Turing Lecture: Data science or data humanities?
Tuesday 05 Mar 2019
Time: 14:00 - 16:00
Bio
Melissa Terras is the Professor of Digital Cultural Heritage at the University of Edinburgh's College of Arts, Humanities, and Social Sciences (CAHSS), leading digital aspects of research within CAHSS as Director of the Centre for Data, Culture and Society, and Director of Research in the Edinburgh Futures Institute. Her research focuses on the use of the digitisation of cultural heritage including capture, usage, and further computational analysis. With a background in Classical Art History and English Literature (MA, University of Glasgow), and Computing Science (MSc IT with distinction in Software and Systems, University of Glasgow), her doctorate (Engineering, University of Oxford) examined how to use image processing and machine learning to interpret and read deteriorated Ancient Roman texts.
She was employed in UCL Department of Information Studies from 2003-2017, Directing UCL Centre for Digital Humanities from 2013, and becoming Professor of Digital Humanities in 2013. Books include 'Image to Interpretation: An Intelligent System to Aid Historians in Reading the Vindolanda Texts' (2006, Oxford University Press), 'Defining Digital Humanities: A Reader' (2013, Ashgate) which has been translated into Russian and Chinese, and 'Electronic Legal Deposit: Shaping the Libraries of the Future’ (2020, Facet). She is on the Board of Directors of Transkribus, the consumer-facing Handwritten Text Recognition infrastructure spun-out from H2020 machine-learning research. She is a Trustee of the National Library of Scotland, served on the Board of Curators of the University of Oxford Libraries, is a Fellow of the Chartered Institute of Library and Information Professionals, and Fellow of the British Computer Society. You can generally find her on Twitter @melissaterras.
Research interests
A growing source of large scale data is emerging from mass-digitisation programs within the Gallery, Library, Archive and Museum (GLAM) sector. However, most GLAM institutions have neither the expertise nor the resources to allow effective mining of this content that moves beyond basic keyword searching of OCRd text provided by standard search interfaces.
Melissa's work will:
- Investigate the establishment of a shared data repository of machine processable historical texts for use in Turing research, accessible by all Turing partners.
- Encourage and facilitate use of available historical data sets in computational research at the Turing, supporting and promoting the benefits of utilising such historical content
- Understand the mechanisms of access for humanities researchers wishing to undertake data driven analysis of historical text material.
- Explore the frameworks for re-purposing digitised content as data.
While there are a range of technical barriers (OCR correction, non-standardised data management) as well as interdisciplinary issues (access to large scale compute has been difficult for humanities researchers as there is too steep a learning curve), the establishment of a shared resource of historical texts from GLAM partners would enable humanities research at the Turing to be truly supercharged, laying the groundwork for the secure sharing and text mining of in-copyright GLAM material (such as journal or web archive content) and can be seen as exploratory research which assists large scale mining of content while engaging with the technical practices and legal frameworks of the GLAM sector.
