Speech re-recognition using external data sources

摘要

Methods comprising computer programs encoded on a computer storage medium for improving speech recognition based on external data sources. In one aspect, the method comprises the steps of obtaining an initial candidate notation of a utterance using an automated speech recognizer and based on a language model not used by the automated speech recognizer in generating the initial candidate notation, the initial And identifying one or more terms that are phonetically similar to one or more terms that occur in the candidate notation. Additional actions include generating one or more additional candidate notations based on the identified one or more terms and selecting a notation among the candidate notations. The described configurations can allow data from an external data source to be used to generate more accurate notations without modifying the existing automated speech recognizer, or avoid re-compiling the automated speech recognizer.