Can I get uhh a better WER: Challenges and Opportunities in Evaluating Conversational Speech Recognition

Simple item page
dc.contributor.advisorDingemanse, M.
dc.contributor.advisorLiesenfeld, A.M.
dc.contributor.authorLopez, Ada
dc.date.issued2023-08-24
dc.description.abstractConversational speech recognition stands as a pivotal area in language technology, yet it still remains a significant challenge in the field despite technological advancements. In this thesis, I argue that the only way to solve this is through representing the foundations of human interaction. In this research, I look into the interactional infrastructure and resources employed in spontaneous conversations and discuss how these are represented – or neglected – in Automatic Speech Recognition (ASR). An analysis on the differences between human and ASR transcriptions shows that current state-of-the-art systems fail to accurately reflect certain essential and characteristic features of conversations: turn-taking, overlaps, and conversational words. The results of this study points towards a necessary paradigm shift, illustrating the importance of using interaction linguistics to inform both conversational ASR system development and evaluation. To address some of these limitations, a new composite metric is proposed to augment the conventional Word Error Rate (WER).
dc.identifier.urihttps://theses.ubn.ru.nl/handle/123456789/16376
dc.language.isoen
dc.thesis.facultyFaculteit der Letteren
dc.thesis.specialisationspecialisations::Faculteit der Letteren::Researchmasters::Researchmaster Language and Communication
dc.thesis.studyprogrammestudyprogrammes::Faculteit der Letteren::Researchmasters
dc.thesis.typeResearchmaster
dc.titleCan I get uhh a better WER: Challenges and Opportunities in Evaluating Conversational Speech Recognition
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
Master thesis ReMa LCS Lopez, Ada.pdf
Size:
1.41 MB
Format:
Adobe Portable Document Format
Download
Collections
Faculteit der Letteren