From 5th – 8th March 2024, the eSTÓR team travelled to Saarbrucken, Germany to attend the LT Bridge Winter School held in Saarland University in the German Research Centre for Artificial Intelligence (DFKI). The eSTÓR project is a research project based in ADAPT DCU which collects English-Irish language data to improve Irish Language Technology.
Day one of the conference explored the evolution of Conversational AI, discussing the evolution from Eliza to ChatGPT. In the afternoon, the eSTÓR team presented their poster submission, which can be seen below, entitled, ‘Unlocking Irish language data for the improvement of Irish technology tools.’
Day two involved discussions on strategies for handling low-resource natural language processing (NLP) scenarios. This lecture was particularly relevant for the eSTÓR team in terms of application to the Irish language. An overview of machine translation (MT), covering various modalities of language translation provided the eSTÓR team with an applied insight into the value of machine translation.
On day three, cross-lingual transfer learning was discussed, focusing on the challenges faced when applying pre-trained models to low-resource languages. Members of the team connected with Maltese colleagues, pin-pointing similarities in our work in processing data and creating resources for a low-resource language, which will possibly lead to further collaboration.
In the afternoon of day three, attendees were taken on a tour of the world’s only fully preserved ironworks from the heyday of industrialisation, and is also the first industrial monument of the same era to be added to the UNESCO World Heritage List.
