检索结果-内蒙古大学图书馆

2024 4th International Conference on Artificial Intelligence and Industrial Technology Applications, AIITA 2024

作者： Tao, Jiaye Hong, Chao Fu, Yun Yang, Yiwei Wei, Lipeng Liang, Zhihong Liu, Junrong Information Centre of Guizhou Power Grid Co Guizhou China Electric Power Research Institute China Southern Power Grid Guangdong Guangzhou China Guangdong Provincial Key Laboratory of Power System Network Security Guangdong Guangzhou China

The existing fuzz testing methods for industrial control protocols suffer from insufficient coverage, false positives, and an inability to handle protocol semantics. This paper proposes a reinforcement learning-based seed scheduling coverage-guided fuzz testing method. Building upon coverage-guided fuzz testing techniques, we integrate reinforcement learning with seed scheduling to optimize the seed selection strategy, thereby enhancing the efficiency of protocol vulnerability detection. Experimental results demonstrate the feasibility and effectiveness of this approach. Through reinforcement learning guidance, seed scheduling is optimized, thereby strengthening the performance of fuzz testing in exploring vulnerabilities in industrial control protocols. © Published under licence by IOP Publishing Ltd.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Method of Computer Automatic Scoring for Subjective Questions 2

A Method of Computer Automatic Scoring for Subjective Questi...

引用

2nd International Conference on Artificial Intelligence, Human-Computer Interaction and Robotics, AIHCIR 2023

作者： Jiang, Yan Liu, Hua Li, Jinping Xia, Yingjie School of Information Science and Engineering China University of Jinan Shandong Provincial Key Laboratory of Network Based Intelligent Computing Jinan250022 China

ISBN: (纸本)9798350360363

In response to the limitations of question types and difficulties of automatic scoring, a method of scoring subjective questions automatically in the computer exams is proposed, and then it is implemented. In general, the meth-ods of solving a certain subjective question are only a few, and their ideas and processes are relatively common and fixed. Our method is to evaluate the idea of solving subjective questions, not to evaluate the process of solving subjective questions, that is, to evaluate the application orders or relationships of relevant knowledge points or operations, including the definitions, formulas, theorems, inferences, lemmas, laws, etc., and their application orders using in the solution process of calculation or proof questions. They have already been numbered when setting the subjective questions. Their contents are displayed according to human habits on the system interface, and their numbers are used during the process of the storages and scoring in the com-puters. This method can change the current form of the computer examination and scoring. It can achieve high effi-ciency with low cost, reducing manual labor intensity. In addition, the scoring of subjective questions will become fairer and more transparent. © 2023 IEEE.

关键词： Automatic Scoring Computer Exams Subjective Questions

来源：评论

学校读者我要写书评

暂无评论

A numerical magnitude aware multi-channel hierarchical encoding network for math word problem solving

引用

Neural computing and Applications 2025年第3期37卷 1651-1672页

作者： Xu, Junjie Chen, Yuzhong Xiao, Lingsheng Liao, Hongmiao Zhong, Jiayuan Dong, Chen College of Computer and Data Science Fuzhou University Fujian Province Fuzhou350108 China Engineering Research Center of Big Data Intelligence Ministry of Education Fuzhou China Fujian Provincial Key Laboratory of Network Computing and Intelligent Information Processing Fujian Province Fuzhou350108 China

Math word problem (MWP) represents a critical research area within reading comprehension, where accurate comprehension of math problem text is crucial for generating math expressions. However, current approaches still grapple with unresolved challenges in grasping the sensitivity of math problem text and delineating distinct roles across various clause types, and enhancing numerical representation. To address these challenges, this paper proposes a Numerical Magnitude Aware Multi-Channel Hierarchical Encoding network (NMA-MHEA) for math expression generation. Firstly, NMA-MHEA implements a multi-channel hierarchical context encoding module to learn context representations at three different channels: intra-clause channel, inter-clause channel, and context-question interaction channel. NMA-MHEA constructs hierarchical constituent-dependency graphs for different levels of sentences and employs a Hierarchical Graph Attention Neural network (HGAT) to learn syntactic and semantic information within these graphs at the intra-clause and inter-clause channels. NMA-MHEA then refines context clauses using question information at the context-question interaction channel. Secondly, NMA-MHEA designs a number encoding module to enhance the relative magnitude information among numerical values and type information of numerical values. Experimental results on two public benchmark datasets demonstrate that NMA-MHEA outperforms other state-of-the-art models. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Lightweight Malicious Encrypted Traffic Detection Method Based on KL Image Compression 9

Lightweight Malicious Encrypted Traffic Detection Method Bas...

引用

9th IEEE International Conference on Data Science in Cyberspace, DSC 2024

作者： Wen, Yuecheng Li, Xiaotong Lv, Haiqing Wang, Menglu Zuo, Wenbo Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9798350391367

Timely and accurate detection of and response to potentially malicious encrypted traffic is crucial in cybersecurity strategies. Convolutional neural network-based approaches that convert traffic loads into images for analysis have made significant progress in malicious encrypted traffic detection, but they suffer from the challenges of large number of participants, large network size, and high demand for computational resources, especially when edge computing devices are limited. Therefore, there is an urgent need for a lightweight, efficient and accurate malicious encrypted traffic detection model. This paper proposes a malicious encrypted traffic detection model whose core idea is to utilize KL image compression technique. The model first converts the raw traffic data into grayscale images and applies KL transform to compress the images efficiently. Subsequently, a convolutional neural network model with lightweight design is used to classify the compressed image for fast and accurate malicious encrypted traffic detection. After validating on both the public traffic dataset and the lab's self-collected dataset, the results indicate that our method requires just 0.0216 seconds to process a single input sample, achieving an average classification accuracy of 98%. These findings highlight its practicality and excellence in addressing the real-world challenges of malicious encrypted traffic. © 2024 IEEE.

关键词： Cybersecurity

来源：评论

学校读者我要写书评

暂无评论

A Patent Fast Recommendation Algorithm Based on User Requirements Combined with an Inverted Table

A Patent Fast Recommendation Algorithm Based on User Require...

引用

2024 lEEE International Conference on Advanced information, Mechanical Engineering, Robotics and Automation, AIMERA 2024

作者： Zheng, Wan Ding, Qingyan Li, Na Pan, Yu Dong, Xuecheng Qilu University of Technology Shandong Academy of Sciences Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Jinan China Shandong Fundamental Research Center for Computer Science Shandong Provincial Key Laboratory of Computer Networks Jinan China

ISBN: (纸本)9798350343335

Patent recommendation has always been an essential service for the innovation and development of all walks of life. The matching of demand and patent information can enable enterprises to quickly obtain the innovation and development trend of the industry. The current patent recommendation method has the problem of slow recommendation speed and insufficient comprehensive recommendation results. According to the existing problems, a patent fast recommendation algorithm based on user needs combined with inverted table is proposed. The algorithm using dual track recommended way for accurate recommendation and extensive recommendation, in the process of recommendation built by patent data inversion table and document list auxiliary recommendation to improve recommendation speed and accuracy, when accurate recommendation when using improved requirements based on GPT-2 compression model recommended after demand text compression, when widely recommended using TF-IDF mechanism for demand text of each paragraph after keyword extraction, finally generate widely recommendation list and accurate recommendation list. This recommendation method can not only realize the fast and accurate recommendation of patents, but also provide users with more innovative ideas for extensive recommendation. The experiments show that the proposed recommendation algorithm has higher speed and accuracy than the original recommendation algorithm. © 2024 IEEE.

关键词： Patents and inventions

来源：评论

学校读者我要写书评

暂无评论

A novel firmware secure transmission scheme based on SM2 threshold key agreement algorithm

A novel firmware secure transmission scheme based on SM2 thr...

引用

2023 International Conference on Computer Graphics, Artificial Intelligence, and Data Processing, ICCAID 2023

作者： Wang, Kun Gao, Shaofeng Chen, Xinze Fu, Yong Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Qilu University of Technology Shandong Academy of Sciences Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (纸本)9781510677685

During the process of embedded firmware upgrade, to protect the security of the firmware during transmission, this paper proposes an improved firmware transmission scheme based on the combination of the SM2 threshold key agreement algorithm. Between the server, client, and downloader, the SM2 key agreement algorithm is used to generate session key pairs to encrypt the firmware information, and the method of reinforcing and verifying re-transmission is adopted to ensure data security. In order to improve the security of the algorithm, a threshold algorithm is used in the key agreement process. The server acts as a trusted center to execute the threshold algorithm and divide the main secret private keys dA and dB. During the phase of key exchange, at least t+1 sub-secrets are required to recover the complete main secret. This paper successfully implements and verifies the feasibility of this scheme. © 2024 SPIE.

关键词： Firmware

来源：评论

学校读者我要写书评

暂无评论

Distributed acoustic sensing signal event recognition and localization based on improved YOLOv7

Distributed acoustic sensing signal event recognition and lo...

引用

2023 International Congress on Ultrasonics, ICU Beijing 2023

作者： Zhang, Hongyu Wang, Chang Zhang, Faxiang Jiang, Shaodong Sun, Zhihui Wang, Xiaodong Duan, Zhenhui Gao, Fengxia Liu, Zhaoying Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Qilu University of Technology Shandong Academy of Sciences Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

The distributed acoustic sensing (DAS) system based on phase-sensitive optical time domain reflection (Φ -OTDR) technology is widely used in pipeline safety monitoring, perimeter security, structure monitoring, etc. Accurate localization and recognition of multiscene events over long distances has always been a challenge. This paper proposes an improved YOLOv7 algorithm for multi-event real-time detection of DAS system. The algorithm employs space-to-depth Conv(SPD-Conv) to replace the strided convolutions and pooling operations in YOLOv7, reducing fine-grained information loss and learning of inefficient feature representations. In addition, the Convolutional Block Attention Module (CBAM) is introduced in YOLOv7 to improve the model performance. We collected spatial-temporal signal data for six types of pipeline safety events, and passed them into the improved YOLOv7 algorithm in the form of data matrixes for training and *** have shown that the proposed method achieves an mAP@.5 (mean Average Precision) of 99.7% for the identification of six pipeline safety event types. Positioning loss reduced to 0.2%,and detection speed can reach 70 frames per second(FPS). Our scheme achieves significant improvements in localization and classification accuracy compared to Faster R-CNN, etc. The event recognition localization method proposed in this paper has the advantage of fast speed and high accuracy. © 2024 Institute of Physics Publishing. All rights reserved.

关键词： Light reflection

来源：评论

学校读者我要写书评

暂无评论

Enhancing Stance Detection on Social Media via Core Views Discovery 27

Enhancing Stance Detection on Social Media via Core Views Di...

引用

27th European Conference on Artificial Intelligence, ECAI 2024

作者： Yan, Yu Shen, Yinghan Liu, Teli Jiang, Xuhui Yin, Dechun School of Information and Network Security People Public Security University of China China Institute of Computing Technology Chinese Academy of Sciences China Key Laboratory of AI Safety Institute of Computing Technology Chinese Academy of Sciences China

ISBN: (纸本)9781643685489

Stance detection aims to identify the expressed attitude towards a target from the text, which is significant for learning public cognition from social media. The short and implicit nature of social media users' expressions potentially results in the stance understanding bias of the model. To address this problem, introducing external background information is helpful to mitigate these biases and enhance explainability. The core view, reflecting the motivations and reasons behind an individual's stance toward the target, can be summarized and extracted from collective tweets, which can serve as a reference for stance detection. In this study, we propose the Stance Detection via Core View Discovery (SD-CVM), where the core views are used for background information modeling. Specifically, we construct a joint classifier combining the semantic understanding of tweets and their relevant core views from the public. We utilize the Large Language Model (LLM) to extract core views with stances from tweets and use these core views as background references for tweets. To further optimize the tweet understanding, we develop the contrastive and rebalancing mechanism by incorporating stance supervision signals for training. Experiments on two representative datasets demonstrate the excellent performance of our method. © 2024 The Authors.

关键词： Tweets

来源：评论

学校读者我要写书评

暂无评论

Patent Recommendation Methods Based on Transformer Encoders and Regularization Strategies

Patent Recommendation Methods Based on Transformer Encoders ...

引用

2024 lEEE International Conference on Advanced information, Mechanical Engineering, Robotics and Automation, AIMERA 2024

ISBN: (纸本)9798350343335

Patent recommendation service has become a major demand of all walks of life in the society. Patent data collection and analysis can not only help enterprises grasp the development status of the industry, but also provide decision support for them in the research and development process of science and technology services. In view of the problem of inaccurate patent recommendation caused by the inconsistency between the search text and the patent text length and the low accuracy of traditional patent recommendation, a word and sentence dual processing model based on Transformer encoder and regularization strategy is proposed to complete the patent recommendation task from the level of sentence granularity and word granularity. The sentence granularity feature extraction module uses the transformer encoder superimposed multi-head attention method to obtain the feature information between the sentence vector groups, so that the sentence vector fuses the feature information of each sentence, and solves the problem of low recommendation accuracy caused by the inconsistent length of the patent text and the search text;The word granularity feature extraction module uses the twin network model to extract the feature information of the sentence granularity vector using the Transformer encoder using the double splicing operation, improve the model performance and model accuracy;Introduce a strategy based on Dropout and symmetric JS divergence regularization, improve the model generalization ability and accuracy. Experiments have proved that it outperforms SBERT, BERT, MGMSN and SVM models in terms of performance. © 2024 IEEE.

关键词： Patents and inventions

来源：评论

学校读者我要写书评

暂无评论

An adaptive time-convolutional network online prediction method for ocean observation data 36

An adaptive time-convolutional network online prediction met...

引用

36th International Conference on Software Engineering and Knowledge Engineering, SEKE 2024

作者： Li, Enjing Li, Xiang Zhao, Zhigang Wu, Lu Wang, Yinglong Key Laboratory of Computing Power Network and Information Security Ministry of Education Shandong Computer Science Center National Supercomputer Center in Jinan Qilu University of Technology Shandong Academy of Sciences Jinan China Shandong Provincial Key Laboratory of Computer Networks Shandong Fundamental Research Center for Computer Science Jinan China

ISBN: (纸本)1891706594

Deep learning is particularly important in the field of time series data analysis, and has been applied to tasks such as marine data prediction. However, there is a’concept drift’ problem in marine observation data, which leads to performance degradation and catastrophic forgetting of traditional deep learning models used in online scenarios. For this reason, this paper proposes OL-TCN, an adaptive temporal convolutional neural network online prediction deep learning model, which is more suitable for marine online learning and inference scenarios. We make the following innovations: Inside the model, model performance is enhanced by incorporating a multi-head attention mechanism and introducing an automatic machine learning enhancement in the residual cell. Outside the model, a model repository approach is used to effectively cope with the complexity and evolutionary challenges of data streams. The experimental results verify that the OL-TCN model is effective and feasible in marine time series processing. © 2024 Knowledge Systems Institute Graduate School. All rights reserved.

关键词： Spatio-temporal data

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：