oriented feature becomes the bottleneck of model performance, and most existing object detectors tend to underperform in drone-vision scenes. To solve these problems, we propose a novel detector named YOLO-Drone. In the proposed detector, the backbone of YOLO is firstly replaced with ConvNeXt, which is the state-of-the-art one to extract more discriminative features. Then, a novel scale-aware attention(SAA) module is designed in detection head to solve the large disparity scale problem. A scale-sensitive loss(SSL) is also introduced to put more emphasis on object scale to enhance the discriminative ability of the proposed detector. Experimental results on the latest VisDrone 2022 test-challenge dataset(detection track) show that our detector can achieve average precision(AP) of 39.43%, which is tied with the previous state-of-the-art, meanwhile,reducing 39.8% of the computational cost.

关键词： YOLO Costs Object oriented modeling Computational modeling Detectors Feature extraction Data models

来源：评论

学校读者我要写书评

暂无评论

A Distributed Self-Tallying Electronic Voting System Using the Smart Contract

引用

Chinese Journal of Electronics 2024年第4期33卷 1063-1076页

作者： Jingyu YAO Bo YANG Tao WANG Wenzheng ZHANG School of Computer Science Shaanxi Normal University State Key Laboratory of Integrated Services Networks Xidian University Science and Technology on Communication Security Laboratory

For electronic voting(e-voting) with a trusted authority, the ballots may be discarded or tampered, so it is attractive to eliminate the dependence on the trusted party. An e-voting protocol, where the final voting result can be calculated by any entity, is known as self-tallying e-voting protocol. To the best of our knowledge, addressing both abortive issue and adaptive issue simultaneously is still an open problem in self-tallying e-voting *** Ethereum blockchain with cryptographic technologies, we present a decentralized self-tallying e-voting protocol. We solve the above problem efficiently by utilizing optimized Group Encryption Scheme and standard Exponential El Gamal Cryptosystem. We use zero-knowledge proof and homomorphic encryption to protect votes' secrecy and achieve self-tallying. All ballots can be verified by anyone and the final voting result can be calculated by any entity. In addition, using the paradigm of score voting and “1-out-of-k” proof, our e-voting system is suitable for a wide range of application scenarios. Experiments show that our protocol is more competitive and more suitable for large-scale voting.

关键词： Electronic voting systems Smart contracts Blockchains Cryptocurrency Electronic voting Homomorphic encryption Standards

来源：评论

学校读者我要写书评

暂无评论

Block-gram:Mining knowledgeable features for efficiently smart contract vulnerability detection

引用

Digital Communications and Networks 2025年第1期11卷 1-12页

作者： Xueshuo Xie Haolong Wang Zhaolong Jian Yaozheng Fang Zichun Wang Tao Li Tianjin Key Laboratory of Network and Data Security Technology TianjinChina College of Computer Science Nankai UniversityTianjinChina Key Laboratory of Blockchain and Cyberspace Governance of Zhejiang Province China State Key Laboratory of Computer Architecture Institute of Computing TechnologyChinese Academy of SciencesChina

Smart contracts are widely used on the blockchain to implement complex transactions,such as decentralized applications on *** vulnerability detection of large-scale smart contracts is critical,as attacks on smart contracts often cause huge economic *** it is difficult to repair and update smart contracts,it is necessary to find the vulnerabilities before they are ***,code analysis,which requires traversal paths,and learning methods,which require many features to be trained,are too time-consuming to detect large-scale on-chain ***-based methods will obtain detection models from a feature space compared to code analysis methods such as symbol *** the existing features lack the interpretability of the detection results and training model,even worse,the large-scale feature space also affects the efficiency of *** paper focuses on improving the detection efficiency by reducing the dimension of the features,combined with expert *** this paper,a feature extraction model Block-gram is proposed to form low-dimensional knowledge-based features from ***,the metadata is separated and the runtime code is converted into a sequence of opcodes,which are divided into segments based on some instructions(jumps,etc.).Then,scalable Block-gram features,including 4-dimensional block features and 8-dimensional attribute features,are mined for the learning-based model ***,feature contributions are calculated from SHAP values to measure the relationship between our features and the results of the detection *** addition,six types of vulnerability labels are made on a dataset containing 33,885 contracts,and these knowledge-based features are evaluated using seven state-of-the-art learning algorithms,which show that the average detection latency speeds up 25×to 650×,compared with the features extracted by N-gram,and also can enhance the interpretability of the detection model.

关键词： Smart contract Bytecode&opcode Knowledgeable features Vulnerability detection Feature contribution

来源：评论

学校读者我要写书评

暂无评论

GPS: graph contrastive learning via multi-scale augmented views from adversarial pooling

引用

science China(Information sciences) 2025年第1期68卷 145-158页

作者： Wei JU Yiyang GU Zhengyang MAO Ziyue QIAO Yifang QIN Xiao LUO Hui XIONG Ming ZHANG School of Computer Science National Key Laboratory for Multimedia Information ProcessingPeking University Artificial Intelligence Thrust The Hong Kong University of Science and Technology Department of Computer Science University of California

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 *** 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.

关键词： graph representation learning graph neural networks graph contrastive learning graph augmentations graph pooling

来源：评论

学校读者我要写书评

暂无评论

The quantum Ising model for perfect matching and solving it with variational quantum eigensolver

引用

science China(Information sciences) 2024年第9期67卷 306-320页

作者： Qilin ZHENG Pingyu ZHU Chao WU Miaomiao YU Weihong LUO Ping XU Institute for Quantum Information and State Key Laboratory of High Performance Computing College of Computer Science and Technology National University of Defense Technology Hefei National Laboratory

Obtaining all perfect matchings of a graph is a tough problem in graph theory, and its complexity belongs to the #P-Complete class. The problem is closely related to combinatorics, marriage matching problems, dense subgraphs, the Gaussian boson sampling, chemical molecular structures, and dimer *** this paper, we propose a quadratic unconstrained binary optimization formula of the perfect matching problem and translate it into the quantum Ising model. We can obtain all perfect matchings by mapping them to the ground state of the quantum Ising Hamiltonian and solving it with the variational quantum eigensolver. Adjusting the model's parameters can also achieve the maximum or minimum weighted perfect matching. The experimental results on a superconducting quantum computer of the Origin Quantum Computing technology Company show that our model can encode 2~n dimensional optimization space with only O（n） qubits consumption and achieve a high success probability of the ground state corresponding to all perfect matchings. In addition, the further simulation results show that the model can support a scale of more than 14 qubits, effectively resist the adverse effects of noise, and obtain a high success probability at a shallow variational depth. This method can be extended to other combinatorial optimization problems.

关键词： perfect matching Ising model quantum Hamiltonian variational quantum eigensolver quadratic unconstrained binary optimization

来源：评论

学校读者我要写书评

暂无评论

For LEO Satellite Networks: Intelligent Interference Sensing and Signal Reconstruction Based on Blind Separation technology

引用

China Communications 2024年第2期21卷 85-95页

作者： Chengjie Li Lidong Zhu Zhen Zhang The Key Laboratory for Computer Systems of State Ethnic Affairs Commission School of Computer Science and Technology(Southwest Minzu University)Chengdu 610000China National Key Laboratory of Science and Technology on Communications(University of Electronic Science and Technology of China) Chengdu 610000China College of Computer Science(Sichuan University) Chengdu 610000China

In LEO satellite communication networks,the number of satellites has increased sharply, the relative velocity of satellites is very fast, then electronic signal aliasing occurs from time to time. Those aliasing signals make the receiving ability of the signal receiver worse, the signal processing ability weaker,and the anti-interference ability of the communication system lower. Aiming at the above problems, to save communication resources and improve communication efficiency, and considering the irregularity of interference signals, the underdetermined blind separation technology can effectively deal with the problem of interference sensing and signal reconstruction in this scenario. In order to improve the stability of source signal separation and the security of information transmission, a greedy optimization algorithm can be executed. At the same time, to improve network information transmission efficiency and prevent algorithms from getting trapped in local optima, delete low-energy points during each iteration process. Ultimately, simulation experiments validate that the algorithm presented in this paper enhances both the transmission efficiency of the network transmission system and the security of the communication system, achieving the process of interference sensing and signal reconstruction in the LEO satellite communication system.

关键词： blind source separation greedy optimization algorithm interference sensing LEO satellite communication networks signal reconstruction

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：