Customizing search queries for information retrieval

Patent metadata

Field	Value
Publication number	US-11775533-B2
Application number	US-202117149487-A
Country	US
Kind code	B2
Filing date	Jan 14, 2021
Priority date	Jan 14, 2021
Publication date	Oct 3, 2023
Grant date	Oct 3, 2023

Abstract

Methods and systems disclosed herein describe customizing searching. Search queries may be customized according to a user's preferences. A user may emphasize or indicate that additional weight should be given to one or more terms in a search query. Terms that are weighted higher may have a larger impact on the results that are returned in response to the search query. In addition to changing the terms in a search query, a user may provide a weight for each term. Each term in a search query may be weighted to varying degrees, giving a user more control over the results that are returned. The weights may be used with machine learning techniques to generate a vector representation of a search query. The vector representation of the search query may be compared with vector representations of search objects to determine results that match the search query.

First claim

What is claimed is:

1. A computer-implemented method comprising:
receiving, by a server and from a user device, a first search query, wherein the first search query comprises a first word and a second word;
receiving, from the user device, a weight associated with the first word in the first search query, wherein the weight is selected by interaction with a user interface of the user device;
generating, using a machine learning model, based on the first search query, and based on the weight, a vector representation of the first search query, wherein the first word is weighted higher than the second word in the vector representation;
executing, by the server and based on the vector representation, a second search query;
receiving, based on the executed second search query, search results comprising one or more documents; and
causing the search results to be displayed by the user device.

2. The method of claim 1 ,
wherein receiving the first search query comprises receiving a document,
wherein the generating the vector representation comprises:
generating, based on the weight received from the user device, a second vector representation of the document, and
comparing the vector representation and the second vector representation, and

wherein the second search query is further based on a comparison of the vector representation and the second vector representation.

3. The method of claim 1 , further comprising:
receiving a user identifier from the user device,
wherein the generating the vector representation is further based on weights of words stored in a user profile associated with the user identifier.

4. The method of claim 1 , wherein the first search query comprises a document.

5. The method of claim 4 , further comprising:
determining, based on highlighting within the document, the weight associated with the first word in the first search query.

6. The method of claim 4 , further comprising:
determining, based on a first color of highlighting within the document, the weight associated with the first word in the first search query; and
determining, based on a second color of highlighting within the document, a second weight associated with the second word in the first search query.

7. The method of claim 1 ,
wherein the weight indicates a value, and
wherein the generating the vector representation further comprises:
generating a second vector representation of the first word, and
generating, by multiplying the second vector representation by the value, the vector representation of the first search query.

8. The method of claim 1 ,
wherein the first search query comprises a third word,
wherein the generating the vector representation of the first search query comprises:
determining, based on weights previously assigned, by other users, to the third word, a weight for the third word,

wherein, the third word is weighted differently from the first word in the vector representation of the first search query.

9. A computer-implemented method comprising:
receiving, by a server and from a user device, a first search query, wherein the first search query comprises a first word and a second word;
receiving a first weight associated with the first word and a second weight associated with the second word, wherein the first weight and the second weight are associated with weights previously assigned by other users that have performed searches similar to the first search query;
determining, based on the received weights and the first search query, a weight associated with the first word in the first search query;
generating, using a machine learning model, based on the first search query, and based on the weight, a vector representation of the first search query, wherein the first word is weighted higher than the second word in the vector representation;
executing, by the server and based on the vector representation, a second search query;
receiving, based on the executed second search query, search results comprising one or more documents; and
causing the search results to be displayed by the user device.

10. The method of claim 9 ,
wherein the determining the weight associated with the first word in the first search query comprises averaging first weights previously assigned by the other users to the first word.

11. The method of claim 9 ,
wherein receiving the first search query comprises receiving a document,
wherein the generating the vector representation comprises:
generating, based on the weights, a second vector representation of the document, and
comparing the vector representation and the second vector representation of the document, and

wherein the second search query is further based on a comparison of the vector representation and the second vector representation.

12. The method of claim 9 , further comprising:
receiving a user identifier from the user device,
wherein the generating the vector representation is further based on weights of words stored in a user profile associated with the user identifier.

13. The method of claim 9 , wherein the first search query comprises a document, the method further comprising:
determining, based on highlighted words within the document, the weight associated with the first word in the first search query and a second weight associated with the second word.

14. The method of claim 9 ,
wherein the weight indicates a value, and
wherein the generating the vector representation further comprises:
generating a second vector representation of the first word, and
generating, by multiplying the second vector representation by the value, vector representation of the first search query.

15. The method of claim 9 , further comprising:
prior to outputting the search results to the user device, removing, based on user preferences associated with the user device, a subset of the one or more documents from the search results.

16. A computer-implemented method comprising:
receiving, by a server and from a user device, a first search query, wherein the first search query comprises a first word and a second word;
receiving, from the user device, a weight associated with the first word in the first search query, wherein the weight is selected by interaction with a user interface of the user device;
generating, based on the first search query and based on the weight, a first vector representation of the first search query;
executing, by the server and based on the first vector representation, a second search query;
receiving, based on the executed second search query, a document;
generating, based on the weight, a second vector representation of the document, wherein the first word is weighted higher than the second word in the second vector representation of the document;
comparing the first vector representation and the second vector representation to determine whether the document matches the first search query; and
outputting, based on a determination that the document matches the first search query, search results comprising the document.

17. The method of claim 16 , further comprising:
receiving a user identifier from the user device, wherein the generating the first vector representation of the first search query is based on weights of words stored in a user profile associated with the user identifier.

18. The method of claim 16 ,
wherein the first search query comprises a second document, and
wherein the receiving the weight comprises determining, based on highlighted words within the second document, the weight associated with th

Assignees

• Capital One Services Llc

Inventors

• Lin Alexander
• Vargas Cruz
• Edwards Joshua
• Miracolo Max
• Rodriguez Mia
• Atkins Phoebe

Classifications

• G06F16/24575 Primary

  using context · CPC title

• G06F16/248

  Presentation of query results · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title