Under the auspices of DARIAH-EU Working Group Text and Data Analytics an expert colloquium was held at Dublin City University’s Adapt Centre, 14th – 15th of December 2015, on the subject of topic models and corpus-analytic approaches in the humanities, with a special focus on literary studies and the philologies.
In the light of increasingly available digital text ressources and suitable quantitative methodologies – which are increasingly able to augment or even reframe research questions traditionally seen as exclusively qualitative – such approaches have found their way into a variety of humanities desciplines and demand a closer look at their domain-specific adaptions, technically as well as epistemologically. As such the workshop was addressed to experienced users, researchers, and developers working with corpus-analytic approaches, especially those geared towards the automatic analysis of semantic content. Contributions ranged from specific experience reports to opinion pieces discussing the broader implications of said approaches, and while the workshop was set up around a group of experts from TDA working group, it was open to the public and well attended by master students and doctoral researchers working in the field.
In the course of the two-day programme, a wide variety of corpus-analytic approaches were discussed – ranging from topic models and other established methods of ‚distant reading‘ to approaches using novel feature combinations. It was a variety of contributions, which were furthermore framed by a unique perspective from translation studies, as presented by Sharon O’Brien. Her keynote pointed at various critical points during the adoption of computer-aided methodologies in translation studies – a discipline that came into contact with language technology early on – showing that machine translation is eventually able to adapt the job to the person, rather than forcing the user to work for the machine. It is a development that led to human-in-the-loop approaches – where the initial machine output acts as a springboard for the actual translation, harboring extensive savings of time – and to the development of an integrated, cognitive perspective on machine translation.
[...]
Quelle: http://dhd-blog.org/?p=6472