Empirical Properties of Multilingual Phone-to-Word Transduction

  • Geoffrey Zweig

In Proceedings of ICASSP |

This paper explores the error-robustness of phone-to-word transduction across a variety of languages. We implement a noisy channel model in which a phonetic input stream is corrupted by an error model, and then transduced back to words using the inverse error model and linguistic constraints. By controlling the error level, we are able to measure the sensitivity of different languages to degradation in the phonetic input stream. This analysis is carried further to measure the importance of each phone in each language individually. We study Arabic, Chinese, English, German and Spanish, and find that they behave similarly in this paradigm: in each case, a phone error produces about 1.4 word errors, and frequently incorrect phones matter slightly less than others. In the absence of phone errors, transduced word errors are still present, and we use the conditional entropy of words given phones to explain the observed behavior.