Enriching Exhibition Scholarship: Reconciling Knowledge Graphs and Social Media from Newspaper Articles to Twitter

点击翻译按钮获取中文摘要

Artwork exhibitions bring diverse cultures together, exposing artists, scholars, and the public to new experiences. These events and the subsequent interactions affect stylistic trends, art markets, and cultural perceptions. Exhibitions are difficult to study because we lack holistic, structured data about co-exhibited objects and the participation of visitors. This project brings together an international and interdisciplinary team of experts in the fields of linked open data (LOD), exhibitions, art history, and AI to create and enhance machine-accessible exhibition descriptions. The Linked Art collaboration (https://linked.art/) has developed a modern Linked Open Data-based (LOD) set of standards to describe and publish art-related knowledge, including which objects were exhibited during which exhibitions. For this data to be truly connected, and allow scholars to seamlessly cross-institutional silos, descriptions of exhibition events and art objects must be "reconciled": the descriptions of the same object can be connected together by matching identifiers in different systems at different institutions.We will combine metadata graphs and cutting-edge natural language processing techniques to rich texts such as newspaper archives and social media. We will identify and extract objects, relevant descriptive information, and audience reactions from social texts across time. We will explore national and international exhibitions using open data, and detailed knowledge from university art galleries, including engagement with a contemporary Ashmolean exhibition via social media.