In language processing, deriving the true intention of a word that has multiple meanings. The word means "to remove ambiguities," and it is used in text-to-speech, speech recognition and text mining applications. For example, the word "r-e-a-d" is pronounced "reed" in the present tense but "red" in the past tense. A text-to-speech program would have to determine the tense of the word by the rest of the sentence.
Sophisticated disambiguation techniques may analyze the entire voice sample or document before deciding on the meaning. See
text-to-speech,
voice recognition and
text mining.