Predicting lexical answer types in open domain question and answering (QA) systems

• US 8,880,388 B2
• Filed: 08/28/2012
• Issued: 11/04/2014
• Est. Priority Date: 08/04/2011
• Status: Active Grant

• Current Assignee: Hyundai Motor Company (Hyundai Motor Group), Kia Corp.
• Original Assignee: International Business Machines Corporation
• Inventors: Ferrucci, David A., Gliozzo, Alfio M., Kalyanpur, Aditya A.
• Primary Examiner(s): Abebe, Daniel D
• Application Number:
  US13/596,636
• Publication Number:
  US 20130035931A1
• Time in Patent Office: 798 Days
• Field of Search: 704/7, 704/9, 707/708
• US Class Current: 704/7
• CPC Class Codes: G06F 16/90332 Natural language query formulation or dialogue systems
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

In an automated Question Answer (QA) system architecture for automatic open-domain Question Answering, a system, method and computer program product for predicting the Lexical Answer Type (LAT) of a question. The approach is completely unsupervised and is based on a large-scale lexical knowledge base automatically extracted from a Web corpus. This approach for predicting the LAT can be implemented as a specific subtask of a QA process, and/or used for general purpose knowledge acquisition tasks such as frame induction from text.

Claims

1. A system for predicting a lexical answer types (LAT) in a question comprising:

a memory storage device including a plurality of syntactic frames;

a processor device operatively connected to said memory storage device and configured to;

receive a question text string;

extract at least one syntactic frame from said question string,

designate, in said syntactic frame, a placeholder for an entity corresponding to a potential lexical answer type; and

query a lexical knowledge database to automatically obtain at least one replacement term for said placeholder of said at least one syntactic frame,

wherein said entity placeholder is a part of a question focus indicating a LAT of the question.

2. The system as claimed in claim 1, wherein to extract said at least one syntactic frame, said processor device is further programmed to:

decompose said question text string into said at least one syntactic frame, each syntactic frame including a corresponding a focus slot-value pair, each said slot representing a syntactic role identified by a dependency relation.

3. The system as claimed in claim 1, wherein said processor device is further configured to:

substitute at least one of said replacement terms with a generalized type information term using a database of entity type knowledge.

4. The system as claimed in claim 3, wherein said entity type knowledge comprises frames having terms with is_a relationships.

5. The system as claimed in claim 3, wherein to query said lexical knowledge database to obtain said at least one replacement term, said processor device is configured to apply a Generalized Frame Model to obtain said at least one replacement term.

6. The system as claimed in claim 1, wherein said processor device is further programmed to:

rank said one or more replacement terms; and

select a top-ranked replacement term as an inferred lexical answer type to said question.

7. The system as claimed in claim 1, wherein said processor device if further configured to:

filter out said replacement terms if said replacement term is not related to the context of said question.

8. The system as claimed in claim 5, wherein said lexical knowledge database includes frame structures identified from a corpus of text, a frame structure having one or more slots-value pairs, a slot representing a syntactic role identified by a dependency relation, wherein to apply a Generalized Frame Model, said processor device is further configured to:

specify a frame cut comprising a sub-set of frames having non-empty slot-values for a given subset of slot-values pairs;

specify a frame abstraction to determine a desired relationship among selected selected slot values from said given subset of slot-values pairs; and

,

generate, from said frame cut, plural frame vectors defining a multi-dimensional vector space from which relationships among selected selected slot values is determined; and

,

processing said frame vectors to determine said desired relationship among selected selected slot values as defined by said abstraction.

9. The system as claimed in claim 8, wherein said desired relationship includes an inferred type of a slot representing one of:

a syntactic subject role or syntactic object in corresponding verb and noun phrases.