Teresa Lynn, ADAPT Centre

I am very honoured to have been nominated for the 2023 ELRA Board Member elections. As a member of ELRA and a strong supporter of the LREC events, there are a number of reasons why I believe I am suitable for this role.

Firstly, I believe my combined background in research, academia and industry in the areas of NLP, MT, localisation and data collection would allow me to contribute to the ELRA association from a number of different perspectives.

Secondly, the main drivers of my research over the past several years have been the development and collection of language resources that would enable key technologies to support our languages, including our minority languages. This has been achieved in my own work for the Irish language at Dublin City University with the support of both national (Irish) funding and EU projects. Furthermore, over the past couple of years, I have been a core member of the European Language Equality project, where thorough assessment has been carried out to evaluate digital language (in)equality across EU languages and establish guidelines for ensuring necessary gaps are filled to this end by 2030. Along with other members of the ELE team, I presented the results of this project at a STOA Workshop at the EU parliament in November 2022, along with a summary of the EU Language Reports, including Irish.

As such, in alignment with the ethos and culture of ELRA, I strongly believe that ensuring the availability of necessary resources is vital to ensuring that today’s technologies can recognise our languages and provide the necessary support required to see them prosper as living languages in the digital age. Moreover, I believe that pan-European and pan-global efforts, along with a collective coordination is the most effective approach we can take as speakers of and advocates for our languages.

For many researchers, what makes ELRA so attractive as both an association and as organisers of LREC, is its strong ethos of diversity and inclusivity. LREC differs in many ways from other large speech and language technology conferences. These differences are reflected in the diversity of the languages and language resources that are usually represented at the conference and as such, the diversity of cultures that are usually represented. This diversity is enhanced both through the main conference and the variety of workshops held alongside the main conference. LREC provides a venue for researchers who otherwise would be overlooked at an empirical and engineering focused conference that usually requires large datasets and computing power resources. I believe this diversity needs to be fostered further to ensure that more researchers are encouraged to pursue research activities that will help to strengthen their language’s digital vibrancy.

As an advocate for making the Speech and Language Technology community a more welcoming place, I would like to support the broadening of LREC’s reach in this respect. I am eager to investigate and propose new ideas or strategies for supporting those working with under-resourced languages. Under-resourced not only in the sense of a lack of linguistic or technical resources, but also with respect to a lack of skilled experts/ researchers, lack of research funding and sometimes a lack of government support. I therefore propose close collaboration with other board members in ensuring that under-represented minorities in our field continue to be supported.

From a commercial perspective, it is worth noting that the Language AI data marketplace is growing in importance in Europe’s digital sphere. ELRA has long been a provider of language resources to ensure commercial entities can build the technologies required by their consumers and users.
One of the largest challenges in ensuring digital language equality in Europe lies in the lack of availability of suitable data upon which commercial technologies can be built - be they for industry use, consumer use or government use. While languages like English, French, Spanish and German are usually prioritised as they are realtively easily resourced, the rest of the EU languages represent a long tail of low technological support. This directly impacts the level of multilinguality that technical products or services can provide. By ensuring that the ELDA catalogue continues to have broad coverage of our languages, we are supporting the EU’s commercial landscape and goals for a digital single market.

https://www.computing.dcu.ie/~tlynn/
https://www.linkedin.com/in/teresa-lynn-41430817/