Systems and methods for generating recommendations using neural network and machine learning techniques

Patent metadata

Field	Value
Publication number	US-11544534-B2
Application number	US-202016779273-A
Country	US
Kind code	B2
Filing date	Jan 31, 2020
Priority date	Aug 23, 2019
Publication date	Jan 3, 2023
Grant date	Jan 3, 2023

Abstract

Systems and methods including one or more processors and one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and perform: receiving an input identifying an anchor item; determining, using a quadruplet network associated with a neural network architecture, one or more item categories corresponding to complementary items associated with the anchor item; generating, using a ranking network associated with the neural network architecture, scores for the complementary items included in the one or more item categories; generating, using the ranking network associated with the neural network architecture, first ranking results for the complementary items based, at least in part, on the scores; and selecting one or more of the complementary items to be displayed based, at least in part, on the first ranking results. Other embodiments are disclosed herein.

First claim

What is claimed is:

1. A system for recommending complementary items comprising:
one or more processors; and
one or more non-transitory computer-readable storage devices storing computing instructions configured to run on the one or more processors and perform:
receiving an input identifying an anchor item;
determining, using a quadruplet network associated with a neural network architecture, one or more item categories corresponding to the complementary items associated with the anchor item;
generating, using a ranking network associated with the neural network architecture, scores for the complementary items included in the one or more item categories;
generating, using the ranking network associated with the neural network architecture, first ranking results for the complementary items based, at least in part, on the scores; and
selecting one or more of the complementary items to be displayed based, at least in part, on the first ranking results.

2. The system of claim 1 , wherein the computing instructions are further configured to run on the one or more processors and perform:
determining whether a user profile is accessible for an individual who selected the anchor item;
in response to determining that the user profile for the individual is accessible, updating the scores for the complementary items using a re-ranking network associated with the neural network architecture; and
updating the first ranking results to create second ranking results for the complementary items based on the scores that are updated by the re-ranking network.

3. The system of claim 2 , wherein the one or more complementary items are selected to be displayed based, at least in part, on the second ranking results updated by the re-ranking network.

4. The system of claim 2 , wherein generating the second ranking results comprises:
generating, using the re-ranking network of the neural network architecture, a user profile embedding corresponding to the user profile;
generating, using the re-ranking network of the neural network architecture, item profile embeddings for the complementary items associated with the second ranking results; and
using the user profile embedding and item profile embeddings to update the scores.

5. The system of claim 1 , wherein generating the first ranking results comprises:
generating, using a text encoder associated with the neural network architecture, dense features comprising title embeddings and category embeddings corresponding to the complementary items; and
generating the first ranking results based, at least in part, on the dense features.

6. The system of claim 5 , wherein generating the first ranking results further comprises:
generating categorical features associated with the complementary items;
generating continuous features associated with the complementary items; and
generating the first ranking results based, at least in part, on the dense features, the categorical features, and the continuous features.

7. The system of claim 1 , wherein the anchor item and the one or more complementary items are accessible via an electronic platform over a network.

8. The system of claim 1 , wherein the computing instructions are further configured to run on the one or more processors and perform:
determining whether a user profile is accessible for an individual who selected the anchor item; and
in response to determining that the user profile for the individual is accessible, customizing the first ranking results based, at least in part, on the user profile.

9. A method for recommending complementary items implemented via execution of computing instructions configured to run at one or more processors and configured to be stored at non-transitory computer-readable media, the method comprising:
receiving an input identifying an anchor item;
determining, using a quadruplet network associated with a neural network architecture, one or more item categories corresponding to the complementary items associated with the anchor item;
generating, using a ranking network associated with the neural network architecture, scores for the complementary items included in the one or more item categories;
generating, using the ranking network associated with the neural network architecture, first ranking results for the complementary items based, at least in part, on the scores; and
selecting one or more of the complementary items to be displayed based, at least in part, on the first ranking results.

10. The method of claim 9 further comprising:
determining whether a user profile is accessible for an individual who selected the anchor item;
in response to determining that the user profile for the individual is accessible, updating the scores for the complementary items using a re-ranking network associated with the neural network architecture; and
updating the first ranking results to create second ranking results for the complementary items based on the scores that are updated by the re-ranking network.

11. The method of claim 10 , wherein the one or more complementary items are selected to be displayed based, at least in part, on the second ranking results updated by the re-ranking network.

12. The method of claim 10 , wherein generating the second ranking results comprises:
generating, using the re-ranking network of the neural network architecture, a user profile embedding corresponding to the user profile;
generating, using the re-ranking network of the neural network architecture, item profile embeddings for the complementary items associated with the second ranking results; and
using the user profile embedding and item profile embeddings to update the scores.

13. The method of claim 9 , wherein generating the first ranking results comprises:
generating, using a text encoder associated with the neural network architecture, dense features comprising title embeddings and category embeddings corresponding to the complementary items; and
generating the first ranking results based, at least in part, on the dense features.

14. The method of claim 13 , wherein generating the first ranking results further comprises:
generating categorical features associated with the complementary items;
generating continuous features associated with the complementary items; and
generating the first ranking results based, at least in part, on the dense features, the categorical features, and the continuous features.

15. The method of claim 9 , wherein the anchor item and the one or more complementary items are accessible via an electronic platform over a network.

16. The method of claim 9 further comprising:
determining whether a user profile is accessible for an individual who selected the anchor item; and
in response to determining that the user profile for the individual is accessible, customizing the first ranking results based, at least in part, on the user profile.

17. A computer program product for recommending complementary items, the computer program product comprising a non-transitory computer-readable medium including instructions for causing a computer to:
receive an input identifying an anchor item;
determine, using a quadruplet network associated with a neural network architecture, one or more item categories corresponding to the complementary items associated with the anchor item;
generate, using a ranking network associated with the neural network architecture, scores for the complementary items included in the one or more item categories;
generate, using the ranking network associated with the neural network architecture,

Assignees

• Walmart Apollo Llc

Inventors

• Mane Mansi Ranjit
• Sundaresan Anirudha
• Guo Stephen Dean
• Mantha Aditya
• Achan Kannan

Classifications

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06N3/08

  Learning methods · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N3/0454 Primary

  Physics · mapped topic

• G06V10/82 Primary

  using neural networks · CPC title