US 10,592,542 B2Granted

Document ranking by contextual vectors from natural language query

AssigneeInternational Business Machines Corporation

InventorsTakashi Fukuda, Hiroaki Kikuchi

Life of this patent

Estimating the term…

Loading patent…

Abstract

A set of keywords is extracted from a query. Natural Language Processing (NLP) is performed on the query to extract a set of contextual words for a keyword from the query. For the query, a first score of a first vector is computed, where the first vector represents a first contextual word. For a first result in a result set, a first result score of a first result vector is computed, where the first vector represents a first result contextual word in a set of result contextual words corresponding to the keyword in the first result. Using the first score and the first result score, a first similarity value is computed for the first result. The first result is re-ranked relative to a second result according to the first similarity value for the first result and a second similarity value for the second result in the result set.

Claims19

Dependent claims

A method comprising:

extracting a set of keywords from a query;
performing, for a keyword in the set of keywords, Natural Language Processing (NLP) on the query to extract a set of contextual words from the query;
computing, for the query, using a processor and a memory, a first score of a first vector, wherein the first vector represents a first contextual word in the set of contextual words, wherein a contextual word in the set of contextual words operates as a reference to a subject-matter domain in which a meaning of the keyword is intended;
performing a search using the keyword;
obtaining, responsive to the search, the result set;
analyzing, to detect the keyword therein, a first result in the result set, the keyword comprising a subset of the first result;
constructing, in the first result, a first result context in a set of result contexts, each of the set of result contexts comprising a context of the keyword within the result, each of the set of result contexts comprising less than the entirety of the first result;
computing, for the first result, a first result score of a first result vector, wherein the first result vector represents the first result context;
computing, using the first score and the first result score, a first result similarity value for the first result;
and
re-ranking, according to the first result similarity value for the first result and a second result similarity value for a second result in the result set, the first result relative to the second result.

8 more elements

A computer usable program product comprising one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices, the stored program instructions comprising:

program instructions to extract a set of keywords from a query;
program instructions to perform, for a keyword in the set of keywords, Natural Language Processing (NLP) on the query to extract a set of contextual words from the query;
program instructions to compute, for the query, using a processor and a memory, a first score of a first vector, wherein the first vector represents a first contextual word in the set of contextual words, wherein a contextual word in the set of contextual words operates as a reference to a subject-matter domain in which a meaning of the keyword is intended;
program instructions to perform a search using the keyword;
program instructions to obtain, responsive to the search, the result set;
program instructions to analyze, to detect the keyword therein, a first result in the result set, the keyword comprising a subset of the first result;
program instructions to construct, in the first result, a first result context in a set of result contexts, each of the set of result contexts comprising a context of the keyword within the result, each of the set of result contexts comprising less than the entirety of the first result;
program instructions to compute, for the first result, a first result score of a first result vector, wherein the first result vector represents the first result context;
program instructions to compute, using the first score and the first result score, a first result similarity value for the first result;
and
program instructions to re-rank, according to the first result similarity value for the first result and a second result similarity value for a second result in the result set, the first result relative to the second result.

8 more elements

A computer system comprising one or more processors, one or more computer-readable memories, and one or more computer-readable storage devices, and program instructions stored on at least one of the one or more storage devices for execution by at least one of the one or more processors via at least one of the one or more memories, the stored program instructions comprising:

program instructions to extract a set of keywords from a query;
program instructions to perform, for a keyword in the set of keywords, Natural Language Processing (NLP) on the query to extract a set of contextual words from the query;
program instructions to compute, for the query, using a processor and a memory, a first score of a first vector, wherein the first vector represents a first contextual word in the set of contextual words, wherein a contextual word in the set of contextual words operates as a reference to a subject-matter domain in which a meaning of the keyword is intended;
program instructions to perform a search using the keyword;
program instructions to obtain, responsive to the search, the result set;
program instructions to analyze, to detect the keyword therein, a first result in the result set, the keyword comprising a subset of the first result;
program instructions to construct, in the first result, a first result context in a set of result contexts, each of the set of result contexts comprising a context of the keyword within the result, each of the set of result contexts comprising less than the entirety of the first result;
program instructions to compute, for the first result, a first result score of a first result vector, wherein the first result vector represents the first result context;
program instructions to compute, using the first score and the first result score, a first result similarity value for the first result;
and
program instructions to re-rank, according to the first result similarity value for the first result and a second result similarity value for a second result in the result set, the first result relative to the second result.

8 more elements

Dependencies are inferred from the claim text.