Projects / Programmes
Computer-assisted multilingual news discourse analysis with contextual embeddings
| Code |
Science |
Field |
Subfield |
| 6.05.02 |
Humanities |
Linguistics |
Theoretical and applied linguistics |
| Code |
Science |
Field |
| 6.02 |
Humanities |
Languages and Literature |
Natural language processing, critical discourse analysis, news analysis, word embeddings, deep learning, diachronic news analysis, comparative news analysis, metaphor
Researchers (23)
Organisations (3)
Abstract
The ability to analyse and understand the news media has never been more important than now. Mass online content surrounds us, with misinformation and bias rife, leading to a wide distrust in broadcast media, and increasing reliance on unattested sources and social media. Qualitative discourse analysis techniques are designed precisely to uncover and understand the biases, viewpoints and framing techniques involved; but are limited by their reliance on complex and time-consuming manual analysis. At the same time, computational natural language processing techniques have been developed which can accurately process and categorise text in a range of ways; but are rarely applied to critical discourse analysis tasks or in less-resourced languages such as Slovene. This project will bridge the gap between these two, by adapting recent NLP tools to Slovene, developing new methods to help identify important discourse analysis phenomena, support multilingual news analysis, and work with discourse analysts to produce tools and visualisation methods that they can use to augment and improve their work.
