The Interface(s) of a Virtual National Collection: Deep Discoveries

Panel Abstract: Towards a National Collection (TaNC) is a five-year £18.9 million investment in the UK’s world-renowned museums, archives, libraries and galleries (GLAM). TaNC will allow to formulate new research questions, increase visitor numbers, and expand and diversify virtual access to UK heritage. Eight small-scale Foundation Projects – made of collaborations between Higher Education Institutions (HEIs) and Independent Research Organisations (IROs) - began on 1 February 2020, aiming to lay the foundations for this future virtual national collection. This panel includes papers by the representatives of three Foundation Projects: Deep Discoveries; Locating a National Collection, and Engaging Crowds. Each project is developing a proof of concept user interface that aims at engaging different audiences (from experts to casual users) at different levels of agency (from participatory methods to serendipitous discovery) - setting an example for organisations dealing with similar digital collections and technologies. We are also submitting the interfaces demos in the poster/ demo section.

Paper Abstract: Deep Discoveries joins HEI-based computer vision experts, GLAM professionals, and UX researchers to explore the opportunities afforded by AI-enabled visual similarity recognition technologies for cross-collection image searching. A central aim of the project is the delivery of a prototype to demonstrate the ‘research’ and ‘discovery’ potentials of a novel technology to multiple types of users. The complex goal of bridging diverse user searching tasks was further complicated by the distinct drivers of our transdisciplinary team. A critical intervention occurred at the interface design stage, with the integration of a Design Research team and the development of a shared vision and a joint vocabulary. The interface design served to cultivate a balance between advancements in computer vision technology and existing end-user knowledge, skills and adaptability. Our teams came to view the interface as a boundary object, a tool that enabled, if not an agreement on the research approach, then certainly a shared understanding and way forward. We propose that the interface can serve as a site for collaboration across the TaNC programme, mediating the distinctive drivers and professional demands of all stakeholders, from the complex needs of specialist researchers and requirements of heritage organisations, to the discovery-driven demands of general audiences.