13th Annual Conference of the International Speech Communication Association

Portland, OR, USA
September 9-13, 2012

Analyzing and Interpreting Automatically Learned Rules Across Dialects

Nancy F. Chen (1,2), Wade Shen (1), Joseph P. Campbell (1)

(1) MIT/Lincoln Laboratory, Lexington, MA, USA
(2) Institute for Infocomm Research, Singapore, Singapore

In this paper, we demonstrate how informative dialect recognition systems such as acoustic pronunciation model (APM) can help speech scientists locate and analyze phonetic rules efficiently. In particular, we analyze dialect-specific characteristics automatically learned from APM across two American English dialects. We show that unsupervised rule retrieval performs similarly to supervised retrieval, indicating that APM is useful in practical scenarios, where word transcriptions are often unavailable. We also demonstrate that the top-ranking rules learned from APM generally correspond to the linguistic literature, and can even pinpoint potential research directions to refine existing linguistic knowledge. The APM system can help phoneticians analyze rules efficiently by characterizing large amounts of data to postulate rule candidates, so phoneticians can save time and manual effort to conduct more targeted investigations.Potential applications of informative dialect recognition systems include forensic phonetics and diagnosis of spoken language disorders.

Index Terms: informative dialect recognition, rule retrieval, phonological rules, forensic phonetics

Full Paper

Bibliographic reference.  Chen, Nancy F. / Shen, Wade / Campbell, Joseph P. (2012): "Analyzing and interpreting automatically learned rules across dialects", In INTERSPEECH-2012, 1307-1310.