SubmissionNumber#=%=#27
FinalPaperTitle#=%=#Bottlenecks of In-Context Learning for Fieldwork ASR: A Case-study of Panãra
ShortPaperTitle#=%=#
NumberOfPages#=%=#10
CopyrightSigned#=%=#Siyu Liang
JobTitle#==#
Organization#==#
Abstract#==#In-context learning (ICL) enables ASR models to transcribe unseen languages by conditioning on a handful of audio-transcript pairs at inference time, with no fine-tuning. This is appealing for language documentation, where transcribed data is scarce and recording conditions vary across sessions. We evaluate ICL on Panãra (Northern Jê, Brazil), a language with a complex practical orthography in which diacritics encode phonemic contrasts, across seven fieldwork recordings varying in speaker, narrative, and recording context. We find substantial within-language variation in transcription accuracy unexplained by any single recording-level factor, and show that diacritics are a systematic bottleneck with pronounced differences across diacritic types. An orthographic manipulation experiment further shows that how diacritics are represented in context transcriptions substantially affects model performance. These results highlight orthographic complexity and recording-level variation as key practical challenges for ICL-assisted fieldwork transcription.
Author{1}{Firstname}#=%=#Siyu
Author{1}{Lastname}#=%=#Liang
Author{1}{Username}#=%=#liangsy
Author{1}{Orcid}#=%=#
Author{1}{Email}#=%=#liangsy@uw.edu
Author{1}{Affiliation}#=%=#University of Washington
Author{2}{Firstname}#=%=#Myriam
Author{2}{Lastname}#=%=#Lapierre
Author{2}{Orcid}#=%=#
Author{2}{Email}#=%=#myriam.lapierre2@mcgill.ca
Author{2}{Affiliation}#=%=#McGill University
Author{3}{Firstname}#=%=#Gina-Anne
Author{3}{Lastname}#=%=#Levow
Author{3}{Username}#=%=#levow
Author{3}{Orcid}#=%=#
Author{3}{Email}#=%=#levow@uw.edu
Author{3}{Affiliation}#=%=#University of Washington

==========
èéáğö