On the 11th of July 2019, we held the second meeting of the Open Dutch FrameNet project. In attendance were Piek Vossen, Johan Bos, Antske Fokkens, Filip Ilievski, Tommaso Caselli, Malvina Nissim, and Marten Postma.
The meeting was held at the Rijksuniversiteit Groningen. The core item of the meeting was the pilot data, which was created before the meeting and processed by both research groups. Filip Ilievski was the first presenter and provided information about how we selected and represented the pilot data.
Marten Postma was next and talked about the type of linguistic processing that we ran on the pilot data. We ran open-sesame, a FrameNet-based SRL system, and determined the most dominant frames in each event type discourse, e.g., all incidents discussing elections.
Also, Marten Postma talked about the first version of Dutch FrameNet, for which we exploited translation relationships between monosemous English and Dutch lemmas.
Finally, Johan Bos and Tommaso Caselli talked about how they inserted the pilot data into the Parallel Meaning Bank and what type of linguistic processing was performed on the data.