检索结果-内蒙古大学图书馆

Interspeech Conference

作者： Li, Yuhang Wei, Xiao Si, Yuke Wang, Longbiao Wang, Xiaobao Dang, Jianwu Tianjin Univ Coll Intelligence & Comp Tianjin Key Lab Cognit Comp & Applicat Tianjin Peoples R China

Slot filling is an essential component in task-oriented dialogue systems. Due to the scarcity of annotated data, zero-shot slot filling has been studied to transfer knowledge from source domains to a target domain. Previous methods adopt slot descriptions or questions as slot semantics, where they utilize slot descriptions to calculate similarity scores, or reformat the task as a question-answering problem. However, these methods do not fully exploit the token-level dependency between the slot semantics and utterances. In this study, we propose a Transformer-based Slot semantics fusion method for Slot Filling (TSSF). We first adopt two encoders with shared weights to obtain the representations of utterances and slot semantics. Then, we design a transformer-based fusion module for effectively integrating slot semantics into utterances. Experimental results on the public benchmark SNIPS show that our model significantly outperforms the state-of-the-art model by 6.09% in terms of slot F1.

关键词： dialogue system zero-shot slot filling transformer-based slot semantics fusion

来源：评论

学校读者我要写书评

暂无评论

DFier: A directed vulnerability verifier for Ethereum smart contracts

引用

JOURNAL OF NETWORK AND compUTER APPLICATIONS 2024年 231卷

作者： Wang, Zeli Dai, Weiqi Li, Ming Choo, Kim-Kwang Raymond Zou, Deqing Chongqing Univ Posts & Telecommun Chongqing Key Lab Computat Intelligence Key Lab Big Data Intelligent Comp Key Lab Cyberspace Big Data Intelligent SecurMini Chongqing 400065 Peoples R China Huazhong Univ Sci & Technol Natl Engn Res Ctr Big Data Technol & Syst Hubei Engn Res Ctr Big Data Secur Serv Comp Technol & Syst LabHubei Key Lab Distrib Wuhan 430074 Peoples R China Univ Texas San Antonio Dept Informat Syst & Cyber Secur San Antonio TX USA

Smart contracts are self-executing digital agreements that automatically enforce the terms between parties, playing a crucial role in blockchain systems. However, due to the potential losses of digital assets caused by vulnerabilities, the security issues of Ethereum smart contracts have garnered widespread attention. To address this, researchers have developed various techniques to detect vulnerabilities in smart contracts, with fuzzing techniques achieving promising results. Nonetheless, current fuzzers are unable to effectively exercise suspicious targets because they overlook two key factors: comprehensively exploring all paths to the targets and providing high-quality directed seed inputs. This paper presents a Directed vulnerability veriFier (DFier), which elaborates effective transaction sequences with directed inputs for the fuzzer. This focuses on exploring target paths and automatically validating whether the specified locations are vulnerable. Specifically, DFier employs static analysis to help locate target paths, facilitating their comprehensive exploration. Additionally, we devise three heuristic strategies to enable our fuzzing technique to generate directed inputs that effectively validate the targets. Extensive experiments demonstrate that DFier is effective in verifying contract security, compared with three existing contract fuzzers (i.e., contractFuzzer, sFuzz, and conFuzzius), while the performance losses are in an acceptable range.

关键词： Blockchain Smart contract Vulnerability detection Fuzzing

来源：评论

学校读者我要写书评

暂无评论

Parallax-aware dual-view feature enhancement and adaptive detail compensation for dual-pixel defocus deblurring

引用

ENGINEERING APPLICATIONS OF ARTIFICIAL intelligence 2025年 139卷

作者： Niu, Yuzhen He, Yuqi Xu, Rui Li, Yuezhou Chen, Yuzhong Fuzhou Univ Coll Comp & Data Sci Fujian Key Lab Network Comp & Intelligent Informat Fuzhou 350108 Peoples R China Minist Educ Engn Res Ctr BigData Intelligence Fuzhou 350108 Peoples R China

Defocus deblurring using dual-pixel sensors has gathered significant attention in recent years. However, current methodologies have not adequately addressed the challenge of defocus disparity between dual views, resulting in suboptimal performance in recovering details from severely defocused pixels. To counteract this limitation, we introduce in this paper a parallax-aware dual-view feature enhancement and adaptive detail compensation network (PA-Net), specifically tailored for dual-pixel defocus deblurring task. Our proposed PA- Net leverages an encoder-decoder architecture augmented with skip connections, designed to initially extract distinct features from the left and right views. A pivotal aspect of our model lies at the network's bottleneck, where we introduce a parallax-aware dual-view feature enhancement based on Transformer blocks, which aims to align and enhance extracted dual-pixel features, aggregating them into a unified feature. Furthermore, taking into account the disparity and the rich details embedded in encoder features, we design an adaptive detail compensation module to adaptively incorporate dual-view encoder features into image reconstruction, aiding in restoring image details. Experimental results demonstrate that our proposed PA-Net exhibits superior performance and visual effects on the real-world dataset.

关键词： Defocus deblurring Dual-pixel Defocus disparity Image restoration Detail restoration

来源：评论

学校读者我要写书评

暂无评论

Effective Value Analysis of Fuzzy Similarity Relation in HQSS for Efficient Granulation

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2024年第9期35卷 12849-12863页

作者： Zhang, Qinghua Zhao, Fan Cheng, Yunlong Gao, Man Wang, Guoyin Xia, Shuyin Ding, Weiping Chongqing Univ Posts & Telecommun Chongqing Key Lab Tourism Multisource Data Percept Key Lab Big Data Intelligence Comp Minist Culture & Tourism Chongqing 400065 Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Computat Intelligence Chongqing 400065 Peoples R China Chongqing Univ Posts & Telecommun Sch Comp Sci & Technol Chongqing 400065 Peoples R China Nantong Univ Sch Informat Sci & Technol Nantong 226019 Peoples R China

Hierarchical quotient space structure (HQSS), as a typical description of granular computing (GrC), focuses on hierarchically granulating fuzzy data and mining hidden knowledge. The key step of constructing HQSS is to transform the fuzzy similarity relation into fuzzy equivalence relation. However, on one hand, the transformation process has high time complexity. On the other hand, it is difficult to mine knowledge directly from fuzzy similarity relation due to its information redundancy, i.e., sparsity of effective information. Therefore, this article mainly focuses on proposing an efficient granulation approach for constructing HQSS by quickly extracting the effective value of fuzzy similarity relation. First, the effective value and effective position of fuzzy similarity relation are defined according to whether they could be retained in fuzzy equivalence relation. Second, the number and composition of effective values are presented to confirm that which elements are effective values. Based on these above theories, redundant information and sparse effective information in fuzzy similarity relation could be completely distinguished. Next, both isomorphism and similarity between two fuzzy similarity relations are researched based on the effective value. The isomorphism between two fuzzy equivalence relations is discussed based on the effective value. Then, the algorithm with low time complexity for extracting effective values of fuzzy similarity relation is introduced. On the basis, the algorithm for constructing HQSS is presented to realize efficient granulation of fuzzy data. The proposed algorithms could accurately extract effective information from the fuzzy similarity relation and construct the same HQSS with the fuzzy equivalence relation while greatly reducing the time complexity. Finally, relevant experiments on 15 UCI datasets, 3 UKB datasets, and 5 image datasets are shown and analyzed to verify the effectiveness and efficiency of the proposed algorithm.

关键词： Granular computing Time complexity data mining Rough sets Fuzzy sets data models Fans Fuzzy set fuzzy similarity relation granular computing (GrC) granulation hierarchical quotient space structure (HQSS)

来源：评论

学校读者我要写书评

暂无评论

Secure and Efficient Industrial Wireless Sensor Networks Protocol Based on Cancelable Biometrics

引用

IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS 2024年第12期20卷 13580-13590页

作者： Hu, Yao Bian, Weixin Xie, Dong Xu, Deqin Xu, Zilong Anhui Normal Univ Sch Comp & Informat Wuhu 241002 Peoples R China Anhui Prov Key Lab Ind Intelligence Data Secur Wuhu 241002 Peoples R China

This article discusses the application of wireless sensor networks (WSNs) in industrial settings, particularly in the context of the Internet of Things. The advantages of WSNs, such as convenience, low energy consumption, and real-time monitoring, make them highly suitable for industrial applications. However, these networks face several challenges, including signal interference, security concerns, privacy issues, and battery life. This article explores these challenges and proposes potential solutions. Current industrial wireless sensor protocols often prioritize lightweight or real-time aspects, potentially neglecting security considerations. This article proposes a strategy for protecting user biometrics during authentication using cancelable biometrics. The approach utilizes sliding window hashing, fuzzy extractors, and physical unclonable functions to generate a cancelable biometric template. The proposed industrial wireless sensor protocols will be used in this strategy. This protocol aims to safeguard users' biometric information in the event of a data breach while detecting and preventing potential attacks on the sensors.

关键词： Cancelable biometrics fuzzy extractors industrial wireless sensor network physical unclonable functions

来源：评论

学校读者我要写书评

暂无评论

Few-Shot Learning With Multi-Granularity Knowledge Fusion and Decision-Making

引用

IEEE TRANSACTIONS ON BIG data 2024年第4期10卷 486-497页

作者： Su, Yuling Zhao, Hong Zheng, Yifeng Wang, Yu Minnan Normal Univ Sch Comp Sci Zhangzhou 363000 Peoples R China Fujian Prov Univ Key Lab Data Sci & Intelligence Applicat Zhangzhou 363000 Peoples R China Tianjin Univ Coll Intelligence & Comp Tianjin 300072 Peoples R China

Few-shot learning (FSL) is a challenging task in classifying new classes from few labelled examples. Many existing models embed class structural knowledge as prior knowledge to enhance FSL against data scarcity. However, they fall short of connecting the class structural knowledge with the limited visual information which plays a decisive role in FSL model performance. In this paper, we propose a unified FSL framework with multi-granularity knowledge fusion and decision-making (MGKFD) to overcome the limitation. We aim to simultaneously explore the visual information and structural knowledge, working in a mutual way to enhance FSL. On the one hand, we strongly connect global and local visual information with multi-granularity class knowledge to explore intra-image and inter-class relationships, generating specific multi-granularity class representations with limited images. On the other hand, a weight fusion strategy is introduced to integrate multi-granularity knowledge and visual information to make the classification decision of FSL. It enables models to learn more effectively from limited labelled examples and allows generalization to new classes. Moreover, considering varying erroneous predictions, a hierarchical loss is established by structural knowledge to minimize the classification loss, where greater degree of misclassification is penalized more. Experimental results on three benchmark datasets show the advantages of MGKFD over several advanced models.

关键词： Visualization Feature extraction data models Decision making Convolutional neural networks Correlation Task analysis Decision-making few-shot learning knowledge fusion multi-granularity

来源：评论

学校读者我要写书评

暂无评论

CCSS: Towards conductance-based community search with size constraints

引用

EXPERT SYSTEMS WITH APPLICATIONS 2024年 250卷

作者： He, Yue Lin, Longlong Yuan, Pingpeng Li, Ronghua Jia, Tao Wang, Zeli Southwest Univ Coll Comp & Informat Sci Chongqing 400715 Peoples R China Huazhong Univ Sci & Technol Natl Engn Res Ctr Big Data Technol & Syst Serv Comp Technol & Syst Lab Sch Comp Sci & TechnolCluster & Grid Comp Lab Wuhan 430074 Peoples R China Beijing Inst Technol Sch Comp Sci Technol Beijing 100081 Peoples R China Chongqing Univ Posts & Telecommun Chongqing Key Lab Computat Intelligence Key Lab Big Data Intelligent Comp Key Lab Cyberspace Big Data Intelligent SecurMini Chongqing 400065 Peoples R China

Size -constrained community search, retrieving a size -bounded high -quality subgraph containing user -specified query vertices, has been extensively studied in graph analysis. However, existing methods mainly focus on the cohesiveness within the community while ignoring its separation from the outside, leading to sub -optimal results. Inspired by this, we adopt the well-known conductance to measure the quality of a community and introduce a novel problem of Conductance -based Community Search with Size Constraints ( CCSS ). CCSS aims to identify a subgraph with the smallest conductance among all connected subgraphs that contain the query vertex q and have at least l and at most h vertices. We prove that CCSS is NP -hard. To efficiently address CCSS , we first propose a computational framework ISC with two -level loops. In the inner loop, ISC starts from q and iteratively adds vertices based on the predefined score function, obtaining a coarse -grained candidate community. In the outer loop, we design the perturbation strategy to repeatedly execute the inner loop until the quality of the candidate community converges. Then, we propose two heuristic algorithms, ISCCI and ISCCP , to concretely implement the score function and perturbation strategy of ISC , which significantly reduce the computational costs. Finally, empirical results demonstrate the superiorities of our solutions. Our source codes and datasets are available at https://***/longlonglin/CCSS.

关键词： Community search Size-constrainted optimization problem Conductance

来源：评论

学校读者我要写书评

暂无评论

Security-Enhanced data Sharing Scheme With Location Privacy Preservation for Internet of Vehicles

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2024年第9期73卷 13751-13764页

作者： Hu, Peng Chu, Xixi Zuo, Kaizhong Ni, Tianjiao Xie, Dong Shen, Zhangyi Chen, Fulong Luo, Yonglong Anhui Normal Univ Sch Comp & Informat Wuhu 241002 Peoples R China Anhui Prov Key Lab Ind Intelligence Data Secur Wuhu 241002 Peoples R China

data sharing plays an important role in assisting traffic management centers to obtain traffic operation status and make accurate traffic decisions. However, there are still many security problems associated with data sharing, such as privacy breaches. To address these problems, we propose a security-enhanced data sharing scheme with location privacy preservation for Internet of Vehicles, called SEDS. Specifically, the scheme first exploits an autonomous and controllable dynamic pseudonym mechanism to protect the privacy of vehicles' location, thereby preventing the leakage of vehicle trajectory information. Secondly, SEDS utilizes homomorphic encryption to aggregate data at the edge of the network to achieve privacy protection of data content. Thirdly, SEDS uses batch verification with privacy protection to achieve integrity and authenticity of data, and can effectively filter false data. Finally, SEDS applies blockchain technology to store shared data, which can guarantee the immutability of shared data. The security proof and analysis show that SEDS has security features, containing confidentiality, location privacy, authentication and data integrity, filtering false data and immutability. Experimental results exhibit that the SEDS is superior to other recently proposed schemes in terms of computational and communication overhead, as well as network performance.

关键词： Internet of vehicles blockchain location privacy protection pseudonym mechanism Internet of vehicles blockchain location privacy protection pseudonym mechanism

来源：评论

学校读者我要写书评

暂无评论

Fingerprint image super-resolution based on multi-class deep dictionary learning and ridge prior

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2024年第6-7期18卷 5491-5501页

作者： Huang, Yi Bian, Weixin Xu, Deqin Jie, Biao Feng, Luo Anhui Normal Univ Sch Comp & Informat Wuhu 241002 Peoples R China Anhui Prov Key Lab Ind Intelligence Data Secur Wuhu 241002 Peoples R China

The identification of low-resolution fingerprints has always been one of the focuses in the field of biometric identification. This paper proposes a method for super-resolving low-resolution fingerprints based on deep dictionary learning. First, it is necessary to obtain a priori based on the fingerprint ridge orientation. After obtaining it, the ridges of the fingerprint are divided into n categories according to the direction. Each class uses deep dictionary learning models to train corresponding high- and low-resolution dictionaries respectively. In the super-resolution part, after extracting features from the patches that require super-resolution, the sparse coefficients are obtained through the deep dictionary learning model, and then combined with the high-resolution dictionary to obtain high-resolution patches, which are combined into high-resolution fingerprints. Experimental results show that the proposed method performs better than some other methods.

关键词： Deep dictionary learning Fingerprint super resolution Ridge orientation

来源：评论

学校读者我要写书评

暂无评论

HJE: Joint Convolutional Representation Learning for Knowledge Hypergraph completion

引用

IEEE TRANSACTIONS ON KNOWLEDGE AND data ENGINEERING 2024年第8期36卷 3879-3892页

作者： Li, Zhao Wang, Chenxu Wang, Xin Chen, Zirui Li, Jianxin Tianjin Univ Coll Intelligence & Comp Tianjin 300354 Peoples R China Tianjin Key Lab Cognit Comp & Applicat Tianjin 300354 Peoples R China Deakin Univ Sch Informat Technol Geelong Vic 3221 Australia

Knowledge hypergraph representation learning, which projects entities and $n$n-ary relations into a low-dimensional vector space, remains a challenging area to be explored despite the ubiquity of $n$n-ary relational facts in the real world. Current methods are always extensions of those used for knowledge graphs with shallow or deep structures. However, shallow and linear models limit the extraction capacity of the latent knowledge, while deep and non-linear models lead to the overabundance of parameters. In this paper, we propose a novel knowledge hypergraph completion model called HJE, which utilizes the powerful capability of convolutional neural networks for efficient representation learning. Interaction-enhanced 3D convolution and relation-aware 2D convolution are jointly utilized by HJE to extract explicit and implicit global knowledge and semantic information effectively without compromising the translation property of the model. Moreover, HJE constructs a unified learnable embedding matrix to capture entity position information in knowledge tuples. The entity mask mechanism can naturally couple the multilinear scoring approach for $n$n-ary facts to speed up the training convergence of the model. Extensive experimental results on real datasets of knowledge hypergraphs and knowledge graphs demonstrate the superior performance of HJE compared with state-of-the-art baselines.

关键词： Knowledge graphs Semantics Representation learning data mining Convolution Three-dimensional displays Tail Knowledge hypergraph knowledge representation learning convolutional neural network

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：