检索结果-内蒙古大学图书馆

5th International Academic Exchange Conference on Science and Technology Innovation, IAECST 2023

作者： Chen, Yuhang Yang, di Wang, Peng Gao, Cheng Changchun University of Science and Technology School of Computer Science and Technology Changchun China Changchun University of Science and Technology College of Computer Science and Technology Changchun China Chengdu Sub-Center of Software R&D Center Postal Savings Bank of China Chengdu China

ISBN: (纸本)9798350357738

Addressing the issues of scarce corpora, semantic sparsity, and polysemy in the field of tender announcements, this paper proposes a bid announcement domain named entity recognition model based on a two-stage enhancement. The model consists of two core stages: context-aware fine-tuning (CFT) and multi-dimensional context semantic synthesis (MdCS). In the CFT stage, a domain-specific dataset is constructed, and ErNIE model is fine-tuned using a multi-stage masking strategy to acquire context-specific information in the tender announcement domain. Subsequently, in the MdCS stage, features are dynamically fused using a bidirectional long short-term memory network (BiLSTM) and a multi-head self-attention mechanism to capture the dependency relationships between complex entities in long texts. Experimental results demonstrate that our proposed model outperforms existing state-of-the-art models in named entity recognition tasks within the tender announcement domain. Additionally, ablation experiments further confirm the indispensability of each component in the two-stage strategy. © 2023 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

design and Implementation of Packet-Aware Scheduling Algorithm for Serverless Architecture

Design and Implementation of Packet-Aware Scheduling Algorit...

引用

International Conference on Artificial Intelligence and Computer Information Technology (AICIT)

作者： Xingliang Chen Yuan Gao Yuheng ding Peng Fu System Platform R&D Center NARI Group Corporation State Grid Electric Power Research Institute Nanjing China

ISBN: (数字)9798350365443

ISBN: (纸本)9798350365450

Serverless is a fine-grained deployment model. Users only need to pay for the time segment of specific function execution to achieve true on-demand charging. However, the cloud function startup process requires downloading dependent packages, which delays the response time of the function and reduces the application's ability to respond quickly to load bursts. Therefore, a packet aware scheduling algorithm is proposed for the cloud function cold start problem. The algorithm establishes a scheduling queue and a packet cache tree at the working node. When the server receives the cloud function, the scheduler senses the optimal matching position of the packet through the improved location sensitive hash algorithm (WCFCM-LSH), and schedules the cloud function to the node queue; At the same time, in order to avoid excessive load on the work node, the scheduler will steal the tasks in the highest load scheduling queue and schedule them to other queues to reduce the load of the node. The experiment shows that the packet aware scheduling algorithm can effectively reduce the cold start time of cloud function, improve the scheduling efficiency, and reduce the waste of platform resources.

关键词： Schedules Scheduling algorithms Clustering algorithms Computer architecture Scheduling delays Time factors Servers resource management Optimal matching

来源：评论

学校读者我要写书评

暂无评论

rectal Cancer Stages T2 and T3 Identification Based on Asymptotic Hybrid Feature Maps

引用

Computer Modeling in Engineering & Sciences 2023年第10期137卷 923-938页

作者： Shujing Sun Jiale Wu Jian Yao Yang Cheng Xin Zhang Zhihua Lu Pengjiang Qian R&D Center of Artificial Intelligence Systems and Applications School of Artificial Intelligence and Computer ScienceJiangnan UniversityWuxi214122China Shanghai Enterprise Information Operation Center China Telecom Group Co.Ltd.Shanghai201315China Department of Radiology Dushu Lake Hospital Affiliated to Soochow UniversitySuzhou215123China Suzhou Yunmai Software Technology Co. Ltd.Suzhou215100China

Many existing intelligent recognition technologies require huge datasets for model ***,it is not easy to collect rectal cancer images,so the performance is usually low with limited training *** addition,traditional rectal cancer staging is time-consuming,error-prone,and susceptible to physicians’subjective awareness as well as professional *** settle these deficiencies,we propose a novel deep-learning model to classify the rectal cancer stages of T2 and ***,a novel deep learning model(rectalNet)is constructed based on residual learning,which combines the squeeze-excitation with the asymptotic output layer and new cross-convolution layer links in the residual block ***,a two-stage data augmentation is designed to increase the number of images and reduce deep learning’s dependence on the volume of *** experiment results demonstrate that the proposed method is superior to many existing ones,with an overall accuracy of ***,other traditional techniques,such as VGG16,denseNet121,EL,and dErNet,have an average accuracy of 0.6981,0.7032,0.7500,and 0.7685,respectively.

关键词： rectal cancer staging T2WI images data expansion asymptotic output layer

来源：评论

学校读者我要写书评

暂无评论

FTM Implementation and ranging Performance Characterization for Indoor Localization 13

FTM Implementation and Ranging Performance Characterization ...

引用

13th IEEE Global Conference on Consumer Electronic, GCCE 2024

作者： Hirata, rei Koizumi, Atsushi Oguma, Hiroshi Sakaguchi, Naotaka Sumi, Takenori Nagai, Yukimasa Nit Toyama College Department of Control Information System Enginering 1-2 Ebie-Neriya Toyama Imizu933-0239 Japan Mitsubishi Electric Corporation Information Technology R&d Center 5-1-1 Ofuna Kanagawa Kamakura247-8501 Japan

ISBN: (纸本)9798350355079

Indoor positioning technologies for IoT sensing are attracting attention. We are currently exploring standard indoor positioning methods using IEEE 802.11 / Wi-Fi. In this paper, we utilize the IEEE802.11-2016 FTM (Fine Timing Measurement) protocol, which enables to calculate distances between devices and access points (APs). The ranging accuracy of FTM, however, varies depending on multipath propagation of wireless signals and device-specific factors. Therefore, we implemented FTM on two models, Google Pixel4 and Zenfone9, and conducted experiment studies in the gymnasium and corridor of our school to evaluate position detection accuracy. In the gymnasium, the average error was 1.42 m for Zenfone9, while 1.07 m for Pixel4. In the corridor, the average error was 0.66m for Zenfone9, whereas 2.47 m for Pixel4. © 2024 IEEE.

关键词： FTM Indoor Localization IoT sensing machine dependence round Trip Time

来源：评论

学校读者我要写书评

暂无评论

Optimal design of WET Etching Bath for 3d Flash Memories Using Multi-Objective Bayesian Optimization

Optimal Design of WET Etching Bath for 3D Flash Memories Usi...

引用

Joint e-Manufacturing and design Collaboration Symposium (eMdC) & International Symposium on Semiconductor Manufacturing (ISSM)

作者： Miyuki Kouda Yumi Mori Tomohiko Sugita Youyang Ng AI & System Research Center Frontier Technology R&D Institute KIOXIA Corporation Yokkaichi-shi Japan AI & System Research Center Frontier Technology R&D Institute KIOXIA Corporation Yokohama-shi Japan Advanced Memory Process Development Center Memory Division KIOXIA Corporation Yokkaichi-shi Japan

ISBN: (数字)9798331510589

ISBN: (纸本)9798331510596

In recent years, the complexity of semiconductor manufacturing processes has increased, leading to a growing need for the high-precision optimization of device structures. For example, in batch-type wet etching devices, the flow of chemical liquids within the process bath can vary depending on the device structure, causing variations in the etching state of the wafer. This issue is addressed using a feedback mechanism that iteratively adjusts the device structure based on the results of an etching experiment, thereby achieving more uniform etching conditions. However, this approach requires a large number of trial experiments. In the fabrication process of 3d flash memories, word-line formation in the silicon substrate requires precise control of the silicon concentration in the etching solution. However, this concentration can fluctuate due to the dissolution of the SiN film during the etching process. This fluctuation may cause various problems. This study introduces an innovative method utilizing multi-objective Bayesian optimization to derive optimal wet etching bath design parameters; this method is informed by image and physical-quantity data from fluid dynamics simulations [1]. This approach is validated through simulation experiments, whose results are used to identify the best possible wet etching bath designs effectively.

关键词： Wet etching Three-dimensional displays Fluid dynamics Semiconductor device manufacture Silicon Bayes methods Flash memories Optimization Silicon dioxide Substrates

来源：评论

学校读者我要写书评

暂无评论

A Product Classification recognition Model Based on Correlation Analysis and a Combination of Multiple Machine Learning Classification Algorithms

A Product Classification Recognition Model Based on Correlat...

引用

2023 International Conference on Intelligent Communication and Computer Engineering, ICICCE 2023

作者： Feng, Qiongzu Huang, Xinran Qu, Tonghui School of Computer Science and Technology Dongguan University of Technology Guangdong China School of Digital Art and Design Chengdu Neusoft University Chengdu China Software Product R&d Center Hangzhou Hikvision Digital Technology Hangzhou China

ISBN: (纸本)9798350395617

This article mainly studies the relationship between the composition and weathering of ancient glass products of different types, patterns and colors. Firstly, the data were preprocessed to exclude unreasonable data, and then the correlation analysis model and difference analysis model were used to fit the relationship between the weathering of the glass relics surface and the types, decorations and colors of the glass. Secondly, the missing data is processed, filled with mathematical expectations, and classified by random forest classification method, based on alumina. Then, by drawing the importance moving average reference graph, calculating the mathematical expectation and standard deviation, the acceptable domain is obtained, and the standard deviation graph is drawn to verify the rationality and sensitivity of the classification results. Then, ridge regression predictive analysis is used to analyze the training group, and the trained model is used to classify the test group, and the classification results are obtained. Finally, the model is verified according to the prediction accuracy obtained by training. This paper mainly introduces the principle of correlation analysis algorithm and ridge regression method, and describes how to apply these methods to solve the weathering problem of ancient relics. Pearson correlation analysis is a statistical method used to measure the strength of the linear relationship between two variables, while ridge regression is a regression analysis method used to deal with multicollinearity problems. These algorithms have been effectively applied in this paper, which can provide reference for other research on the weathering of ancient relics. © 2023 IEEE.

关键词： Weathering

来源：评论

学校读者我要写书评

暂无评论

dynamic Characteristics of Thin Film Structure with Concentrated Mass 3

Dynamic Characteristics of Thin Film Structure with Concentr...

引用

2023 3rd International Conference on Industrial Manufacturing and Structural Materials, IMSM 2023

作者： Cheng, Jingsong Wang, Sheng Nie, Ying Song, Lin Center for LTA System R&D Aerospace Information Research Institute Chinese Academy of Sciences No.9 Dengzhuang South Road Haidian District Beijing100094 China Yuquan Road Shijingshan District Beijing100049 China

Aiming at the design and analysis of flexible film building rigid flexible large complex structure system, this paper carried out the simulation analysis of dynamic characteristics of thin film structure with concentrated mass. The structural connectors with high stiffness and relatively concentrated mass are simplified as concentrated mass. The flexible thin film structure is simplified as a biaxially tensioned cross film material sample. The influences of pretension, quality, and concentrated mass of the size on the natural frequency of cross film material samples are analyzed. The results show that the natural frequency of the sample increases with pretension, decreases with the quality of the concentrated mass, and increases with the side length of the connection between the concentrated mass and the sample. The fundamental frequency of the sample is independent of the height of the concentrated mass, but the second frequency decreases with the height of the concentrated mass. The research results provide a theoretical basis for the overall design and analysis of building film structure and the further dynamic experiments of film material samples with concentrated mass. © 2023 Institute of Physics Publishing. All rights reserved.

关键词： Natural frequencies

来源：评论

学校读者我要写书评

暂无评论

On challenges of AI to cognitive security and safety

引用

Security and Safety 2023年第2期2卷 44-56页

作者： ruiyang Huang Xiaoqing Zheng Yuting Shang Xiangyang Xue National Digital Switching System Engineering&Technological R&D Center Zhengzhou 450002China Fudan University Shanghai 200433China

recent advances in deep learning have led to disruptive breakthroughs in artificial intelligence(AI),fueling the jump in ChatGPT-like large language models(LLMs).As with any emerging technology,it is a two-sided coin,bringing not only vast social impacts but also significant security concerns,especially in the socio-cognitive *** this back-ground,this work starts with an inherent mechanism analysis of cognitive domain games,from which it proceeds to explore the security concerns facing the cognitive domain as well as to analyze the formation mechanisms of a cognitive immune ***,inspired by behavioral mimicry in biology,this work will elaborate on new approaches to cognitive security from three aspects:Mimicry Computing,Mimicry defense,and Mimicry Intelligence.

关键词： Cognitive security Artificial intelligence Endogenous safety and security Mimicry intelligence

来源：评论

学校读者我要写书评

暂无评论

Interpretable Metric Learning: A Weight distribution Optimization Perspective 11th

Interpretable Metric Learning: A Weight Distribution Optimiz...

引用

11th International Conference on Applications and Techniques in Cyber Intelligence, ATCI 2023

作者： Miao, Xinyue Xu, Le Wen, Zhang Zhang, Lei Wei, Zhiqiang Wu, Changtao Shang, Zhengkai Wang, Kelin deng, Qianxing Wei, Fuyuan Liu, Liyan R&D Center for Ultra Deep Complex Reservior Exploration and Development CNPC Xinjiang841000 China Network and Information Center Southwest Petroleum University Sichuan Province Chengdu610500 China Tarim Oilfield Branch Company Petrochina Xinjiang841000 China School of Computer Science and Software Engineering Southwest Petroleum University Chengdu China

Metric learning aims to reduce intra-class distance and expand inter-class distance. This paper proposes a distribution optimization algorithm that incorporates interpretable feature engineering (dOIFE). dOIFE explores the local interpretability of each instance, reconstructs the feature space of the training set with reverse weighting, and optimizes the feature space of the target domain to address metric issues. Our approach innovatively derives interpretability weights from the classifier's perspective and applies fine-grained weighting to the instances to optimize the spatial metric. For any classifier, we conduct local interpretability analysis on its results to obtain fine-grained sample weights. In the training space, we designed an interpretable feature engineering method to expand inter-class distance and reduce intra-class distance. For the test space, we developed a unified method that combines ensemble inference and distance metrics to optimize feature distribution. We conducted experiments on classical balanced datasets, long-tail datasets, and image datasets. Comparative experiments with nine distance metric learning algorithms demonstrated the superiority of our algorithm. Statistical tests show that our algorithm can be applied to various classification models, improve classification performance, and offer a certain level of interpretability. © 2024 The Authors. Published by Elsevier B.V.

关键词： Optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

SecMVX:Analysis on the Vulnerability of Multi-Variant Execution

引用

China Communications 2021年第8期18卷 85-95页

作者： Bingzheng Li Zheng Zhang Xiaomei Wang Sheng Qu Jiangxing Wu State Key Laboratory of Mathematical Engineering and Advanced Computing Zhengzhou 450001China China National Digital Switching System Engineering and Technological R&D Center Zhengzhou 450002China

As an active defenses technique,multivariant execution(MVX)can detect attacks by monitoring the consistency of heterogeneous variants with parallel *** with patch-style passive defense,MVX can defend against known and even unknown vulnerability-based attacks without relying on attack feature ***,variants generated with software diversity technologies will introduce new vulnerabilities when they execute in ***,we analyze the security of MVX theory from the perspective of formal *** we summarize the general forms and techniques for attacks against MVX,and analyze the new vulnerabilities arising from the combination of variant generation *** propose SecMVX,a secure MVX architecture and variant generation *** evaluations based on CVEs and SPEC 2006 benchmark show that SecMVX introduces 11.29%of the average time overhead,and avoids vulnerabilities caused by the improper combination of variant generation technologies while keeping the defensive ability of MVX.

关键词： multi-variant execution software diversity cyberspace security

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：