SUPR
Recognition of Child Language
Dnr:

NAISS 2024/22-51

Type:

NAISS Small Compute

Principal Investigator:

Peter Andersson Lilja

Affiliation:

Högskolan i Borås

Start Date:

2024-01-12

End Date:

2025-02-01

Primary Classification:

10208: Language Technology (Computational Linguistics)

Webpage:

Allocation

Abstract

In a larger project funded by the Swedish Research Council 2022-03500, we are investigating different aspects of children's storytelling. About 250 audio recordings of children telling (and retelling) stories are collected on three different occasions. In the transcription phase, in step 1, we want to use software/resources that can recognize children's language and make the manual transcription more efficient. Experiments with speech recognition of child language are relatively few and present a number of challenges. This project should therefore also be seen as part of developing methods for this purpose. The audio files do not contain sensitive personal data but only children telling a story from out six pictures.