Novel application of computational approaches in addressing problematic terminology and descriptions within V&A Museum catalogues

“计算方法在解决V&A博物馆目录中有问题的术语和描述中的新应用”研究了计算方法在解决维多利亚和阿尔伯特博物馆目录中有问题的术语和描述中的应用。它应用并扩展了计算语言学和信息科学领域的现有方法，以识别、分析和构建有关博物馆目录数据中术语使用和描述的见解。该项目发生在V&A博物馆的工作背景下，其中需要更好地理解和处理藏品中的描述，目的是协助计算检测有问题的术语和描述。它由牛津大学的一个多学科团队提供支持，该团队拥有来自语言学和工程系以及牛津电子研究中心的专业知识。该研究将调查现有的（数字和非数字）方法和整个博物馆部门的实践，提出一个连贯的概念信息框架，在这个框架中，上述任务可以表示为需求，并与计算算法的能力相匹配。这一框架还将为馆藏和机构之间交换术语和说明分析提供基础，以便对一个目录的分析可以引导或促进对另一个目录的理解。最后，该项目将开发数字工具和方法，这些工具和方法可以在计算上协助检测和潜在地纠正目录数据中有问题的术语和描述。能够在不同的目录和机构之间交换和比较研究和数据将是未来数字实践的关键考虑因素，这是本项目将探索的。然而，在实践中开始这项工作之前，有必要更好地了解算法应用在不同目录和集合中的差异，否则任何结构化数据都有可能被误解为简化。通过描述不同目录和藏品的信息结构特征，该研究有可能为Linked Art提供一个超集。同样，这种对应用计算语言学如何受数据集特征影响的更好理解，将为使用来自其他地方的比较数据提供信息。该项目预期的适当的计算语言学方法是成熟的和容易理解的；因此，这些领域的创新并不是主要的结果。相反，研究的贡献是这些方法的应用和对实践的影响。这将是该项目的主要贡献：提出解决方案，将数字创新带入目录。该研究的智力重点将是确定和发展未来的数字实践，将计算语言学方法纳入机构内，特别是在编目中。殖民意识形态经常以难以使用计算工具和方法检测的方式嵌入博物馆：该项目旨在发现这些边界的限制，并试图询问哪些问题可以使用计算方法检测和纠正，哪些问题需要详细的手工工作。这个项目属于AHRC语言学-词汇研究领域。

"Novel application of computational approaches in addressing problematic terminology and descriptions within V&A Museum catalogues" investigates the application of computational methods in addressing problematic terminology and descriptions within Victoria & Albert Museum catalogues. It applies and extends existing methods from the fields of computational linguistics and information science to identify, analyse, and structure insights about terminology use and description within museum catalogue data. The project takes place in the context of work at the V&A in which descriptions within collections need to be better understood and addressed, with the aim of assisting computationally in the detection of problematic terminology and descriptions. It is supported by a multidisciplinary team at the University of Oxford, with expertise from the Departments of Linguistics and Engineering, and the Oxford e-Research Centre.The study will survey existing (digital and non-digital) approaches and practice from across the museum sector, proposing a coherent conceptual information framework within which the above tasks can be expressed as requirements, and matched to the affordances of computational algorithms. This framework will also provide the basis for exchange of terminology and description analyses between collections and institutions, so that analysis of one catalogue can bootstrap or inform understanding of another. Finally, this project will develop digital tools and methods which can computationally assist in the detection, and potentially correction, of problematic terminology and descriptions within catalogue data.Being able to exchange and compare research and data between different catalogues and institutions will be a key consideration for future digital practice, which this project will explore. However, before embarking on this work in practice, it is essential to gain a better understanding of how algorithmic application varies across different catalogues and collections, since otherwise any structured data risks being misleadingly reductive. By characterising the information structures for different catalogues and collections, the study has the potential to propose a superset for use by Linked Art. Similarly, this improved understanding of how applied computational linguistics is affected by the characteristics of the dataset will inform the use of comparator data from elsewhere.The appropriate computational linguistics methods anticipated for this project are mature and well-understood; therefore, innovation in these is not intended to be a primary outcome. Rather, the research contribution is the application of these methods and the implications for practice. This will be the principal contribution of the project: to propose solutions bringing digital innovations to catalogues as they really are.The intellectual focus of the study will be in identifying and developing future digital practice to incorporate computational linguistics methods within the institution, particularly in cataloging. Colonial ideology is often embedded in museums in ways that make it very difficult to detect using computational tools and methods: this project aims to discover the limits of these boundaries and seeks to ask which issues can be detected and corrected using computational approaches and which require detailed manual work.This project falls within the AHRC Linguistics - Lexicon research area.