How Good is Automatic Segmentation as a Multimodal Discourse Annotation Aid?

Corbyn Terpstra, Ibrahim Khebour, Mariah Bradford, Brett Wisniewski, Nikhil Krishnaswamy, Nathaniel Blanchard

Abstract
In this work, we assess the quality of different utterance segmentation techniques as an aid in annotating collaborative problem solving in teams and the creation of shared meaning between participants in a situated, collaborative task. We manually transcribe utterances in a dataset of triads collaboratively solving a problem involving dialogue and physical object manipulation, annotate collaborative moves according to these gold-standard transcripts, and then apply these annotations to utterances that have been automatically segmented using toolkits from Google and Open-AI’s Whisper. We show that the oracle utterances have minimal correspondence to automatically segmented speech, and that automatically segmented speech using different segmentation methods is also inconsistent. We also show that annotating automatically segmented speech has distinct implications compared with annotating oracle utterances — since most annotation schemes are designed for oracle cases, when annotating automatically-segmented utterances, annotators must make arbitrary judgements which other annotators may not replicate. We conclude with a discussion of how future annotation specs can account for these needs.
Anthology ID:
2023.isa-1.10
Volume:
Proceedings of the 19th Joint ACL-ISO Workshop on Interoperable Semantics (ISA-19)
Month:
June
Year:
2023
Address:
Nancy, France
Editor:
Harry Bunt
Venues:
ISA | WS
SIG:
SIGSEM
Publisher:
Association for Computational Linguistics
Note:
Pages:
75–81
Language:
URL:
https://aclanthology.org/2023.isa-1.10
DOI:
Bibkey:
Cite (ACL):
Corbyn Terpstra, Ibrahim Khebour, Mariah Bradford, Brett Wisniewski, Nikhil Krishnaswamy, and Nathaniel Blanchard. 2023. How Good is Automatic Segmentation as a Multimodal Discourse Annotation Aid?. In Proceedings of the 19th Joint ACL-ISO Workshop on Interoperable Semantics (ISA-19), pages 75–81, Nancy, France. Association for Computational Linguistics.
Cite (Informal):
How Good is Automatic Segmentation as a Multimodal Discourse Annotation Aid? (Terpstra et al., ISA-WS 2023)
Copy Citation:
PDF:
https://preview.aclanthology.org/add_acl24_videos/2023.isa-1.10.pdf
PDF Search