Language Complexity and Speech Recognition Accuracy: Orthographic Complexity Hurts, Phonological Complexity Doesn't

Chihiro Taguchi, David Chiang
6/13/2024
cs.CLcs.AI

No official code found. Best implementation: ctaguchi/ASRcomplexity (3 stars, last commit 2024-05-30).

Abstract

We investigate what linguistic factors affect the performance of Automatic Speech Recognition (ASR) models. We hypothesize that orthographic and phonological complexities both degrade accuracy. To examine this, we fine-tune the multilingual self-supervised pretrained model Wav2Vec2-XLSR-53 on 25 languages with 15 writing systems, and we compare their ASR accuracy, number of graphemes, unigram grapheme entropy, logographicity (how much word/morpheme-level information is encoded in the writing system), and number of phonemes. The results demonstrate that orthographic complexities significantly correlate with low ASR accuracy, while phonological complexity shows no significant correlation.

View on arXivView PDF

Code Implementations(1)

ctaguchi/ASRcomplexityCommunity100%
30PythonMar 6, 20242 years ago

Similar Papers

Investigating Transcription Normalization in the Faetar ASR Benchmark

82% match2025

Quantifying the Role of Textual Predictability in Automatic Speech Recognition

81% match2024

Correlation Does Not Imply Compensation: Complexity and Irregularity in the Lexicon

81% match2024

Linguistically Informed Tokenization Improves ASR for Underresourced Languages

80% match2025

What is lost in Normalization? Exploring Pitfalls in Multilingual ASR Model Evaluations

80% match2024

Advocating Character Error Rate for Multilingual ASR Evaluation

80% match2024

Cite this paper

@article{taguchi2024language,
  title  = {Language Complexity and Speech Recognition Accuracy: Orthographic Complexity Hurts, Phonological Complexity Doesn't},
  author = {Chihiro Taguchi and David Chiang},
  year   = {2024},
  eprint = {2406.09202},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2406.09202v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2406.09202.svg)](https://codeofpaper.com/paper/2406.09202)

Discussion