当期目录

15 November 2020, Volume 6 Issue 6

Previous Issue    Next Issue

APPLICATION

Application and exploration of automatic generation technology in artificial intelligence platform

Zhengxun XIA,Yifan YANG,Shengmei LUO,Dachao ZHAO,Yan ZHANG,Jianfei TANG

2020, 6(6):  0.  doi:10.11959/j.issn.2096-0271.2020058

Asbtract ( 205 )   HTML ( 36)   PDF (1449KB) ( 625 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

With the development of artificial intelligence (AI) technology,AI applications have entered a period of rapid popularization,facing the rapidly growing market demand,it is necessary for AI platforms to use automated methods to improve development efficiency.Based on the analysis of the research progress of generation technology,the status quo and challenges of AI platform,based on generation technology,the automation of AI platform’s front and rear end adaptation,performance optimization,and model security enhancement were realized,which can generate data or code according to the needs of the context and to meet the requirements in a more flexible way.It can also avoid a lot of manual work and can effectively improve development efficiency and reduce development cost.

Research on person-position relationship based on relation graph

Xiaoping WANG,Mengjie GUO,Jingwen YUE

2020, 6(6):  0.  doi:10.11959/j.issn.2096-0271.2020059

Asbtract ( 292 )   HTML ( 55)   PDF (1370KB) ( 648 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Utilizing the existing data to further analyze and help leaders organize their work is a potential and challenging direction.According to the characteristics of leader information data,the leadership team was analyzed using the person-position relationship judgment method based on relation graph,a relation graph was built by integrating the leader resume and the multi-source information from the database.Then the information such as nodes and relationships in the relation graph extracted by the network representation learning method was used as features to input into the classification model.By using the proposed model,the relationship between people and positions can be inferred.The experimental results show that the method based on relation graph can well capture the complex relationship between people and positions,and can accurately judge the person-position relationship.

FORUM

The required authorization to the data-centric economic activities

Yangyong ZHU,Yun XIONG

2020, 6(6):  0.  doi:10.11959/j.issn.2096-0271.2020060

Asbtract ( 274 )   HTML ( 55)   PDF (1440KB) ( 707 )   Knowledge map   

References | Related Articles | Metrics

With the rapid development of the digital economy for which data is the key elements,there are more and more disputes and cases involving data.Since the data element is a newly concept,it is necessary to clarify the types of the data-centric economic activity,and then what rights need to be authorized to whom executing these data-centric economic activities should be designed.The data-centric economic activities were divided into five categories according the characteristics of data:data production and reproduction,data right confirmation,data usage and service,data sharing,and scientific research.Correspondingly,four types of rights were proposed including data production right,data proprietary right,data access right,and data sharing right,to provide a reference for legislation of data use and development of dagital economy.

TOPIC：EDUCATIONAL BIG DATA

Research and practice on education data standard system for data sharing

Lin YANG, Wei WANG, Ji ZHU, Mingzheng WANG

2020, 6(6):  3-13.  doi:10.11959/j.issn.2096-0271.2020050

Asbtract ( 553 )   HTML ( 95)   PDF (1575KB) ( 720 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Under the background of accelerating education informatization and gradual opening of government data,the sharing and opening of education data has become an urgent problem,and the construction demand of education data standard system for sharing has emerged.Through the analysis of education data sharing requirements and standardization construction requirements,as well as the research on the positioning of data sharing standards in the education data standard system,the contents of education data sharing standards were clarified.In addition,the construction methods and characteristics of education data standard system in Shanghai was introduced as a case to provide reference for the education data sharing in other regions.

Research on the mechanism and key technologies for big data collection in education

Huanyou CHAI, Sannyuya LIU, Lingyun KANG, Yaxian ZHANG, Qing LI, Zhi LIU

2020, 6(6):  14-25.  doi:10.11959/j.issn.2096-0271.2020051

Asbtract ( 483 )   HTML ( 115)   PDF (712KB) ( 769 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

The mechanism and its related technologies of data collection are the foundation of realizing the valuable potential of applying big data in education,and therefore are vital for the construction and application of big data in education.The contents,methods,means,standards and specifications of data collection of big data in education were outlined.Combining with the practical problems in the construction and application of big data in education,guidance on research on big data in education was provided,from the perspectives about balancing data sharing and privacy protection,driving data governance and talent acquisition,and improving the mechanisms and related technologies respectively.

Mining prerequisite relations among learning objects

Kexin MA, Bifan WEI, Jie MA, Jun LIU, Yi HUANG, Min HU, Junlan FENG

2020, 6(6):  26-39.  doi:10.11959/j.issn.2096-0271.2020052

Asbtract ( 356 )   HTML ( 38)   PDF (2018KB) ( 482 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Prerequisite relation refers to the learning dependency between learning objects.Most previous works mined prerequisite relations in a pipelined way and heavily relied on hyperlinks,which lead to the accumulation of errors.To address these issues,prerequisite relations among knowledge topics were analyzed,and the asymmetry feature of prerequisite relation was found out.An end-to-end prerequisite relation model for mining prerequisite relations from texts was proposed.Based on the hyponymy relations between terms extracted from texts,this model calculates the asymmetry of prerequisite relation among related terms of learning objects,and then predicts the prerequisite relation betweens learning objects.The experimental results show that the proposed method achieves the state-of-the-art performance.

Construction and application of regional education platform based on data intelligence

Xiangchun HE, Shaoqing GUO

2020, 6(6):  40-51.  doi:10.11959/j.issn.2096-0271.2020053

Asbtract ( 262 )   HTML ( 42)   PDF (1637KB) ( 546 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Data intelligence leading is an important direction of regional education platform construction and application in the new era.Starting from the current situation and challenges of the construction and application of regional education platform,the development stage of regional education platform based on data intelligence was clarified from the perspective of concept.The construction ideas of using iterative thinking to promote platform construction,adhere to three guidance to overall planning platform construction,focus on three key points to promote platform construction,innovate construction mechanism to boost ecological construction,were put forward.The overall framework and key technologies of the platform were analyzed from the perspective of technology.The application promotion path,application focus and application promotion mode were put forward from the perspective of application,and typical practice cases were given,which provides reference for researchers and practitioners in related fields.

Research on privacy protection mechanism and technology of educational big data

Jieyu YUE, Chaoyang LUO, Jingshu DING, Qing LI

2020, 6(6):  52-63.  doi:10.11959/j.issn.2096-0271.2020054

Asbtract ( 334 )   HTML ( 63)   PDF (748KB) ( 662 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

With the deepening application of big data technology in the field of education,the potential risks such as personal privacy security that may be brought by the collection,analysis and sharing of educational data can not be ignored.How to ensure the safety of educational data,reasonable and compliance analysis and utilization of educational data are the urgent problems to be solved.Based on the characteristics of education big data,the privacy connotation of education big data was focused on.The privacy protection framework of education big data was put forward around the privacy protection needs of stakeholders in the life cycle of education data.And the current privacy protection technology system that can be used in education field was combed,so as to provide support for various application links of education big data,promote the standardized and orderly development of education big data.

STUDY

Research review of federated learning algorithms

Jianzong WANG, Lingwei KONG, Zhangcheng HUANG, Linjie CHEN, Yi LIU, Anxun HE, Jing XIAO

2020, 6(6):  64-82.  doi:10.11959/j.issn.2096-0271.2020055

Asbtract ( 2184 )   HTML ( 357)   PDF (1224KB) ( 2949 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

In recent years,federated learning has been proposed and received widespread attention to overcome data isolated island challenge.Federated learning related researches were adopted in areas such as financial field,healthcare domain and smart city related application.Federated learning concept was introduced into three different layers.The first layer introduced the definition,architecture,classification of federated learning and compared the federated learning with traditional distributed learning.The second layer presented comparison and analysis of federated learning algorithms from machine learning and deep learning aspects.The third layer separated federated learning optimization algorithms into three aspects to optimize federated learning algorithm through reducing communication cost,selecting proper clients and different aggregation method.Finally,the current research status and three main challenges on communication,heterogeneity of system and data to be solved were concluded,and the future prospects in federated learning domain were proposed.

State of the art and future perspectives of the applications of deep learning in the medical image analysis

Lihui WANG, Yongbin QIN

2020, 6(6):  83-104.  doi:10.11959/j.issn.2096-0271.2020056

Asbtract ( 1068 )   HTML ( 163)   PDF (1732KB) ( 1650 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

Medical imaging is an important auxiliary tool for clinical diagnosis.Medical images occupy almost 90% of clinical data.Therefore,mining medical image information will be beneficial for intelligent diagnosis,decision-making and prognosis prediction.With the emergence of deep learning,using deep neural networks to analyze medical images has become a hot research topic.Following the process of medical image analysis,from the image acquisition,the image pre-processing to the classification and prediction,the state-of-the-art applications of deep learning in each step of medical image analysis was elaborated,and according to the existed issues and the challenges,the future perspectives were finally discussed.

Research on demand identification for customized bus based on multi-source mobility data

Xi CHEN, Yinhai WANG, Zhuang DAI, Xiaolei MA

2020, 6(6):  105-118.  doi:10.11959/j.issn.2096-0271.2020057

Asbtract ( 292 )   HTML ( 45)   PDF (1927KB) ( 564 )   Knowledge map   

Figures and Tables | References | Related Articles | Metrics

As a new type of public transit services,the demand identification of customized bus (CB) has great practical significance,as well as the basis of route design of CB.Under the context of big data,a methodology framework of CB demand identification based on multi-source data was proposed by mining spatial-temporal characteristics from large scale mobility data.The proposed framework includes several phases,which are the identification of commuters from transit and Internet users,data fusion of travel demands and stop deployment method.This study takes Chengdu city as an example to verify the effectiveness of the proposed methods.The results can provide a theoretical support of CB route design.