Gaelic Algorithmic Research Group (GARG)

Description

The Gaelic Algorithmic Research Group (GARG) is an international research group devoted to developing modern technologies for Gaelic and utilising them both to strengthen and better understand the language and its associated culture.

Given the importance of technology in modern communication, members believe that it is central to the language revitalisation effort. They also believe that it is crucial to develop technologies for Gaelic users, with Gaelic users. Through cooperation between universities, public bodies and the Gaelic community itself, the Group aims to create a growing platform for Gaelic in technologically mediated spaces. Ultimately, its ambition is to create new domains for Gaelic-usage and bolster the future stability of the language.

The group's work is built on speech and language technology. Its ÈIST project, short for Ecosystem for Interactive Speech Technologies, launched in 2023 and develops Gaelic speech-to-text that also handles English in order to cope with code-switching; the first released model is based on an open-weight Whisper model and is published so that developers and researchers worldwide can embed Gaelic speech recognition in applications ranging from educational software to digital assistants. The group has also built an orthographic normaliser that updates historical Gaelic writing to modern spelling using an AI model developed at Edinburgh, a linguistic analyser, and a crowdsourcing platform on which volunteers transcribe Gaelic recordings held on the Tobar an Dualchais portal, generating training data that improves future speech recognition models. A related strand used automated text and handwriting recognition to convert thousands of folklore manuscripts into digital text at error rates below five per cent, and the group argues publicly for ethical practice in this work, including transparent training data and returning outputs to the Gaelic community through the Digital Archive of Scottish Gaelic rather than depending on large technology companies.

Currently active

Yes

Offers funding

No, this infrastructure does not provide funding.

Founding year

2020

Contact details

On the map

Go to larger version of this map

Categorisation

Type

Tags

University affiliation(s)

University of Edinburgh
Edinburgh

University of Glasgow
Glasgow

Partner Infrastructures

Centre for Scottish and Celtic Studies

The Centre for Scottish and Celtic Studies (CSCS) is a thriving multidisciplinary and interdisciplinary collaboration that brings together scholars and postgraduate communities from various department… read more about Centre for Scottish and Celtic Studies

Glasgow

Language Technology Group (LTG)

The Edinburgh Language Technology Group (LTG) is a research and development (R&D) group working in the area of natural language engineering since the early 1990s. It now functions cross-college at… read more about Language Technology Group (LTG)

Edinburgh

Bòrd na Gàidhlig

Bòrd na Gàidhlig is the principal public body in Scotland responsible for promoting the development of Gaelic, including advising Scottish Ministers on Gaelic issues. Its stated aim is f… read more about Bòrd na Gàidhlig

Inverness

Additional Partnerships

BBC ALBA; MG ALBA; NVIDIA; Tobar an Dualchais / Kist o Riches; Digital Archive of Scottish Gaelic (DASG); National Trust for Scotland; University of the Highlands and Islands; School of Scottish Studies Archives