Attribute detection

• US 9,098,543 B2
• Filed: 03/14/2013
• Issued: 08/04/2015
• Est. Priority Date: 03/14/2013
• Status: Active Grant

• Current Assignee: Walmart Apollo, LLC (WalMart Inc.)
• Original Assignee: Wal-Mart Stores Texas LLC (WalMart Inc.)
• Inventors: Vuong, Ba-Quy, Gattani, Abhishek
• Primary Examiner(s): Hoang, Son T
• Application Number:
  US13/804,806
• Publication Number:
  US 20140280201A1
• Time in Patent Office: 873 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2255 Hash tables
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24522 Translation of natural language queries to structured queries

Abstract

Computer-implemented methods and systems may be used for identifying an attribute and/or attribute value in a text string. In embodiments, the text string comprises a search query submitted by a user. Embodiments of the present disclosure include identifying an attribute value from a search query by comparing the search query string to a list of known attribute values and comparing the candidate attribute value to a knowledge base to confirm that the string represents an attribute value rather than a non-attribute concept. In embodiments, a Bloom filter is employed to execute a relatively efficient comparison between a candidate attribute value and known non-attribute concepts.

Claims

1. A computer-implemented method of identifying an item attribute in a search query, comprising:

at a search query attribute module, receiving the search query from a user;

at the search query attribute module, comparing a string from the search query to a value in an attribute database, thereby resulting in a matched attribute string;

at a disambiguation module, processing the matched attribute string through a hash function, thereby resulting in a hashed attribute string;

at the disambiguation module, comparing the hashed attribute string to a Bloom filter table to determine if the matched attribute string does not belong to a set of known non-attribute concepts; and

if the matched attribute string does not belong to the set of known non-attribute concepts;

transmitting an item object to the user, wherein the item object comprises an attribute value that corresponds to the matched attribute string.

2. The method of claim 1, wherein receiving the search query comprises processing a voice command.

3. The method of claim 1, further comprising:

performing a keyword search of the search query in a product database.

4. The method of claim 1, further comprising:

performing a search of the matched attribute string in a product database.

5. The method of claim 1, further comprising:

at a Bloom filter module, retrieving an attribute value from a product database;

at the Bloom filter module, searching for a match to the attribute value in the attribute database, thereby resulting in a matched attribute value;

at the Bloom filter module, searching for the matched attribute value in a knowledge base, thereby resulting in a matched non-attribute concept;

at the Bloom filter module, processing a text string representing the matched non-attribute concept through the hash function, thereby resulting in a hashed non-attribute concept; and

inserting the hashed non-attribute concept into the Bloom filter table.

6. The method of claim 5, wherein the knowledge base comprises Wikipedia.

7. A computer-implemented method of identifying an item attribute in a search query, comprising:

at a disambiguation module of a user computing device, receiving the search query from a user;

at the disambiguation module, processing at least a portion of the search query through a hash function, thereby resulting in a hashed attribute string, wherein the at least the portion of the search query comprises a queried attribute value;

at the disambiguation module, comparing the hashed attribute string to a Bloom filter table to determine if the search query belongs to a set of known non-attribute concepts; and

if the hashed attribute string does not belong to the set of known non-attribute concepts;

displaying an item object to the user, wherein the item object comprises the queried attribute value that corresponds to the search query.

8. The method of claim 7, wherein receiving the search query comprises processing a voice command.

9. The method of claim 7, further comprising:

at a search query attribute module, comparing the search query to a value in an attribute database.

10. The method of claim 7, further comprising:

performing a keyword search for the search query in a product database.

11. The method of claim 7, further comprising:

performing a search for an attribute object in a product database, wherein the attribute object comprises the queried attribute value.

12. The method of claim 7, further comprising:

at a Bloom filter module, retrieving an attribute value from a product database;

at the Bloom filter module, searching for a match to the attribute value in an attribute database, thereby resulting in a matched attribute value;

at the Bloom filter module, searching for the matched attribute value in a knowledge base, thereby resulting in a matched non-attribute concept;

at the Bloom filter module, processing a text string representing the matched non-attribute concept through the hash function, thereby resulting in a hashed non-attribute concept; and

inserting the hashed non-attribute concept into the Bloom filter table.

13. A system for identifying an item attribute in a search query, comprising:

an attribute database adapted to store attribute values;

a search query attribute module comprising a search query attribute memory device and a search query attribute processing device, the search query attribute memory device storing computer-readable instructions directing the search query attribute processing device to query the attribute database for matches between a query string of the search query received from a user and the attribute values in the attribute database; and

a disambiguation module comprising a disambiguation memory device and a disambiguation processing device, the disambiguation memory device storing computer-readable instructions directing the disambiguation processing device to;

process the query string through a hash function, thereby resulting in a hashed query;

compare the hashed query against a Bloom filter table, wherein the Bloom filter table comprises a representative set of non-attribute concepts; and

in response to not finding a match in the Bloom filter table, return a search result to the user, wherein the search result comprises an item object with a matching attribute value to the query string.

14. The system of claim 13, wherein the system comprises computer-readable instructions stored in a memory of a mobile computing device.

15. The system of claim 14, further comprising a voice command module comprising a voice command memory device and a voice command processing device, the voice command memory device storing computer-readable instructions directing the voice command processing device to receive a voice command from the user and translate the voice command to the query string.

16. The system of claim 13, wherein the system comprises a network server.

17. The system of claim 16, further comprising a voice command module comprising a voice command memory device and a voice command processing device, the voice command memory device storing computer-readable instructions directing the voice command processing device to receive a voice command from the user and translate the voice command to the query string.

18. The system of claim 13, wherein the Bloom filter table is stored on the disambiguation memory device.

19. The system of claim 13, further comprising a Bloom filter module comprising a Bloom filter memory device and a Bloom filter processing device, the Bloom filter memory device storing computer-readable instructions directing the Bloom filter processing device to:

crawl a knowledge base;

process a non-attribute concept from the knowledge base through the hash function, thereby resulting in a hashed non-attribute concept; and

insert the hashed non-attribute concept into the Bloom filter table.

20. The system of claim 19, wherein the knowledge base comprises Wikipedia.