Intelligent query engine for processing voice based queries

A natural language processing system adapted for assisting recognition of a speech utterance, the system comprising:

• a text parsing routine for receiving text recognized from the speech utterance by a continuous speech recognition routine, and for generating a first linguistic analysis of said text, including an identification of any word phrases present in said text;
• wherein said first linguistic analysis is used to identify a candidate set of potential matches for said speech utterance, and
• said text parsing routine being further configured for:
• (i) generating one or more second linguistic analyses of text contained in said candidate set of potential matches;
• and
• (ii) comparing said first linguistic analysis with each of said one or more second linguistic analyses to identify a best match for said speech utterance in real time.

A method of recognizing a speech query comprising the steps of:

• (a) recognizing text in an articulated speech utterance;
• and
• (b) linguistically processing said recognized text to generate search terms based on phrase analysis of said articulate speech utterance;
• (c) generating a preliminary query to identify a potential match for said speech utterance, said preliminary query being based on said recognized text;
• (d) generating a final query to identify a potential match for said speech utterance, said final query being based on said recognized text and said search terms;
• (e) determining a final match for said speech utterance by linguistically comparing any potential matches identified by said preliminary and/or final query with said articulated speech utterance;
• wherein said preliminary query is made based on said recognized text to determine a final match during a time when linguistic processing is still being done on said recognized text.

A distributed method of recognizing and responding in real time to a speech utterance that corresponds to a user question posed within an application program executing on a client device, the method comprising the steps of:

• (a) receiving continuous speech data generated by the client device over a network;
• (b) processing said speech data at a server device to recognize words and form a recognized text associated with the user question;
• wherein during step (b) a context experienced by the user within the application program is considered for loading appropriate dictionaries and grammars to facilitate recognizing said words;
• (c) submitting said recognized text to a database query engine and a natural language engine;
• (d) linguistically processing said recognized text with said natural language engine during a first linguistic process to identify word phrases contained therein, including noun phrases;
• (e) forming a query to a question/answer pair database based on said recognized text and said word phrases;
• wherein said query includes a preliminary query based on said recognized text, and a final query based on said recognized text and said word phrases;
• (f) locating a set of question/answer pairs based on said query, including an initial set of question/answer pairs based on said preliminary query;
• (g) evaluating said set of question/answer pairs, using a second linguistic analysis to compare word phrases in said set of question/answer pairs with word phrases identified in said recognized text;
• (h) identifying a single question/answer pair best matching the user question based on step (g);
• (i) providing an answer determined from said single question/answer pair in step (h) to the user over the network;
• wherein an answer is provided to the user question in real-time over a distributed speech processing system.