Building a Swedish Open-Domain Conversational Language Model

Tobias Norlund, Agnes Stenbom


Abstract
We present on-going work of evaluating the, to our knowledge, first large generative language model trained to converse in Swedish, using data from the online discussion forum Flashback. We conduct a human evaluation pilot study that indicates the model is often able to respond to conversations in both a human-like and informative manner, on a diverse set of topics. While data from online forums can be useful to build conversational systems, we reflect on the negative consequences that incautious application might have, and the need for taking active measures to safeguard against them.
Anthology ID:
2021.nodalida-main.38
Volume:
Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa)
Month:
May 31--2 June
Year:
2021
Address:
Reykjavik, Iceland (Online)
Editors:
Simon Dobnik, Lilja Øvrelid
Venue:
NoDaLiDa
SIG:
Publisher:
Linköping University Electronic Press, Sweden
Note:
Pages:
357–366
Language:
URL:
https://aclanthology.org/2021.nodalida-main.38
DOI:
Bibkey:
Cite (ACL):
Tobias Norlund and Agnes Stenbom. 2021. Building a Swedish Open-Domain Conversational Language Model. In Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa), pages 357–366, Reykjavik, Iceland (Online). Linköping University Electronic Press, Sweden.
Cite (Informal):
Building a Swedish Open-Domain Conversational Language Model (Norlund & Stenbom, NoDaLiDa 2021)
Copy Citation:
PDF:
https://aclanthology.org/2021.nodalida-main.38.pdf