Augmented text search with syntactic information

Patent metadata

Field	Value
Publication number	US-9904674-B2
Application number	US-201514859712-A
Country	US
Kind code	B2
Filing date	Sep 21, 2015
Priority date	Jun 22, 2015
Publication date	Feb 27, 2018
Grant date	Feb 27, 2018

Abstract

An approach is provided in which a knowledge manager generates syntactic annotation tokens that correspond to syntactic relationships between terms included in a source document. The knowledge manager creates a knowledge structure that stores the syntactic annotation tokens in parallel fields and stores the source document terms in original text fields, which align to the parallel fields. In turn, the knowledge manager utilizes the knowledge structure to generate answers to questions based upon the syntactic annotation tokens.

First claim

The invention claimed is:

1. A method implemented by an information handling system that includes a memory and a processor, the method comprising:
generating a plurality of syntactic annotation tokens based upon a plurality of syntactic relationships between a plurality of terms included in a source document;
creating a knowledge structure that includes the plurality of terms in a plurality of original text fields and includes the plurality of syntactic annotation tokens in a plurality of parallel fields, wherein each of the plurality of syntactic annotation tokens align to at least one of the plurality of original text fields;
generating one or more question-based syntactic annotation tokens corresponding to a question; and
utilizing the knowledge structure in a question answer system to generate one or more answers to the question, wherein the question answer system matches at least one of the one or more question-based syntactic annotation tokens to at least one of the plurality of syntactic annotation tokens during the generation of at least one of the one or more answers.

2. The method of claim 1 further comprising:
analyzing the question and identifying one or more question-based syntactic relationships between question terms in the question;
generating the one or more question-based syntactic annotation tokens based upon the question-based syntactic relationships;
including the question-based syntactic annotation tokens in a query; and
using the query in the querying of the knowledge structure.

3. The method of claim 2 wherein the querying of the knowledge structure further comprises:
identifying a selected one of the plurality of original text fields that align to the selected parallel field; and
generating one or more candidate answers utilizing one or more of the plurality of terms included in the selected original text field.

4. The method of claim 1 wherein, during the generation of the knowledge structure, the method further comprises:
matching a selected one of the plurality of terms to an abstract concept entry, wherein the abstract concept entry includes the selected term and a different term;
for each of the plurality of syntactic annotation tokens that include the selected term, creating an abstract syntactic annotation token that replaces the selected term with the different term, resulting in one or more abstract syntactic annotation tokens; and
including one or more abstract syntactic annotation tokens in the knowledge structure.

5. The method of claim 1 wherein, during the generation of the knowledge structure, the method further comprises:
selecting at least one term from the plurality of terms;
for each of the plurality of syntactic annotation tokens that include the selected term, creating a variable syntactic annotation token that replaces the selected term with a variable, resulting in a plurality of variable syntactic annotation tokens; and
including the plurality of variable syntactic annotation tokens in the knowledge structure.

6. The method of claim 1 wherein, during the generation of the knowledge structure, the method further comprises:
for one or more of the plurality of syntactic annotation tokens, creating a relaxed syntactic annotation token that replaces a syntactic relationship identifier with a variable, resulting in a plurality of relaxed syntactic annotation tokens; and
including the plurality of relaxed syntactic annotation tokens in the knowledge structure.

7. The method of claim 1 wherein, prior to the generating of the plurality of syntactic annotation tokens, the method further comprises:
parsing the source document using an English Slot Grammar (ESG) parser, resulting in a plurality of syntactic items that each identify a syntactic relationship between a first one of the plurality of terms and a second one of the plurality of terms, wherein the plurality of syntactic annotation tokes are generated from the plurality of syntactic items.

Assignees

• IBM

Inventors

• Boxwell Stephen A
• Nolan Christopher M
• Visotski William G
• Wilson David E

Classifications

• G06F40/211

  Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars · CPC title

• G06F16/24575

  using context · CPC title

• G06F16/243

  Natural language query formulation · CPC title

• G06F40/253

  Grammatical analysis; Style critique · CPC title

• G06F40/295

  Named entity recognition · CPC title