检索结果-内蒙古大学图书馆

International Conference on Networking and Network Applications (NaNA)

作者： Zhihui Wang Jianrui Chen Peijie Wang Tingting Zhu School of Computer Science Shaanxi Normal University Xi’an China Ministry of Education Key Laboratory of Modem Teaching Technology Xi’an China

Graph convolutional network (GCN) and dynamic evolutionary model are the mainstream collaborative filtering technologies in recent years. Nevertheless, the initial feature vectors selection problem of the existing recommendation algorithms based on dynamic evolutionary models leads to unstable recommendation accuracy. In addition, the collaborative filtering method of GCN does not take into account the dynamic evolution law of graph networks. Based on this, this research adopts GCN to train the initial embedding of the dynamic evolution model to perform collaborative filtering recommendation. First of all, a heterogeneous graph network is constructed by applying explicit feedback information (rating scores) of users. Secondly, the embedding of users and items are propagated through the dynamic evolution model. Then, the final embedding is obtained by weighting the embedding of each layer, and the scores are predicted. Finally, according to the Adam optimizer, the initial embedding of the dynamic evolution model is trained in the form of mini-batch to minimize the loss function. Experimental results show that the proposed algorithm is superior to several compared excellent algorithms in recommendation performance.

关键词： Knowledge engineering Technological innovation Heuristic algorithms Collaborative filtering Prediction algorithms Mathematical models

来源：评论

学校读者我要写书评

暂无评论

Automatic Calculation of Loss Rate of Cotton Pickers Based on Deep Neural Network and Image Processing 22

Automatic Calculation of Loss Rate of Cotton Pickers Based o...

引用

Proceedings of the 2022 5th International Conference on Machine Learning and Machine Intelligence

作者： Jialue Tian Zhongmin Qian Qing Xie Yongfei Sun Lei Shi College of Computer Science and Technology Nanjing University of Aeronautics and Astronautics China Fundamental Experiment Teaching Department Nanjing University of Aeronautics and Astronautics China Key Laboratory of Modern Agricultural Equipment Ministry of Agriculture and Rural Affairs China and Nanjing Institute of Agricultural Mechanization Ministry of Agriculture and Rural Affairs China

ISBN: (纸本)9781450397551

The loss rate of cotton pickers is an important indicator to measure the quality of cotton pickers, which is directly the vital interests of cotton farmers. To obtain the loss rate of cotton pickers in real-time, this paper takes the cotton field images after a cotton picker harvesting as the research object, combines the deep learning method and image processing technology, and proposes a method for calculating the loss rate of cotton pickers. First, nearly 5,000 ha-rvested cotton field images were collected manually to establish a data set. Then a lightweight Yolov5s neural network model was constructed to detect the cotton areas in the image and obtain the regions of interest (ROI). After that, each ROI was processed by binarization, dilation, erosion, and masking to obtain a more accurate cotton area, which excludes cotton husks, branches, and leaves in the entire image. Finally, a convolutional neural network was used to build a regression model to calculate the cotton loss rate for the whole image. The experimental results show that: 1) The constructed lightweight Yolov5s model can detect cotton areas at a speed of 43 frames per second and an average accuracy of 84.8%, which meets the needs of cotton pickers. 2) For the calculation of the image cotton loss rate, the difference between the calculated loss rate and the actual loss rate is less than 1% for 90% of the images; only 1.1% of the images have a difference greater than 2%, and the maximum difference does not exceed 3%. Therefore, the method proposed in this paper can realize the real-time calculation of the loss rate of cotton pickers.

关键词： Convolutional neural network Cotton picker loss rate Deep learning Yolov5s

来源：评论

学校读者我要写书评

暂无评论

Buffer-Aided Relay Selection for Wireless Cooperative Relay Networks with Untrusted Relays

Buffer-Aided Relay Selection for Wireless Cooperative Relay ...

引用

International Conference on Networking and Network Applications (NaNA)

作者： Cheng Zhang Xuening Liao Zhenqiang Wu Guoyong Qiu Key Laboratory of Modern Teaching Technology Ministry of Education Xi’an China School of Computer Science Shaanxi Normal University Xi’an China

Wireless cooperative relay networks are widely used in various fields. Wireless channels are vulnerable to eavesdroppers because of the broadcast characteristics of wireless channels, the security problem of wireless cooperative relay networks needs to be solved urgently, and the trust level of relays is rarely considered in existing work. Therefore, the paper considers a wireless cooperative relay network with a passive eavesdropping node and untrusted relays and designs a trusted buffer-aided relay selection (TBARS) scheme based on the fact that the channel state information (CSI) of eavesdropping channel is hard to obtain. The paper uses the secrecy outage probability (SOP) to measure the security performance, derives the expression of the SOP based on Markov Process, and verify the feasibility of the TBARS scheme by simulations and analyzes the impact of various factors on the SOP of the TBARS scheme. Finally, the paper compares the TBARS scheme with max-link scheme in terms of the SOP, and the simulations demonstrate that the proposed scheme outperforms the other schemes.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Combining Transformers with Natural Language Explanations

arXiv

引用

arXiv 2021年

作者： Ruggeri, Federico Lippi, Marco Torroni, Paolo Department of Computer Science and Engineering viale Risorgimento 2 Bologna Italy Department of Sciences and Methods for Engineering via Amendola 2 Reggio Emilia Italy

Many NLP applications require models to be interpretable. However, many successful neural architectures, including transformers, still lack effective interpretation methods. A possible solution could rely on building explanations from domain knowledge, which is often available as plain, natural language text. We thus propose an extension to transformer models that makes use of external memories to store natural language explanations and use them to explain classification outputs. We conduct an experimental evaluation on two domains, legal text analysis and argument mining, to show that our approach can produce relevant explanations while retaining or even improving classification performance. © 2021, CC BY-SA.

关键词： Text processing

来源：评论

学校读者我要写书评

暂无评论

WHEN TO TRUST AGGREGATED GRADIENTS: ADDRESSING NEGATIVE CLIENT SAMPLING IN FEDERATED LEARNING

arXiv

引用

arXiv 2023年

作者： Yang, Wenkai Lin, Yankai Zhao, Guangxiang Li, Peng Zhou, Jie Sun, Xu Center for Data Science Peking University China Gaoling School of Artificial Intelligence Renmin University of China Beijing China Beijing Key Laboratory of Big Data Management and Analysis Methods Beijing China Shanghai AI Lab China Tsinghua University Beijing China Pattern Recognition Center WeChat AI Tencent Inc. China MOE Key Lab of Computational Linguistics School of Computer Science Peking University China

Federated Learning has become a widely-used framework which allows learning a global model on decentralized local datasets under the condition of protecting local data privacy. However, federated learning faces severe optimization difficulty when training samples are not independently and identically distributed (non-i.i.d.). In this paper, we point out that the client sampling practice plays a decisive role in the aforementioned optimization difficulty. We find that the negative client sampling will cause the merged data distribution of currently sampled clients heavily inconsistent with that of all available clients, and further make the aggregated gradient unreliable. To address this issue, we propose a novel learning rate adaptation mechanism to adaptively adjust the server learning rate for the aggregated gradient in each round, according to the consistency between the merged data distribution of currently sampled clients and that of all available clients. Specifically, we make theoretical deductions to find a meaningful and robust indicator that is positively related to the optimal server learning rate and can effectively reflect the merged data distribution of sampled clients, and we utilize it for the server learning rate adaptation. Extensive experiments on multiple image and text classification tasks validate the great effectiveness of our method. © 2023, CC BY.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

A Novel Hybrid Makeup Transfer Method Based on Latent Vector Manipulation

A Novel Hybrid Makeup Transfer Method Based on Latent Vector...

引用

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

作者： Yumiao Li Peng Li Ziqi Shen Wenhui Li Ministry of Culture and Tourism Key Laboratory of Intelligent Computing and Service Technology for Folk Song Xi'an China School of Computer Science Shaanxi Normal University Xi'an China Key Laboratory of Modern Teaching Technology Ministry of Education Xi'an China Engineering Laboratory of Teaching Information Technology of Shaanxi Province Xi'an China Xi'an Key Laboratory of Cultural Tourism Resources Development and Utilization Xi'an China

With the development of image processing technology and the improvement of people's living standards, there is a wide demand for makeup services in society. In this paper, we focus on style transfer and its application in automatic makeup generation, and propose a facial feature fusion recovery technique that utilizes an adaptive spatial feature fusion attention mechanism to solve the problem of inconsistent clarity between the reference image and the source image. By using the moving least squares method to calculate similarity on each sample image, the reference image set achieving the optimal makeup transfer effect can be efficiently selected. At the same time, the latent vector can be manipulated to achieve on-demand hybrid makeup generation and de-makeup image generation. By assigning different weights, the proportion of makeup in specific reference image in the final makeup effect can be controlled, and a new mixed makeup style can be obtained. The experiment shows that the proposed model can achieve state-of-the-art makeup transfer effect, and can also generate suitable makeup effects for low resolution input images.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Blockchain in Digital Education Resource Sharing in Colleges and Universities

Research on Blockchain in Digital Education Resource Sharing...

引用

2021 IEEE International Conference on computer science, Electronic Information Engineering and Intelligent Control Technology, CEI 2021

作者： Zhu, Yongchuan Dang, Jianwu Wang, Yangping Yong, Jiu School of Electronic and Information Engineering Lanzhou Jiaotong University Lanzhou730070 China Gansu Provincial Engineering Research Center for Artificial Intelligence and Graphic Image Proc. Lanzhou730070 China Experimental Teaching Cener on Computer Science Lanzhou Jitotong University Lanzhou730070 China

ISBN: (纸本)9780738146492

Blockchain has been widely concerned by the society since its birth. With the development of society, it has shown a wide application prospect in finance, supply chain, medical treatment, education and other fields. The co-construction and sharing of digital education resources is an important way to promote the rapid development of education information. Firstly, the problems encountered in the co-construction and sharing of digital education resources in current universities are analyzed. Then, the model is constructed from the perspective of education through the network structure of blockchain. Building a resource sharing model of university education based on block chain technology can not only increase the sharing range of educational resources, improve the quality of educational resources, but also deposit the intellectual property rights of the uploader. In this research, an improved PBFT algorithm (SPBFT) is designed to reduce the time complexity of the consensus process from O(N2) to O(N), which effectively reduces the communication overhead. The average delay is from 57ms to 34ms, and the average throughput is from 330TPS to 680TPS. © 2021 IEEE.

关键词： Blockchain

来源：评论

学校读者我要写书评

暂无评论

A Sentiment Classification Model Based on Attention TIG-CNN-BiGRU

A Sentiment Classification Model Based on Attention TIG-CNN-...

引用

Culture-Oriented science and Technology (CoST), International Conference on

作者： Peng Li Tianmian Liang Ning An Lichen Zhang Xiaojun Wu Xiaoming Wang Junling Lu Key Laboratory of Intelligent Computing and Service Technology for Folk Song Ministry of Culture and Tourism Xi'an China School of Computer Science Shaanxi Normal University Xi'an China Key Laboratory of Modern Teaching Technology Ministry of Education Xi'an China Engineering Laboratory of Teaching Information Technology of Shaanxi Province Xi'an China Xi'an Key Laboratory of Cultural Tourism Resources Development and Utilization Xi'an China

The uniqueness of folk song expression forms and complex semantic relationships, and the existing models are not good at recognizing the emotion category of folk song lyrics, so an attention TIG-CNN-BiGRU based emotion classification model is proposed. That is, we first propose a text vector representation based on TF-IDF feature weighted pretrained Glove word embedding to extract important features by assigning different weights, and then combine CNN, BiGRU and attention mechanism to extract local features using multi-scale convolutional kernel CNN, BiGRU to capture semantic dependencies of long sequences and attention mechanism to assign high weights to key features to make the model have better lyrics recognition ability. Experiments comparing the model with various baseline methods on the lyrics dataset show that the model outperforms other lyrics sentiment classification models in terms of sentiment classification accuracy and various other metrics.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Reform and practice of web design and production based on OBE concept in the context of new engineering 8

Reform and practice of web design and production based on OB...

引用

8th Annual International Conference on Geo-Spatial Knowledge and Intelligence, GSKI 2020

作者： Guo, En Xu, Weizhi Yu, Hui Zhong, Yujie School of Computer Science Shandong Xiehe University Room 610 No.5 Teaching Building Yuaoqiang Campus Licheng District Jinan Shandong China School of Information Engineering Shandong Normal University 1 University Road University Science park Changqing District Jinan Shandong China Xia Zhang Town Government Tai’an City Daiyue district Shandong China

In this paper, based on the characteristics of web design and facture course, from the course introduction, teaching existence question analysis, training mode reform aspects and so on three aspects to illustrate how to improve the teaching effect of this course, and to explore the concept of how to use the OBE complete curriculum reform in teaching practice, thus improve the computer professional and visual communication professional social practice ability of students, in order to conform to the demand of the current web design position. © 2021 Institute of Physics Publishing. All rights reserved.

关键词： Web Design

来源：评论

学校读者我要写书评

暂无评论

Pattern Recognition for Driving Scenario Detection in Real Driving Data 31

Pattern Recognition for Driving Scenario Detection in Real D...

引用

31st IEEE Intelligent Vehicles Symposium, IV 2020

作者： Montanari, Francesco German, Reinhard Djanatliev, Anatoli Department for Processes Methods Tools Virtual Environment Audi Ag Ingolstadt85045 Germany Friedrich-Alexander-University Department of Computer Science 7 Erlangen-Nuremberg91058 Germany

For the scenario-based development and testing of automated and connected driving an unknown huge number of different driving scenarios is needed. In this paper we propose an approach that extracts driving scenarios from real driving data without any requirement of predefinitions or rules. Instead of searching for specific scenarios in the data, we cluster recurring patterns and interpret the resulting clusters as potential scenario groups. The method shows promising results. In the exemplary clustering we are able to detect four main scenario groups and corner cases within the clusters. With an huge amount of data this method could be used in the future to set up a scenario database in an automatic manner. © 2020 IEEE.

关键词： Intelligent systems

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：