PROVIDING QUESTION AND ANSWERS WITH DEFERRED TYPE EVALUATION USING TEXT WITH LIMITED STRUCTURE

• US 20120330934A1
• Filed: 09/06/2012
• Published: 12/27/2012
• Est. Priority Date: 09/24/2010
• Status: Abandoned Application

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Duboue, Pablo A., Fan, James J., Ferrucci, David A., Murdock, James W. IV, Welty, Christopher A., Zadrozny, Wlodek W.
• Application Number:
  US13/605,231
• Publication Number:
  US 20120330934A1
• Time in Patent Office: Days
• US Class Current: 707/722
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence

Abstract

A system, method and computer program product for conducting questions and answers with deferred type evaluation based on any corpus of data. The method includes processing a query including waiting until a “Type” (i.e. a descriptor) is determined AND a candidate answer is provided. Then, a search is conducted to look (search) for evidence that the candidate answer has the required Lexical Answer Type (e.g., as determined by a matching function that can leverage a parser, a semantic interpreter and/or a simple pattern matcher). Prior to or during candidate answer evaluation, a process is provided for extracting and storing collections of entity-type pairs from semi-structured text documents. During QA processing and candidate answer scoring, a process is implemented to match the query LAT against the lexical type of each provided candidate answer and generate a score judging a degree of match.

Claims

1. A system for generating answers to questions comprising:

a memory storage device;

one or more processing devices, each operatively connected to said memory device and configured for performing a method comprising;

determining a lexical answer type (LAT) associated with an input query received at an automated questions and answer (QA) system;

obtaining one or more candidate answers to the input query using a data source having semi-structured content;

determining a lexical type (LT) for each said one or more obtained candidate answers from said semi-structured content;

comparing, at said one or more processing devices, said query LAT with said candidate answer LT; and

generating a score representing a degree of match between said query LAT with said candidate answer LT, said score indicative of a quality of said obtained candidate answer.

2. The system of claim 1, wherein said processor device is further configured to:

identify, in said semi-structured content, one or more entities and associated lexical type information; and

store, in a data storage device in communication with said QA system, entity-type structures, each entity-type structure representing said one or more entities and associated lexical type information, wherein said determining a lexical type (LT) includes accessing said stored entity-type structures to identify a lexical type (LT) from a type associated with said one or more entities stored in said entity-type data structures.

3. The system of claim 2, wherein said identifying and storing are performed off-line as a pre-processing step.

4. The system of claim 2, wherein said identifying and storing are performed on-demand by a QA system at a time the input query is received.

5. The system of claim 2, wherein said processor device is configured to parse each respective said query LAT and said candidate answer LT to obtain respective terms or phrases for each.

6. The system of claim 1, wherein said comparing comprises matching individual terms of respective query LAT and candidate answer LT, or matching entire phrases of each respective query LAT and candidate answer LT.

7. The system of claim 5, wherein processing device is further configured to:

determine which terms of respective said query LAT and candidate answer LT are to be used for said term matching;

obtain individual scores based on degree of match between terms of respective said LAT and candidate answer LT; and

combine obtained scores of said matches determined for said phrase matching.

9. The system as claimed in claim 1, wherein said semi-structured text comprises item-delimited markup, said automatically identifying of semi-structured text data comprising:

parsing content of said data source to identify said item-delimiting markup, said item delimited mark-up specifying said type information and entities forming an entity-type data structure.

10. The system as claimed in claim 9, wherein said item-delimiting markup includes a title, a header, a recitation of the word “

list”

of entities of a specified type, bullet markers, parentheses, a hypertext link, a Uniform Resource Locator, a table in said data source.

8. A system for providing content to a database used by an automatic question and answering (QA) system comprising:

a memory storage device;

one or more processing devices, each operatively connected to said memory device and configured for performing a method to;

automatically identify semi-structured text data from a data source;

automatically identify one or more entity-type relations from said semi-structured text data, said entity-type relation including one or more entities associated with a type;

automatically extract said identified entity-type relations; and

store said extracted entity-type relations as entity-type data structures in said database,

wherein a processing device is configured to perform said automatic identifying of semi-structured text and entity-type relations, said extracting and said storing.