Systems and methods for altering a graphical user interface

Patent metadata

Field	Value
Publication number	US-12079455-B2
Application number	US-202318103142-A
Country	US
Kind code	B2
Filing date	Jan 30, 2023
Priority date	Jan 31, 2021
Publication date	Sep 3, 2024
Grant date	Sep 3, 2024

Abstract

A system comprising one or more processors and one or more non-transitory computer-readable storage devices storing computing instructions that, when executed to run on the one or more processors, cause the one or more processors to perform: receiving in-session user activity entered into on a graphical user interface (GUI) from a user electronic device of a user; generating, using a predictive algorithm, a ranked list of one or more likely intents of the user to perform one or more actions on the GUI; processing the in-session user activity to determine one or more intents of the in-session user activity; comparing the one or more intents of the in-session user activity; and coordinating a display of a likely-to-be-used GUI element on the GUI. Other embodiments are disclosed herein.

First claim

What is claimed is:

1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable storage devices storing computing instructions that, when executed to run on the one or more processors, cause the one or more processors to perform functions comprising:
receiving in-session user activity entered into on a graphical user interface (GUI) from a user electronic device of a user;
generating, using a predictive algorithm, a ranked list of one or more probabilities of one or more intents of the user to perform one or more actions during a session on the GUI;
processing the in-session user activity to determine one or more intents of the in-session user activity;
filtering out, from the one or more intents of the in-session user activity, one or more intents already performed during the in-session user activity;
reranking the one or more intents of the in-session user activity, as filtered;
comparing the one or more intents of the in-session user activity with the one or more intents of the user, wherein comparing the one or more intents of the in-session user activity with the one or more intents of the user further comprises:
generating one or more embeddings for at least part of the in-session user activity; and
comparing the one or more embeddings for the at least the part of the in-session user activity to one or more embeddings for the one or more intents of the in-session user activity; and

coordinating a display of a GUI element on the GUI of the user electronic device of the user based on the one or more intents of the user that most closely match the one or more intents of the in-session user activity.

2. The system of claim 1 , wherein the computing instructions when executed to run on the one or more processors, further cause the one or more processors to perform an operation comprising:
constructing a vector using the in-session user activity by incrementing a count for a specific action corresponding to a category of one or more items in a hierarchical categorization, wherein:
the count is stored in a database; and
the category is a level in the hierarchical categorization.

3. The system of claim 1 , wherein processing the in-session user activity comprises:
determining a normalized point-wise mutual information (NPMI) score between one or more GUI interactions in the in-session user activity.

4. The system of claim 3 , wherein the NPMI score is calculated as a function of:
point-wise mutual information between at least two intents of the one or more intents of the in-session user activity; and
joint self-information of the at least two intents of the one or more intents of the in-session user activity.

5. The system of claim 1 , wherein processing the in-session user activity comprises:
down sampling one or more popular intents of the one or more intents of the in-session user activity, wherein the one or more popular intents is determined using historical activity data.

6. The system of claim 1 , wherein:
processing the in-session user activity comprises:
constructing one or more sequences of the one or more intents of the in-session user activity from one or more GUI interactions of the in-session user activity.

7. The system of claim 1 , wherein comparing the one or more intents of the in-session user activity with the one or more intents of the user further comprises:
comparing (a) one or more sequences of the one or more intents of the user from one or more GUI interactions and (b) one or more sequences of one or more GUI interactions of the one or more intents of the user; and
determining a similarity metric between at least two intents of the one or more intents of the user, wherein the at least two intents are similar to each other when the similarity metric between the at least two intents exceeds a predetermined threshold.

8. The system of claim 1 , wherein generating the one or more embeddings comprises:
inputting one or more sequences of one or more GUI interactions of the in-session user activity into a neural network.

9. The system of claim 8 , wherein the neural network uses a Word2Vec skip-gram model.

10. The system of claim 8 , wherein an output layer of the neural network comprises a SoftMax classifier, and wherein the neural network comprises a 2-layer neural network having at least one hidden layer.

11. A method 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 in-session user activity entered into on a graphical user interface (GUI) from a user electronic device of a user;
generating, using a predictive algorithm, a ranked list of one or more probabilities of one or more intents of the user to perform one or more actions during a session on the GUI;
processing the in-session user activity to determine the one or more intents of the in-session user activity;
filtering out, from the one or more intents of the in-session user activity, one or more intents already performed during the in-session user activity;
reranking the one or more intents of the in-session user activity, as filtered;
comparing the one or more intents of the in-session user activity with the one or more intents of the user, wherein comparing the one or more intents of the in-session user activity with the one or more intents of the user further comprises:
generating one or more embeddings for at least part of the in-session user activity; and
comparing the one or more embeddings for the at least the part of the in-session user activity to one or more embeddings for the one or more intents of the in-session user activity; and

coordinating a display of a GUI element on the GUI of the user electronic device of the user based on the one or more intents of the user that most closely match the one or more intents of the in-session user activity.

12. The method of claim 11 further comprising:
constructing a vector using the in-session user activity by incrementing a count for a specific action corresponding to a category of one or more items in a hierarchical categorization, wherein:
the count is stored in a database; and
the category is a level in the hierarchical categorization.

13. The method of claim 11 , wherein processing the in-session user activity comprises:
determining a normalized point-wise mutual information (NPMI) score between one or more GUI interactions in the in-session user activity.

14. The method of claim 13 , wherein the NPMI score is calculated as a function of:
point-wise mutual information between at least two intents of the one or more intents of the in-session user activity; and
joint self-information of the at least two intents of the one or more intents of the in-session user activity.

15. The method of claim 11 , wherein processing the in-session user activity comprises:
down sampling one or more popular intents of the one or more intents of the in-session user activity, wherein the one or more popular intents is determined using historical activity data.

16. The method of claim 11 , wherein:
pre-processing the in-session user activity comprises:
constructing one or more sequences of the one or more intents of the in-session user activity from one or more GUI interactions of the in-session user activity.

17. The method of claim 11 , wherein comparing the one or more intents of the in-session user activity with the one or more intents of the user further comprises:
comparing (a) one or more sequences of the

Assignees

• Walmart Apollo Llc

Inventors

• You Jiwen
• Subramaniam Sinduja
• Cerekovic Aleksandra
• Korpeoglu Evren
• Achan Kannan

Classifications

• G06N3/0495

  Quantised networks; Sparse networks; Compressed networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06F9/451 Primary

  Execution arrangements for user interfaces · CPC title

• G06N3/02

  Neural networks · CPC title