情报研究中多源异构数据的小算力聚合及应用研究  

Research on Small Computing Power Aggregation and Application of Multi-source Heterogeneous Data in Intelligence Analysis

作　　者：王曰芬 张华俊[1] 何劲 Wang Yuefen;Zhang Huajun;He Jin(Institute for Big Data Science,Tianjin Normal University,Tianjin 300387;School of Economics and Management,Nanjing University of Science and Technology,Jiangsu Nanjing 210094)

机构地区：[1]天津师范大学大数据科学研究院,天津300387 [2]南京理工大学经济管理学院,江苏南京210094

出　　处：《情报理论与实践》2024年第7期81-87,共7页Information Studies:Theory & Application

基　　金：国家社会科学基金一般项目“全文计量分析视角下学科交叉的多元体系化测度研究”的成果之一,项目编号:22BTQ098。

摘　　要：[目的/意义]探索采用低能耗、高效率的算力,应对情报研究中处理语义环境高复杂、知识边界不确定、信息博弈强对抗、数据样本稀疏为主要特征的多源异构数据聚合面临的挑战。[方法/过程]剖析在大数据、人工智能技术快速发展背景下,多源异构数据、小算力、聚合等数据处理相关的概念内涵及其情报研究发展趋势,基于I/O模型和数据流转换分析人机混合增强算力作用路径及其支撑数据价值实现的层次,提出情报研究中人机混合增强算力的多源异构数据聚合模式(MHDAM-HAC),并以动态情报研究业务为对象进行实例应用。[结果/结论]技术领域发展状态概要画像、聚焦某类技术主题挖掘发现隐含知识的两个实例,验证了所构建的多源异构数据的小算力聚合模式可以有效改善情报研究工作面临的问题,实现一部分情报生成过程的可重现,并促进情报研究的新发现,可作为新实践当中的实现方式。[Purpose/significance]This paper explores a new pattern.It is used to address such a challenge that using low energy consumption and high efficiency computing power to process multi-source heterogeneous data characterized by high complexity of semantic environments,uncertain knowledge boundaries,strong adversarial information games,and sparse data samples.[Method/process]In the context of the rapid development of big data and artificial intelligence technology,the paper analyzes the concepts of multi-source heterogeneous data,small computing power,aggregation and development trends in intelligence analysis.Based on I/O model and data flow conversion,studies the path of human-machine augmented computing power and the hierarchy of supporting data value implementation,proposes a multi-source heterogeneous data aggregation model with human-machine augmented computing power in intelligence analysis(MHDAM-HAC),and apply it to the dynamic intelligence analysis work.[Result/conclusion]By summarizing the development status of the technical field and focusing on certain technical topics to discover the hidden knowledge,it is verified that the small computing power aggregation mode of multi-source heterogeneous data constructed in this study can effectively improve the problems faced by intelligence analysis work,achieve the reproducibility of some intelligence generation processes,and promote new discoveries in intelligence analysis,which can be used as an implementation method in new practices.

关 键 词：多源异构数据 聚合模式 人机混合增强算力 人机协作 情报研究 

分 类 号：G350[文化科学—情报学]