检索结果-内蒙古大学图书馆

SSRN 2024年

作者： Li, Yahui Sun, Qi Zhou, Chunjie Liu, Lu Tian, Yu-Chu School of Artificial Intelligence and Automation Huazhong University of Science and Technology Hubei Wuhan430074 China Key Laboratory of Image Processing and Intelligent Control Ministry of Education Huazhong University of Science and Technology Hubei Wuhan430074 China School of Cyber Science and Engineering Huazhong University of Science and Technology Hubei Wuhan430074 China School of Computer Science Queensland University of Technology BrisbaneQLD4001 Australia

Intelligent industrial manufacturing heavily relies on structured knowledge. Named Entity Recognition (NER), an essential technique for extracting structured knowledge from text, has garnered significant research interest. However, current NER approaches face difficulties in handling multiple levels of entity nesting in knowledge extraction, especially within the manufacturing context. To address this issue, we present an Anchor Span-based NER (ASNER) approach for accurately and efficiently extracting manufacturing text knowledge. This method combines token and span classification. Initially, the head and tail features of the entity are extracted, with the corresponding anchor span of potential entities generated based on a boundary match neural network to localize the entities. Subsequently, token spatial contextual features are extracted using biaffine attention and convolutional neural networks, and classification categories are assigned to the anchor span-filtered entity features. Experimental studies demonstrate the effectiveness of the proposed ASNER approach. © 2024, The Authors. All rights reserved.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

AoI-Reliability Analysis with Block Length over Erasure Channels Using Rateless Codes

AoI-Reliability Analysis with Block Length over Erasure Chan...

引用

International Conference on Communication Technology (ICCT)

作者： Haobo Huang Yusun Fu Yue Qiao Junpeng Yin Weiwu Yan Ningbo Artificial Intelligence Institute Shanghai Jiao Tong University Ningbo China School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

As a metric to measure the information freshness, Age of Information (AoI) is widely used to evaluate low-latency high-reliability communication. Since the communication in Industrial Internet of Things (IIoT) often has the special requirements, the relationship among AoI, reliability and block length in time-critical scenarios with bounded latency and non-time-critical scenarios without bounded latency is explored respectively in this paper, based on erasure channels using rateless codes. We build a model using LT codes over erasure channels and consider two metrics: average AoI and bounded AoI. Theoretical analysis and simulation results show that AoI has a minimum value with different block lengths. Furthermore, we draw a conclusion that the minimum value of AoI under different reliability is monotonically decreasing functioned with reliability. Finally, the impact of overhead on reliability and AoI of LT codes in time-critical scenarios is studied.

关键词： Measurement Codes Simulation Reliability theory Information age Communications technology Time factors Low latency communication Industrial Internet of Things

来源：评论

学校读者我要写书评

暂无评论

Addr-Net: Attention Mechanism-Based Dual-Stream Deformable Medical image Registration Network

SSRN

引用

SSRN 2024年

作者： Fan, Chao Zhu, Xinru Peng, Bincheng Xuan, Zhihui Zhu, Zhentong School of Artificial Intelligence and Big Data Henan University of Technology Henan Province Zhengzhou City China Key Laboratory of Grain Information Processing and Control Ministry of Education Henan Province Zhengzhou City China School of Information Science and Engineering Henan University of Technology Henan Province Zhengzhou City450001 China

Medical image registration is crucial for tumor growth monitoring, radiation therapy, and disease diagnosis. Recently, U-Net type networks have become widely used in unsupervised image registration to predict dense deformation fields. However, for image data with large deformations, these networks are prone to interference from outliers and task-irrelevant regions, leading to suboptimal registration accuracy and difficulty in maintaining the continuity of the deformation field. To address this issue, we propose a novel model called ADDR-Net. ADDR-Net leverages dual encoders and incorporates a Cross-Perceptual Attention Module (CPA) alongside a Multi-dimensional Attention Module (MDA) to enhance registration. Unlike traditional methods that concatenate fixed and moving images into a single encoder, ADDR-Net uses dual encoders to separately extract features from each image, facilitating explicit feature matching between image pairs. The CPA module, embedded within the encoder, enhances the extraction of global features over long distances. Meanwhile, the MDA module, embedded within the skip connection, filters redundant features and extracts local refined features, thereby improving the U-shaped network architecture. Experimental results demonstrate that ADDR-Net outperforms state-of-the-art methods. Specifically, compared to the Transformer-based TransMorph method, ADDR-Net increases the Dice scores on two publicly available brain datasets, IXI and OASIS, by 1.4% and 3.5%, respectively. © 2024, The Authors. All rights reserved.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Enhancing the McEliece Scheme Based on Concatenation of Polar Codes and Blocked QC-LDPC Codes

Enhancing the McEliece Scheme Based on Concatenation of Pola...

引用

IEEE Conference on Communications and Network Security (CNS)

作者： Xin Lin Yusun Fu Zihao Wang Junpeng Yin Ningbo Artificial Intelligence Institute Shanghai Jiao Tong University Ningbo China School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350375961

ISBN: (纸本)9798350375978

This paper proposes an enhanced McEliece cryptosystem, leveraging a concatenation scheme of Polar codes and blocked QC-LDPC codes. During the key generation phase, the generator matrices for both Polar and blocked QC-LDPC codes are employed as private keys. These matrices are subsequently multiplied by reversible and random permutation matrices respectively. The products of these multiplications are then combined to form the public key, effectively concealing the structural properties of the generator matrices. In the encryption phase, the plaintext is encrypted sequentially through the Polar codes and the blocked QC-LDPC codes. This process introduces two layers of random error vectors, culminating in the generation of the encrypted ciphertext. The proposed scheme reduces the public key length by 54% and doubles the security level compared to the original McEliece scheme with similar parameters. It also significantly reduces the total key length compared to improved schemes based on Polar codes and QC-LDPC/MDPC codes at an 80-bit security level, and shortens the private key length compared to schemes based on Polar-LDPC concatenated codes. Notably, the scheme is secure against known structural and decoding attacks while increases resistance to information set decoding (ISD) attacks. Through the randomization and hashing of the plaintext, the proposed scheme can achieve IND-CCA2 security.

关键词： Resistance Codes Public key Network security Generators Vectors Decoding Encryption Concatenated codes Polar codes

来源：评论

学校读者我要写书评

暂无评论

Boltzmann Robust Soliton Distribution for Rateless Codes

Boltzmann Robust Soliton Distribution for Rateless Codes

引用

International Conference on Communication Technology (ICCT)

作者： Jinhui Tang Yusun Fu Yue Qiao Junpeng Yin Haobo Huang School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University Shanghai China Ningbo Artificial Intelligence Institute Shanghai Jiao Tong University Ningbo China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China Shanghai Engineering Research Center of Intelligent Control and Management Shanghai Jiao Tong University Shanghai China

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

The decoding success rate may be worse when classical degree distributions are used in short-packet transmission of rateless codes, especially in industrial communication scenarios. A novel Boltzmann degree distribution is proposed for Luby Transform (LT) codes. The Boltzmann degree distribution has a high decoding success rate for short-packet transmission but a lower decoding success rate in long-packet transmission. To balance the performance of rateless codes for short and long packet transmission, the Boltzmann degree distribution and the robust soliton distribution (RSD) are combined with a new joint degree distribution, called Boltzmann Robust Soliton Distribution (BRSD). Simulation results show that compared with RSD and Poisson robust soliton distribution (PRSD), the proposed joint distribution can achieve a higher decoding success rate and reduce the average degree, especially in short-packet transmission of industrial communication.

关键词： Codes Simulation Industrial communication Symbols Solitons Transforms Robustness Decoding Complexity theory Physics

来源：评论

学校读者我要写书评

暂无评论

Causal Invariant Hierarchical Molecular Representation for Out-of-distribution Molecular Property Prediction

Causal Invariant Hierarchical Molecular Representation for O...

引用

IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

作者： Xinlong Wen Yifei Guo Shuoying Wei Wenhan Long Lida Zhu Rongbo Zhu Huazhong Agricultural University Wuhan China College of Informatics Huazhong Agricultural University Wuhan China Shenzhen Institute of Nutrition and Health Huazhong Agricultural University Shenzhen China Shenzhen Branch Guangdong Laboratory for Lingnan Modern Agriculture Genome Analysis Laboratory of the Ministry of Agriculture and Agricultural Genomics Institute at Shenzhen Chinese Academy of Agricultural Sciences Shenzhen China Engineering Research Center of Intelligent Technology for Agriculture Ministry of Education Wuhan China Huazhong University of Science and Technology Wuhan China Key Laboratory of Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China

ISBN: (数字)9798350386226

ISBN: (纸本)9798350386233

Molecular representation learning is widely used in the field of drug discovery, due to its ability to accurately capture the complex features of compounds in high-dimensional space. However, existing molecular representation learning models are prone to be influenced by spurious parts during distribution shifts (also known as out-of-distribution, or OOD), which results in models mistakenly treating these spurious parts as crucial features of molecules, thereby limiting the generalization capability of the models. To tackle this issue, a novel invariant molecular representation learning model, called Causal Invariant Hierarchical Molecular Representation Graph Neural Networks (CHiMoGNN), is proposed for OOD molecular property prediction. In CHiMoGNN, a Feature Enhancement (FE) module is designed to leverage the multi-level molecular parts to enhance the expression of invariant features, thereby enhancing the model’s capability to capture key molecular information. In addition, a Cartesian Product based Environmental Impact (EI) module is adopted to generate counterfactual samples with environmental diversity. Consequently, these samples are utilized to train a classifier that maintains consistent performance across various environments. Extensive experiments on seven real-world datasets demonstrate that CHiMoGNN outperforms 9 state-of-the-art models, achieving a 5.73% increase in average ROC-AUC, and the results also show that CHiMoGNN can effectively maintain generalization in various distribution shifts. Code and datasets are available at https://***/Chertuion/CHiMoGNN.

关键词： Representation learning Limiting Codes Predictive models Feature extraction Iron Graph neural networks Drug discovery Data mining Compounds

来源：评论

学校读者我要写书评

暂无评论

ESNet: evolution and succession network for high-resolution salient object detection 24

ESNet: evolution and succession network for high-resolution ...

引用

Proceedings of the 41st International Conference on Machine Learning

作者： Hongyu Liu Runmin Cong Hua Li Qianqian Xu Qingming Huang Wei Zhang Institute of Information Science Beijing Jiaotong University & Beijing Key Laboratory of Advanced Information Science and Network Technology Beijing China School of Control Science and Engineering Shandong University & Key Laboratory of Machine Intelligence and System Control Ministry of Education Jinan China School of Computer Science and Technology Hainan University Hainan China Key Laboratory of Intelligent Information Processing Institute of Computing Technology CAS Beijing China School of Computer Science and Technology University of Chinese Academy of Sciences Beijing China

Preserving details and avoiding high computational costs are the two main challenges for the High-Resolution Salient Object Detection (HRSOD) task. In this paper, we propose a two-stage HRSOD model from the perspective of evolution and succession, including an evolution stage with Low-resolution Location Model (LrLM) and a succession stage with High-resolution Refinement Model (HrRM). The evolution stage achieves detail-preserving salient objects localization on the low-resolution image through the evolution mechanisms on supervision and feature; the succession stage utilizes the shallow high-resolution features to complement and enhance the features inherited from the first stage in a lightweight manner and generate the final high-resolution saliency prediction. Besides, a new metric named Boundary-Detail-aware Mean Absolute Error (MAEBD) is designed to evaluate the ability to detect details in high-resolution scenes. Extensive experiments on five datasets demonstrate that our network achieves superior performance at real-time speed (49 FPS) compared to state-of-the-art methods. Our code is publicly available at: https://***/rmcong/ESNet_ICML24.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Improved black-box attack based on query and perturbation distribution

Improved black-box attack based on query and perturbation di...

引用

International Workshop on Advanced Computational intelligence (IWACI)

作者： Weiwei Zhao Zhigang Zeng School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan China Key Laboratory of Image Processing and Intelligent Control of Education Ministry of China Wuhan China

Adversarial examples cause the deep neural network prediction error, which is a great threat to the deep neural network. How to generate more natural adversarial examples and improve the robustness of deep neural networks has received attention. In this paper, we propose an improved blackbox attack (IBBA) algorithm based on query and perturbation distribution. This algorithm only needs the top-l label of the attacked model to generate the adversarial examples. Based on the existing black-box attacks, we optimize the performance of the algorithm from two aspects: query distribution and perturbation distribution. In the aspect of query distribution, we adopt different strategies for nontargeted attack and targeted attack; in the aspect of perturbation distribution, we choose different low-frequency noise according to the difference between the targeted attack and nontargeted attack. The experimental results on imageNet show that the proposed algorithm is better than the existing algorithms in low query number, and the targeted attack is better in each specified query number.

关键词： Perturbation methods Neural networks Prediction algorithms Robustness Low-frequency noise Discrete cosine transforms Computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Modality-aware triplet hard mining for zero-shot sketch-based image retrieval

arXiv

引用

arXiv 2021年

作者： Huang, Zongheng Sun, Yifan Han, Chuchu Gao, Changxin Sang, Nong Key Laboratory of Ministry of Education for Image Processing and Intelligent Control School of Artificial Intelligence and Automation Huazhong University of Science and Technology China Baidu Research United States

This paper tackles the Zero-Shot Sketch-Based image Retrieval (ZS-SBIR) problem from the viewpoint of cross-modality metric learning. This task has two characteristics: 1) the zero-shot setting requires a metric space with good within-class compactness and the between-class discrepancy for recognizing the novel classes and 2) the sketch query and the photo gallery are in different modalities. The metric learning viewpoint benefits ZS-SBIR from two aspects. First, it facilitates improvement through recent good practices in deep metric learning (DML). By combining two fundamental learning approaches in DML, e.g., classification training and pairwise training, we set up a strong baseline for ZS-SBIR. Without bells and whistles, this baseline achieves competitive retrieval accuracy. Second, it provides an insight that properly suppressing the modality gap is critical. To this end, we design a novel method named Modality-Aware Triplet Hard Mining (MATHM). MATHM enhances the baseline with three types of pairwise learning, e.g., a cross-modality sample pair, a within-modality sample pair, and their combination. We also design an adaptive weighting method to balance these three components during training dynamically. Experimental results confirm that MATHM brings another round of significant improvement based on the strong baseline and sets up new state-of-the-art performance. For example, on the TU-Berlin dataset, we achieve 47.88+2.94% mAP@all and 58.28+2.34% Prec@100. Code will be publicly available. Copyright © 2021, The Authors. All rights reserved.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

Depth-Aware Multi-Modal Fusion for Generalized Zero-Shot Learning 22

Depth-Aware Multi-Modal Fusion for Generalized Zero-Shot Lea...

引用

22nd IEEE International Conference on Industrial Informatics, INDIN 2024

作者： Cao, Weipeng Yao, Xuyang Xu, Zhiwu Pan, Yinghui Sun, Yixuan Li, Dachuan Qiu, Bohua Wei, Muheng Shenzhen China National Engineering Laboratory for Big Data System Computing Technology Shenzhen University Shenzhen China College of Computer Science and Software Engineering Shenzhen University Shenzhen China Stony Brook University New York United States Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology Shenzhen China Department of Computer Science and Engineering Southern University of Science and Technology Shenzhen China ZhenDui Industry Artificial Intelligence Co. Ltd Shenzhen China Department of Automation Shanghai Jiao Tong University Shanghai China Key Laboratory of System Control and Information Processing Ministry of Education of China Shanghai China

ISBN: (纸本)9798331527471

Realizing Generalized Zero-Shot Learning (GZSL) based on large models is emerging as a prevailing trend. However, most existing methods merely regard large models as black boxes, solely leveraging the features output by the final layer while disregarding potential performance enhancements from other layers. Indeed, numerous researchers have visually depicted variations in the features learned across different layers of neural networks. Motivated by this observation, we propose a Vision Transformer (ViT)-based GZSL method named Depth-Aware Multi-Modal ViT (DAM2ViT), which exploits multi-level features of ViT. DAM2ViT incorporates a multi-modal interaction block to align semantic information of categories across multiple layers, thereby augmenting the model’s capacity to learn associations between visual and semantic spaces. Extensive experiments conducted on three benchmark datasets (i.e., CUB, SUN, AWA2) have showcased that DAM2ViT achieves competitive results compared to state-of-the-art methods. © 2024 IEEE.

关键词： Zero-shot learning

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：