Frequently asked questions

Does the shared task involve transcribing sign language, using, for instance HamNoSys notation?

No, there will be no transcription of sign language. Creating such annotations is enormously laborious and clearly beyond the scope of this shared task, in terms of both time and cost.

In the largest corpus acquisition project to date, the DGS Corpus Project, the ratio between transcribed time and transcription time was shown to be one to several hundred, i.e., one minute of signing takes several hundred minutes to transcribe.

Can systems for other language pairs be submitted to this shared task?

No, systems for language pairs that are not covered in this shared task  (other than DSGS, LIS and LSF) cannot be submitted. We hope to cover more language pairs in future iterations of the shared task.