A Visual Analytic Observatory of Scientific Knowledge

该项目通过建立一个相互参照的科学断言储存库、研究影响和趋势的衡量标准和指标以及获取和探索科学知识的工具，增加了科学知识的可及性。该项目通过直接处理不确定因素和差异在理解和交流科学知识方面的整体作用，扩大了科学知识的研究范围。天文台跟踪科学文献中的科学断言如何随着时间的推移而演变。该项目通过提供参考资料提高了科学知识研究的能力，使研究人员能够定期和系统地比较和评估旨在捕捉科学贡献的影响的衡量标准和算法。科学知识的可获得性的增加提高了研究效率以及对科学和相关政策的理解。该项目将科学知识表现为一个由科学断言、推理和论证的结构以及从不断增长的科学文献中提取的相关证据组成的语义组织和适应性网络。使用来自机器学习、计算语言学、科学定量研究和视觉分析的技术，在科学出版物中综合提取的概念、语义关系和论证结构。此外，还确定了与科学断言及其影响有关的不确定性和差异，并在储存库中相互参照，以便跟踪其随时间的演变。现有和新开发的研究影响度量和指标与断言级别的证据和模式直接相关。天文台为研究人员和工具开发人员提供了一个可持续的平台，使他们能够获取科学知识和相关资源，以便在科学科学和创新政策研究的广泛背景下进行自己的研究和评估新的指标和算法的质量。

This project increase the accessibility of scientific knowledge by creating a cross-referenced repository of scientific assertions, metrics and indicators of research impact and trends, and tools for accessing and exploring scientific knowledge. The project expands the scope of the study of scientific knowledge by directly addressing the integral role of uncertainties and discrepancies in understanding and communicating scientific knowledge. The observatory tracks how a scientific assertion evolves over time in the scientific literature. The project advances the capabilities of the study of scientific knowledge by providing references so that researchers can routinely and systematically compare and evaluate metrics and algorithms designed to capture the impact of scientific contributions. The increased accessibility of scientific knowledge improves research productivity as well as the understanding of science and relevant policies.The project represents scientific knowledge as a semantically organized and adaptive network of scientific assertions, structures of reasoning and argumentation, and relevant evidence extracted from the growing body of scientific literature. Extracted concepts, semantic relations, and argumentative structures are synthesized across scientific publications using techniques from machine learning, computational linguistics, quantitative studies of science, and visual analytics. In addition, uncertainties and discrepancies associated with scientific assertions and their impacts are identified and cross-referenced in the repository so as to keep track of their evolution over time. Existing and newly developed metrics and indicators of research impact are linked directly to assertion-level evidence and patterns. The observatory serves as a sustainable platform for researchers and tool developers to access scientific knowledge and relevant resources for conducting their own studies and evaluating the quality of new metrics and algorithms in the broad context of the science of science and innovation policy research.