Posted on Dec. 14, 2020
Last week, the 28th International Conference on Computational Linguistics took place online.
Various LT3 colleagues presented work at one of the workshops co-located with this conference:
- Gilles Jacobs and Veronique Hoste presented new work on economic event extraction at the FNP-FNS workshop. They created a dataset and system that automatically gets “what, who, when, how much, and why” is happening from economic news text.
- Orphée De Clercq and Veronique Hoste presented work on coreference resolution at the CRAC workshop. They explored what the added value is of adding coreferential information to the task of fine-grained sentiment analysis in both English and Dutch.
- Nina Bauwelinck and Els Lefever presented a pilot study at the Arg Mining workshop. They reported on annotation efforts related to annotating topics, stance, argumentativeness and claims in Dutch social media comments.
- Bram Vanroy, Sofie Labat, Olha Kaminska, Els Lefever and Veronique Hoste also participated in Task 7 of SemEval2020: "Assessing Humor in Edited News Headlines". They compared fine-tuning RoBERTa to a fully feature-engineered approach for humor assessment. This transformer-based approach finished 7th out of 49 submissions.
Have a look at the publications section to learn more about these papers.