16th Annual Conference of the International Speech Communication Association

Dresden, Germany
September 6-10, 2015

Context-Dependent Error Correction of Spoken Referring Expressions

Ingrid Zukerman, Andisheh Partovi, Su Nam Kim

Monash University, Australia

We integrate a supervised machine learning mechanism for detecting erroneous words in the output of a speech recognizer with a two-tier error-correction approach that features (1) a noisy-channel model that replaces erroneous words with generic words, and (2) a phonetic-similarity mechanism that refines the generic words based on a short list of candidate interpretations. Our results, obtained on a corpus of 341 referring expressions, show that the first tier improves interpretation performance, and the second tier yields further improvements.

Full Paper

Bibliographic reference.  Zukerman, Ingrid / Partovi, Andisheh / Kim, Su Nam (2015): "Context-dependent error correction of spoken referring expressions", In INTERSPEECH-2015, 2032-2036.