Applying PC and FCI causal discovery to children's ASR shows that gender does not cause word errors, age mainly causes substitution errors, and fine-tuning does not reduce the impact of utterance length.
Causal Structure Discovery for Error Diagnostics of Children's ASR
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
Children's automatic speech recognition (ASR) often underperforms compared to that of adults due to a confluence of interdependent factors: physiological (e.g., smaller vocal tracts), cognitive (e.g., underdeveloped pronunciation), and extrinsic (e.g., vocabulary limitations, background noise). Existing analysis methods examine the impact of these factors in isolation, neglecting interdependencies-such as age affecting ASR accuracy both directly and indirectly via pronunciation skills. In this paper, we introduce a causal structure discovery to unravel these interdependent relationships among physiology, cognition, extrinsic factors, and ASR errors. Then, we employ causal quantification to measure each factor's impact on children's ASR. We extend the analysis to fine-tuned models to identify which factors are mitigated by fine-tuning and which remain largely unaffected. Experiments on Whisper and Wav2Vec2.0 demonstrate the generalizability of our findings across different ASR systems.
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Causal Structure Discovery for Error Diagnostics of Children's ASR
Applying PC and FCI causal discovery to children's ASR shows that gender does not cause word errors, age mainly causes substitution errors, and fine-tuning does not reduce the impact of utterance length.