摘要 |
PROBLEM TO BE SOLVED: To provide a system for speech recognition using re-recognition and statistical classification.SOLUTION: In architecture that employs an overall grammar as a set of context-specific grammars for recognition of an input, each of the context-specific grammars corresponds to a specific context, such as a subtask category and a geographic region. The grammars together cover the entire domain. Moreover, multiple recognitions can be run in parallel against the same input, where each recognition path uses one or more of the context-specific grammars. Multiple intermediate recognition results from different recognizer-grammars are reconciled by running re-recognition using a dynamically composed grammar based on the multiple recognition results or other domain knowledge, or by selecting a winner using a statistical classifier operating based on classification features extracted from the multiple recognition results and other domain knowledge.SELECTED DRAWING: Figure 1 |