×

Two-pass decoding for speech recognition of search and action requests

  • US 8,645,138 B1
  • Filed: 12/20/2012
  • Issued: 02/04/2014
  • Est. Priority Date: 12/20/2012
  • Status: Active Grant
First Claim
Patent Images

1. A method, comprising:

  • receiving input speech at a computing system;

    during a first pass of speech recognition, determining a plurality of outputs from a plurality of language models by;

    providing the input speech as an input to each of a plurality of language models, wherein the plurality of language models comprises a query language model and an action language model, andreceiving an output from each language model;

    determining a selected language model of the plurality of language models using a classifier operating on the plurality of outputs from the plurality of language models, wherein the classifier is configured to utilize a support vector machine (SVM) to select the selected language model, and wherein the SVM is configured to determine a plane or hyperplane related to the plurality of language-model outputs and to select the selected language model based on the plane or hyperplane;

    during a second pass of speech recognition, determining a revised output by;

    providing the input speech and the output from the selected language model as inputs to the selected language model, andreceiving the revised output from the selected language model; and

    generating a result based on the revised output using the computing system.

View all claims
  • 2 Assignments
Timeline View
Assignment View
    ×
    ×