AUTOMATED SCHEME FOR IDENTIFYING USER INTENT IN REAL-TIME

• US 20080104037A1
• Filed: 12/18/2007
• Published: 05/01/2008
• Est. Priority Date: 04/07/2004
• Status: Active Grant

• Current Assignee: Oracle OTC Subsidiary LLC (Oracle Corporation)
• Original Assignee: Inquira Incorporated (Oracle Corporation)
• Inventors: Bierner, Gann
• Granted Patent: US 8,082,264 B2
• Time in Patent Office: Days
• US Class Current: 707/3
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence

Abstract

An intent guessing system receives partial user queries as they are entered by the user. The partial user queries are compared with different intents derived from previously logged queries. Guesses are made as to which of the intents are associated with the partial user query. The intent guesses are then provided as responses to the user query. Features are identified for the earlier logged queries and associated with the derived intents. The derived intents and associated features are then used to identify intents for the partial user queries.

Claims

1. A method, comprising:

receiving a user query;

comparing the user query with different intents derived from previously logged queries;

guessing which of the intents are associated with the user query; and

responding to the user query with the intent guesses.

2. The method according to claim 1 including:

extracting features from the user query;

comparing the features from the user query with features from the previously logged queries; and

selecting the intents that have associated features most similar to the features from the user query.

3. The method according to claim 2 wherein the features comprise query tokens, query token stems, and concepts.

4. The method according to claim 1 including:

receiving a first portion of the user query;

guessing the entire intent of the user query according to the first portion of the user query while a remainder of the user query is being received; and

providing the intent guesses for the first portion of the user query prior to receiving an entire user query.

5. The method according to claim 1 further comprising:

identifying one or more parameterized intent guesses that include at least one concept;

identifying words in the user query that are associated with the concept;

modifying the parameterized intent guesses by replacing the concept with the identified words from the user query; and

responding to the user query with the modified parameterized intent guesses.

6. The method according to claim 1 further comprising:

identifying responses associated with the intent guesses;

either displaying the responses along with the associated intent guesses or displaying the responses instead of the associated intent guesses.

7. The method according to claim 6 wherein the responses include links to web pages or additional information associated with the intent guesses.

8. The method according to claim 1 further comprising:

logging a corpus of earlier submitted queries;

deriving intents associated with the corpus queries;

identifying features in the corpus queries;

associating the features with the derived intents; and

using the derived intents and associated features to make intent guesses for the user query.

9. The method according to claim 1 further comprising:

identifying one or more intent guesses associated with the user query;

identifying an order for the intent guesses according to how well the intent guesses match the user query; and

displaying the intent guesses in the identified order.

10. A computer readable medium having instructions stored thereon that, if executed by a computing platform, are adapted to:

receive a user query;

submit the user query to a context generator to identify features associated with the user query;

submit the features to a model;

receive one or more intents from the model associated with the features; and

respond to the user query with the one or more intents.

11. The computer readable medium according to claim 10 further adapted to train the model according to a query corpus of previously logged queries and a set of intents derived from the query corpus.

12. The computer readable medium according to claim 11 further adapted to:

submit the queries from the query corpus to a context generator to identify associated features; and

associate the identified features with the derived intents to create the model.

13. The computer readable medium according to claim 12 further adapted to use a language analyzer along with the context generator to identify the features in the query corpus.

14. The computer readable medium according to claim 10 further adapted to:

identify a first subset of words from the user query while the user query is being entered;

make one or more guesses as to an intent of a completed user query based on the first subset of words; and

replying to the first subset of words with the one or more intent guesses while the user query is still being entered.

15. The computer readable medium according to claim 10 further adapted to:

identify at least one intent that includes at least one concept;

identify words in the user query that are associated with the identified concept;

modify the identified intent by replacing the concept with the identified words in the user query; and

responding to the user query with the modified intent.

16. The computer readable medium according to claim 10 further adapted to:

identify additional information associated with the intents;

either display the identified additional information with the intents or display the additional information instead of the intents.

17. The computer readable medium according to claim 10 further adapted to:

identify confidence factors for the intents; and

display the intents according to the identified confidence factors.

18. The computer readable medium according to clam 10 wherein the intents include associated probability values that indicate a probability that the intents correctly respond to the user query and the intents are sent or not sent in response to the user query according to the associated probability values.

19. A system for identifying query intents, comprising:

an intent guesser configured to receive user queries from a query system;

a context generator configured to receive the user queries from the intent guesser and return back features of the user queries that comprise query tokens, query token stems, and concepts; and

an intent model configured to receive the features from the intent guesser and return back intents associated with the received features.

20. The system according to claim 19 further comprising:

a query corpus of previously logged queries;

an intent identifier configured to identify intents for the query corpus;

a model trainer configured to submit the queries from the query corpus to the context generator for identify features in the query corpus and create the intent model by associating the identified features from the context generator with the intents from the intent identifier.

21. The system according to claim 20 wherein at least some of the intents output from the intent model include concepts that cause the intent guesser to replace the concepts with words from the user query.