🧪 Test?View on arXiv
What Counts as a Mistake? Annotating Recitation Events in Quran Memorization Transcripts
Not provided in the content
annotationASRerror detectionmachine learning
2609.12085
Builder Relevance
2h ago60%
Abstract
The paper discusses the challenges of distinguishing various types of mistakes in Quran recitation transcripts and presents a human annotation study with an executable evaluator.
Reality Card
Core Claim
The study achieved a label-aware F1 score of 0.525 and localization F1 score of 0.826, demonstrating the effectiveness of the annotation process for identifying mistakes in Quran recitation.
Method / Result
The best-performing model achieved a label-aware F1 score ranging from 0.143 to 0.892 across different runs.
Limitations
The main limitation is that no run was annotated before building, which affects the reproducibility of the results.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.