Inferring informational goals and preferred level of detail of answers based on application being employed by the user

• US 7,409,335 B1
• Filed: 06/29/2001
• Issued: 08/05/2008
• Est. Priority Date: 06/29/2001
• Status: Expired due to Term

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Horvitz, Eric J., Zukerman, Ingrid
• Primary Examiner(s): SMITS, TALIVALDIS IVARS
• Application Number:
  US09/896,187
• Time in Patent Office: 2,594 Days
• Field of Search: 704/9, 704/7, 707/5
• US Class Current: 704/9
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A system and method for inferring informational goals and preferred level of details in answers in response to questions posed to computer-based information retrieval or question-answering systems is provided. The system includes a query subsystem that can receive an input query and extrinsic data associated with the query and which can output an answer to the query. The query subsystem accesses an inference model to retrieve conditional probabilities that certain informational goals are present. One application of the system includes determining a user'"'"'s likely informational goals and then accessing a knowledge data store to retrieve responsive information. Determining a user'"'"'s likely informational goals can include inferring a desired level of detail of answers to the query based on the application being employed by the user at the time the query is submitted. The system includes a natural language processor that parses queries into observable linguistic features and embedded semantic components that can be employed to retrieve the conditional probabilities from the inference model. The inference model is built by employing supervised learning and statistical analysis on a set of queries suitable to be presented to a question-answering system. Such a set of queries can be manipulated to produce different inference models based on demographic and/or localized linguistic data.

Claims

1. A system for inferring an information goal, comprising:

a query subsystem that receives at least one of a query and an extrinsic data, the query subsystem is operatively coupled to an inference model and a knowledge data store, the query subsystem comprising;

a natural language processor that parses the query; and

an inference engine that infers one or more informational goals based, at least in part, on at least one of the parsed query, the extrinsic data and an inference data stored in the inference model, the inference engine further inferring one or more preferred levels of detail for an answer based on an application being employed by the user.

2. The system of claim 1, the informational goals include at least one of, a type of information requested in the query, a topic of the query, or a focal point of the query.

3. The system of claim 2 comprising:

an input query log that stores at least one of, one or more queries or one or more pieces of extrinsic data; and

a learning system operatively coupled to the input query log, the learning system operable to produce the inference model.

4. The system of claim 3, where the learning system comprises:

the natural language processor further produces linguistic data concerning one or more linguistic features;

a tagging tool that facilitates manipulating the linguistic data;

one or more taggers that manipulates the linguistic data; and

wherein the inference model stores information concerning conditional probabilities associated with the likelihood that one or more informational goals exist, where the conditional probabilities of the informational goals are determined, at least in part, from Bayesian statistical analysis performed on the linguistic data.

5. The system of claim 4, the linguistic data comprises a parse tree, where the parse tree contains extractable information concerning the nature of and relationships between observable linguistic features.

6. The system of claim 5, the observable linguistic features in the extractable information comprise word-based features, structural features and hybrid linguistic features.

7. The system of claim 6, the word-based features indicate the presence of one or more candidate terms that can be employed in predicting an informational goal.

8. The system of claim 4, the taggers manipulate the linguistic data to conform with one or more schemas associated with reasoning concerning the relevance of a part of a query based on one or more language models.

9. The system of claim 8, the taggers supervise learning associated with computing probabilities associated with the informational goals.

10. The system of claim 4, the inference model represents a probabilistic dependency model.

11. The system of claim 4, the inference model comprises one or more decision trees, the decision trees store conditional probabilities associated with one or more informational goals, the decision trees being traversable by the linguistic data.

12. The system of claim 3, the input query log is at least one of a data store or a manual store.

13. The system of claim 3, the natural language processor parses a query into one or more parts suitable for retrieving one or more conditional probabilities stored in the inference model.

14. The system of claim 13, the one or more parts comprise at least one of, logical forms, adjectival phrases, adverbial phrases, noun phrases, verb phrases, prepositional phrases or parse trees.

15. The system of claim 14, the inference engine further infers one or more informational goals based, at least in part, on at least one of the query, the extrinsic data, the one or more parts, or the one or more conditional probabilities stored in the inference model.

16. The system of claim 3, the query subsystem further comprising:

an answer generator that produces a response to the query and produces an error message.

17. The system of claim 1, the knowledge data store is searchable for information responsive to a new query and where the information retrieved will depend, at least in part, on the inferred informational goals.

18. The system of claim 1, the query subsystem is compiled into an executable, and where the executable accepts as input one or more query distinctions.

19. A computer readable medium storing computer executable components of a system for inferring an information goal, the system comprising:

a query component that receives a new query and a new extrinsic data, the query component operatively coupled to an inference model and a knowledge data store, the query component comprising;

a natural language processing component that parses the new query; and

an inference component that infers one or more informational goals based, at least in part, on at least one of, the new query, the new extrinsic data and an inference data stored in the inference model, the inference engine further inferring one or more preferred levels of detail for an answer based on an application being employed by the user.

20. A computer readable medium storing computer executable components of a system for learning how to infer information goals from queries, the system comprising:

a natural language processing component that produces a linguistic data concerning one or more linguistic features;

a tagging component that manipulates the linguistic data;

one or more taggers that manipulates the linguistic data; and

an inference model component that stores information concerning conditional probabilities associated with the likelihood that one or more informational goals exist, where the conditional probabilities of the informational goals are determined, at least in part, from Bayesian statistical analysis performed on the linguistic data, the inference engine further inferring one or more preferred levels of detail for an answer based on an application being employed by the user.