Method for variable resolution and error control in spoken language understanding

• US 20080243505A1
• Filed: 03/12/2008
• Published: 10/02/2008
• Est. Priority Date: 03/14/2007
• Status: Active Grant

• Current Assignee: Polytechnic University (New York University), Verint Americas Incorporated (Verint Systems Incorporated)
• Original Assignee: Polytechnic University (New York University)
• Inventors: Pieraccini, Roberto, Levon, Kalle, Dayanidhi, Krishna, Barinov, Victor, Dabrowski, Robert
• Granted Patent: US 7,962,339 B2
• Time in Patent Office: Days
• US Class Current: 704/251
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue

Abstract

A method for variable resolution and error control in spoken language understanding (SLU) allows arranging the categories of the SLU into a hierarchy of different levels of specificity. The pre-determined hierarchy is used to identify different types of errors such as high-cost errors and low-cost errors and trade, if necessary, high cost errors for low cost errors.

Claims

1. A method of resolving errors in an automated spoken language recognition system comprising the steps of:

providing a hierarchical set of predetermined spoken language categories arranged in respective paths of progressively increasing specificity, each path associated with a separate category;

receiving voice utterances from a speaker;

recognizing a received voice utterance and classifying the recognized voice utterance in one of said predetermined spoken language categories;

determining, when the speaker identifies an error between his voice utterance and said classified category, whether said error is a high-cost-error or a low-cost-error; and

instructing the speaker to select a category of lesser specificity and superior to said classified category in said hierarchical path if said error is determined to be a low-cost-error.

2. The method of claim 1, wherein a low-cost-error is an error between a superior category and an inferior category in the same hierarchical path.

3. A method of resolving errors in an automated spoken language recognition system comprising the steps of:

providing a hierarchical set of predetermined spoken language categories arranged in respective paths of progressively increasing specificity, each path associated with a separate category;

receiving voice utterances from a speaker;

recognizing a received voice utterance and preliminarily assigning the recognized voice utterance to one of said predetermined spoken language categories;

assigning a confidence level to the recognition of said received voice utterance;

sensing when said confidence level is below a predetermined threshold;

classifying the recognized voice utterance in a spoken language category of immediately lesser specificity than the preliminarily assigned category; and

instructing the speaker to select a category of lesser specificity and superior to said classified category in said hierarchical set.