Temporal Text Classification: Computer-Based Dating of Medieval Irish Texts

The Unit for Linguistic Data at the Insight SFI Research Centre for Data Analytics / Data Science Institute, National University of Ireland Galway is delighted to welcome Greg Toner, professor of Irish at Queen’s University Belfast, to be the next speaker in our seminar series. He will present his research on the application of machine-learning techniques to the dating of medieval Irish texts. Register here.

Abstract:

This paper presents the results of a Leverhulme-funded project which explored the applicability of machine-learning techniques to the dating of medieval Irish texts. We will explore the methods developed in order to construct a model which achieved a 75% success rate in dating individual annual entries to within +/- 25 years. We will then examine the application of this model to non-annalistic texts, including the methods that we have developed to cope with genre variation, multiple strata and bias within the data.

About the Speaker:

Greg Toner holds the position of Professor of Irish at Queen’s University, Belfast, having previously held the Professorship of Irish Language and Literature at the University of Ulster. He is not only a Member of the Royal Irish Academy but chairs the management committee of the RIA’s Foclóir Stairiúil na Gaeilge. He is also a member of the Arts and Humanities Research Council Peer Review College and Honorary General Editor of the Irish Texts Society.

Professor Toner is a trailblazer in the area of digital Early Irish research, having directed the Electronic Dictionary of the Irish Language project (eDIL) which saw the digitisation and further development of the original hard-copy Dictionary of the Irish language (DIL). Since the completion of the project, eDIL has become a central resource in medieval Irish studies for language learners and seasoned academics alike. He also developed the widely used Computer-Assisted Practice Activities for Irish Language Learners (CAPAILL). More recently, he led a Leverhulme-funded project to explore the use of computational methods for dating medieval Irish texts. This investigation may represent the first use of machine learning in Early Irish research and forms the basis of his presentation today.

Host:

The seminar series is led by the Cardamom project team. The Cardamom project aims to close the resource gap for minority and under-resourced languages using deep-learning-based natural language processing (NLP) and exploiting similarities of closely related languages. The project further extends this idea to historical languages, which can be considered closely related to their modern form. It aims to provide NLP through both space and time for languages that current approaches have ignored.

Registration link: