Shared Task on Lexical Normalization for Swiss German
Lexical normalization describes the task of transforming word variations to a standard form. This usually involves contractions, abbreviations, and spelling variants. Written Swiss German is mainly used in informal communication on Social Media and text messaging. There is no standardized spelling and therefore there are many spellings of the same word depending on the author and their dialect. This can make it difficult to analyze Swiss German corpora.
In this shared task, participants are challenged to build systems that normalize Swiss German utterances word for word to their Standard German form. This is different from translation since the Swiss German word order will be preserved.
Workshop Important Dates
- Now: Training Data available
- May 16: Release Test Data
- May 20: System Outputs Due
- May 24: Publication of Evaluation Results
- June 03: System Descriptions Due
- June 08: Shared Task Session at SwissText 2022
- June 17: Review of System Descriptions
- July: Camera-ready system descriptions due
Workshop Schedule
Note: This is a tentative schedule
- Introduction to the shared task (10 minutes)
- Participants present their systems (15 minutes each)
- Outlook and Closing (10 minutes)
Organizers
- von Däniken Pius, ZHAW, vode@zhaw.ch
- Hürlimann Manuela, ZHAW, hueu@zhaw.ch
- Cieliebak Mark, ZHAW, ciel@zhaw.ch
