Fact-based indexing for natural language search

• US 8,639,708 B2
• Filed: 08/29/2008
• Issued: 01/28/2014
• Est. Priority Date: 08/31/2007
• Status: Active Grant

• Current Assignee: Zhigu Holdings Limited
• Original Assignee: Microsoft Corporation
• Inventors: Van Den Berg, Martin Henk, Bobrow, Daniel, Cheslow, Robert D., Pell, Barney, Thione, Giovanni Lorenzo, Walters, Chad
• Primary Examiner(s): LIN, SHEW FEN
• Application Number:
  US12/201,596
• Publication Number:
  US 20090063550A1
• Time in Patent Office: 1,978 Days
• Field of Search: 707/739, 707/741, 707/755, 707/756, 704/4, 704/9
• US Class Current: 707/755
• CPC Class Codes: G06F 16/3344 using natural language analysis
• CPC Class Codes: G06F 16/338 Presentation of query results

Abstract

Computer-readable media and a computer system for implementing a natural language search using fact-based structures and for generating such fact-based structures are provided. A fact-based structure is generated using a semantic structure, which represents information, such as text, from a document, such as a web page. Typically, a natural language parser is used to create a semantic structure of the information, and the parser identifies terms, as well as the relationship between the terms. A fact-based structure of a semantic structure allows for a linear structure of these terms and their relationships to be created, while also maintaining identifiers of the terms to convey the dependency of one fact-based structure on another fact-based structure. Additionally, synonyms and hypernyms are identified while generating the fact-based structure to improve the accuracy of the overall search.

Claims

1. One or more computer-storage media having computer-executable instructions embodied thereon that, when executed, perform a method for generating a semantic index that employs a fact-based structure for a search, the method comprising:

generating one or more semantic relationships from textual content within online documents, wherein the process of generating comprises;

(a) identifying a sentence within from the textual content of the online documents;

(b) distilling a syntactic structure of the sentence;

(c) separating the sentence into two or more logical assertions based on, in part, a grammatical organization of the syntactic structure;

(d) analyzing the two or more logical assertions to determine a linguistic functionality of one or more terms within the two or more logical assertions; and

(e) deriving the one or more semantic relationships from the linguistic functionality of the one or more terms; and

representing the one or more semantic relationships as two or more fact-based structures, respectively, via the process comprising;

(a) representing the one or more terms as nodes;

(b) representing the linguistic functionality of one or more terms as edges that interconnect the nodes, respectively, wherein each edge involves a direction of dependency that targets a single node;

(c) identifying from the nodes one or more head nodes, wherein the one or more head nodes are nodes being targeted by two or more of the edges;

(d) generating the two or more fact-based structures from the one or more head nodes, respectively, wherein a fact-based structure is directed toward one of the one or more head nodes, respectively; and

(e) storing the two or more fact-based structures in the semantic index.

2. The media of claim 1, wherein the one or more syntactic structures between the one or more terms from the textual content reveals the one or more semantic relationships between the one or more terms.

3. The media of claim 1, wherein the nodes include expansions of the terms represented thereby, wherein the expansions include synonyms of the terms.

4. The media of claim 1, wherein the textual content within the online documents that is associated with the two or more fact-based structures is accessed when a proposition of a search query matches at least one of the two or more fact-based structures.

5. One or more computer-storage media having computer-executable instructions embodied thereon that, when executed, perform a method for implementing an index system that employs a fact-based structure for a search, the method comprising:

generating one or more semantic relationships from textual content within online documents, wherein the process of generating comprises;

(a) identifying a sentence within the textual content of the online documents;

(b) distilling a syntactic structure of the sentence

(c) separating the sentence into two or more logical assertions based on, in part, a grammatical organization of the syntactic structure

(d) analyzing the two or more logical assertions to determine a linguistic functionality of one or more terms within the two or more logical assertions; and

(e) deriving the one or more semantic relationships from the linguistic functionality of the one or more terms; and

representing the one or more semantic relationships as two or more fact-based structures, respectively, via the process comprising;

(a) representing the one or more terms as nodes;

(b) representing the linguistic functionality of one or more terms as edges that interconnect the nodes, respectively, wherein each edge targets a single node;

(c) identifying from the nodes one or more head nodes, wherein the one or more head nodes are nodes being targeted by two or more of the edges;

(d) generating the two or more fact-based structures from the one or more head nodes, respectively, wherein a fact-based structure is directed toward one of the one or more head nodes, respectively; and

(e) storing the two or more fact-based structures in the index system.

6. The media of claim 5, wherein, when a first term and a second term of the terms within different fact-based structures are assigned to a same identifier when the first term and the second term relate to a same entity.

7. The media of claim 5, wherein the online documents include web pages.

8. The media of claim 5, wherein the method further comprises identifying matches between a search query and the two or more fact-based structures.

9. The media of claim 8, wherein the method further comprises retrieving the two or more matched fact-based structures.

10. A computer system embodied on one or more computer storage-media having computer-executable instructions provided thereon for performing a method for generating two or more fact-based structures, the system comprising:

a server that invokes a procedure configured for generating one or more fact-based structures, wherein the server comprises a processing unit coupled to the storage media, the storage having stored thereon a plurality of computer software components executable by the processing unit, the computer software components comprising;

(a) a fact-based structure generating component for generating the two or more fact-based structures using a process comprising;

(i) identifying one or more terms from textual content of documents;

(ii) determining a linguistic functionality of one or more terms and deriving one or more semantic relationships from the linguistic functionality of one or more terms;

(iii) representing the one or more terms as nodes;

(iv) representing the linguistic functionality of one or more terms as edges that interconnect the nodes, respectively, wherein each edge involves a direction of dependency that targets a single node;

(v) identifying from the nodes one or more head nodes, wherein the one or more head nodes are nodes being targeted by two or more of the edges; and

(vi) generating the two or more fact-based structures from the one or more head nodes, respectively, wherein a fact-based structure is directed toward one of the one or more head nodes, respectively; and

(b) an identifying component for assigning identifiers to the nodes within the two or more fact-based structures, respectively, wherein common nodes across the two or more fact-based structures are assigned a common identifier, and wherein the assigned identifiers are stored in the semantic index promoting retrieval across the two or more fact-based structures.

11. The computer system of claim 10, wherein the computer software components further comprising a receiving component to receive a search query from a user.

12. The computer system of claim 11, wherein the computer software components further comprising a matching component to match a proposition of the search query to the fact-based structure, wherein the proposition is a structural representation of terms within the search query derived from, in part, semantic relationships between the terms of the search query.