Fact · Accessibility and health
In a 2026 preprint on child-adult recordings, children with high ADOS severity scores had higher multi-talker word error rates than children with low scores, within each age group.
Checked against its source on October 16, 2026
In the source's own words
“Across both age groups, children with high CSS consistently exhibit higher mtWER, whereas mtWER remains comparable within each age group at the same severity level.”
End-to-End Joint ASR and Speaker Role Diarization with Child-Adult Interactions, Anfeng Xu et al., arXiv, 2026, Experimental setup, Table V discussion.
Open the original- What it measures
- CSS is the ADOS Calibrated Severity Score. The authors group children as low (1 to 3) or high (6 to 10) severity and read the result as severity being a main driver of recognition difficulty.
- What it does not show
- Preprint. The error rate is for multi-talker transcripts of child-adult sessions, using Whisper-based models, not single-speaker dictation by an adult.