| 索传军,盖双双,周志超.认知计算——单篇学术论文评价的新视角[J].中国图书馆学报,2018,44(1):50~61 |
| 认知计算——单篇学术论文评价的新视角 |
| Cognitive Computing: A New Perspective for Evaluating the Individual Academic Paper |
| 投稿时间：2017-09-20 修订日期：2017-10-18 |
| DOI： |
| 中文关键词: 学术论文 论文评价 评价本质 评价方法 认知计算 |
| 英文关键词: Academic paper Paper evaluation Evaluation essence Evaluation method Cognitive computing |
| 基金项目: |
|
| 摘要点击次数: 3211 |
| 全文下载次数: 1433 |
| 中文摘要: |
长期以来,同行评议的主观性和低效率、文献计量学的人为操纵和不针对内容等问题始终没有得到很好的解决,学术论文评价面临不同评价方法之间难以相互弥补的矛盾,亟需新的评价理论和方法。认知计算是一个以一定的规模进行学习,带着某种目的进行推理,以此来与人类进行交互的系统。将认知计算引入学术论文评价领域,有望同时解决专家定性评价中的主观性和低效率,以及定量评价中不针对内容等缺陷。本文介绍了认知计算的相关研究和实践,探讨了学术论文评价的发展历程及其困境,揭示了学术论文评价的本质,重点论述了基于认知计算的学术论文评价新思路。研究发现,学术论文评价的发展与科学交流方式和技术变革密切相关。学术论文的评价本质是对其学术价值的评价,具体体现为创新性的测度。从学术论文语义内容角度,构建基于认知计算的学术论文评价系统是完善现有论文评价理论和方法的新视角,有望成为未来学术论文评价的重要发展方向之一。图3。参考文献43。
|
| 英文摘要: |
Academic paper evaluation is a classic problem in the field of library science. Its main purpose is to help users find the excellent papers they need. The qualitative evaluation based on peer review and the quantitative evaluation based on bibliometrics are the most accepted evaluation methods of academic papers. Although qualitative evaluation is based on contents in each paper, it is inefficient and susceptible to expert subjectivity or other non scientific factors, not suitable for the highly efficient evaluation of massive papers. Although quantitative evaluation is objective, efficient and operable, it is not directly related to the content of the paper and it is easily manipulated. All the time, although scholars at home and abroad have never stopped investigating the evaluative methods of papers, they have never achieved satisfactory results. There is contradiction that different evaluation methods could not be complemented by each other. Therefore, new theory and method for evaluation are in urgent need.
Cognitive computing is a system that interacts with human beings by learning with a certain scale and reasoning with some purpose. It has obvious advantages in understanding and processing unstructured text. Cognitive computing is mainly used to solve problems with ambiguity and uncertainty. The introduction of cognitive computing into the evaluation field of academic papers is expected to solve both subjectivity and inefficiency in qualitative evaluation, as well as the lack of analyzing content in quantitative evaluation. Through the relevant document surveys, this paper discusses the research and practice of cognitive computing, analyzes the development process of academic paper evaluation, and discusses the development bottleneck that the current evaluation methods face with and evaluation essence. Mainly, this paper focuses on the new perspective based on cognitive calculation to evaluate academic papers, and we analyze the realization path of cognitive computing in the evaluation of academic papers and key issues in constructing a cognitive computing system for academic papers.
It is found that the development of academic paper evaluation is closely related to the changes in the way of scientific communication. The essence of academic paper evaluation is the evaluation of its academic value. Social value and economic value are the results of the application of academic value in different fields. The academic value and quality of an academic paper depend on its originality. From the perspective of semantic content, combing the multisource data sets and related knowledge base to construct the academic evaluation system based on the cognitive computing is expected to become one of the most important development directions of the evaluation of academic papers in the future.
The cognitive computing system proposed in this paper is a new idea for the evaluation of academic papers, and makes full use of the current advanced technology and big data thinking, which is of great value to optimize the existing evaluation theory and practice of academic papers. Firstly, the cognitive computing system of academic papers takes into account the multiple features including paper content, reference and citing papers, to realize the perfect combination of peer review and bibliometrics, and can make up its respective defects simultaneously. Secondly, the evaluation of academic papers based on cognitive computing is helpful in the fields of research management, discipline construction, reviews by editor and expert, and user's literature acquisition and reading experience et al. However, we are still in the early stage of cognitive system development, and there are several challenges to apply cognitive computing to academic paper evaluation. Among them, the datamation and semantization of the academic papers, and the machine learning model for the evaluation of academic papers are the two key issues that need to be solved in the construction of the cognitive computing system for academic papers. 3 figs. 43 refs.
|
|
查看全文
查看/发表评论 下载PDF阅读器 |
|
|
|
