US 10,769,140 B2Granted

Concept expansion using tables

AssigneeMicrosoft Technology Licensing, LLC

InventorsPhilip A. Bernstein, Kaushik Chakrabarti, Zhimin Chen, Yeye He, Chi Wang, Kris K. Ganjam

Life of this patent

Estimating the term…

Loading patent…

Abstract

Concept expansion using tables, such as web tables, can return entities belonging to a concept based on an input of the concept and at least one seed entity that belongs to the concept. A concept expansion frontend can receive the concept and seed entity and provide them to a concept expansion framework. The concept expansion framework can expand the coverage of entities for concepts, including tail concepts, using tables by leveraging rich content signals corresponding to concept names. Such content signals can include content matching the concept that appear in captions, early headings, page titles, surrounding text, anchor text, and queries for which the page has been clicked. The concept expansion framework can use the structured entities in tables to infer exclusive tables. Such inference differs from previous label propagation methods and involves modeling a table-entity relationship. The table-entity relationship reduces semantic drift without using a reference ontology.

Claims18

Dependent claims

A device comprising:

a processing unit;
a computer-readable medium including modules configured for execution by the processor, the modules comprising:
a receiver module configured to receive a query including a query concept name and at least one seed entity;
an expansion module configured to perform concept expansion by:
searching a database for tables of entities based on the query concept name;
ranking the tables based on at least one keyword of the query concept name that is found in at least one of column headers, captions, and surrounding text of the tables;
ranking data entities of the tables based on the ranking of the tables and the at least one seed entity matching at least one of the data entities, the data entities including table cell values and data associated with the tables;
identifying at least one of the tables, wherein each data entity in the at least one of the tables is relevant to the at least one seed entity, having no data entity that has a rank less than a predetermined parameter value, as an exclusive table;
determining one or more data entities relevant to the query concept name from only the exclusive table;
producing an expanded concept including the query concept name, the at least one seed entity, and the one or more data entities relevant to the query concept name;
building an expanded query based on the expanded concept;
and
an output module configured to expose results for the expanded concept as an expanded query.

10 more elements

A computer-readable memory having stored therein computer executable instructions, arranged to, upon execution, configure a computer to perform operations comprising:

receiving a query including a query concept name and at least one seed entity;
identifying a plurality of tables, each table including the query concept name or the at least one seed entity as data entities;
ranking the tables based on at least one keyword of the query concept name that is found in at least one of column headers, captions, and surrounding text of the tables;
ranking data entities of the tables based on the ranking of the tables and the at least one seed entity matching at least one of the data entities, the data entities including table cell values and data associated with the tables;
identifying at least one of the tables, wherein each data entity in the at least one of the tables is relevant to the at least one seed entity, having no data entity that has a rank less than a predetermined parameter value, as an exclusive table;
determining one or more data entities relevant to the query concept name from only the exclusive table;
generating a graph between entities including the at least one seed entity and the data entities of the tables;
performing concept expansion to generate an expanded concept by processing the query based at least in part on the graph and the exclusive table, wherein the expanded concept includes the query concept name, and the one or more data entities relevant to the query concept name;
and
generating an expanded query from the expanded concept.

7 more elements

Dependencies are inferred from the claim text.