Utilizing a graph neural network to generate visualization and attribute recommendations

Patent metadata

Field	Value
Publication number	US-12093322-B2
Application number	US-202217654933-A
Country	US
Kind code	B2
Filing date	Mar 15, 2022
Priority date	Mar 15, 2022
Publication date	Sep 17, 2024
Grant date	Sep 17, 2024

Abstract

The present disclosure relates to systems, methods, and non-transitory computer readable media that utilize a graph neural network to generate data recommendations. The disclosed systems generate a digital graph representation comprising user nodes corresponding to users, data attribute nodes corresponding to data attributes, and edges reflecting historical interactions between the users and the data attributes; Moreover, the disclosed systems generate, utilizing a graph neural network, user embeddings for the user nodes and data attribute embeddings for the data attribute nodes from the digital graph representation. In addition, the disclosed systems generate, utilizing a graph neural network, user embeddings for the user nodes and data attribute embeddings for the data attribute nodes from the digital graph representation. Furthermore, the disclosed systems determine a data recommendation for a target user utilizing the data attribute embeddings and a target user embedding corresponding to the target user from the user embeddings.

First claim

What is claimed is:

1. A non-transitory computer readable medium comprising instructions that, when executed by a processing device, cause the processing device to perform operations comprising:
generating, from one or more data sets associated with users, a data attribute graph from data attributes utilized by the users, a visual configuration parameters graph from visual configuration parameters of data visualizations portraying the data attributes, and a combined data attribute visual configuration parameters graph from the data attributes and the visual configuration parameters;
generating, based on the data attribute graph, the visual configuration parameters graph, and the combined data attribute visual configuration parameters graph, a digital graph representation by:
generating visual configuration nodes reflecting the visual configuration parameters of the data visualizations,
generating user nodes corresponding to the users,
generating data attribute nodes corresponding to the data attributes, and
generating edges connecting the visual configuration nodes, the user nodes, and the data attribute nodes within the digital graph representation, wherein the edges reflect historical interactions between the users, the data attributes, and the visual configuration parameters of the data visualizations;

generating, utilizing a graph neural network, user embeddings for the user nodes, data attribute embeddings for the data attribute nodes, and visual configuration embeddings for the visual configuration parameters from the digital graph representation; and
determining a data recommendation comprising:
a target data attribute for a target user utilizing the data attribute embeddings and a target user embedding corresponding to the target user from the user embeddings, or
a target data visualization comprising a set of target visual configuration parameters for the target user utilizing the visual configuration embeddings and the target user embedding corresponding to the target user.

2. The non-transitory computer readable medium of claim 1 , wherein generating the user embeddings and the data attribute embeddings comprises, for a first node:
determining neighborhood nodes corresponding to the first node within the digital graph representation;
generating an aggregated neighbor embedding by combining embeddings corresponding to the neighborhood nodes utilizing an aggregator model; and
generating an embedding for the first node utilizing learned parameters of the graph neural network from the aggregated neighbor embedding.

3. The non-transitory computer readable medium of claim 1 , further comprising instructions that, when executed by the processing device, cause the processing device to perform operations comprising:
determining the historical interactions by monitoring client device interactions of the users with the data visualizations, wherein the data visualizations comprise one or more of the data attributes and have one or more of the visual configuration parameters.

4. The non-transitory computer readable medium of claim 1 , further comprising instructions that, when executed by the processing device, cause the processing device to perform operations comprising identifying a target data visualization comprising a set of target data attributes and target visual configuration parameters.

5. The non-transitory computer readable medium of claim 4 , further comprising instructions that, when executed by the processing device, cause the processing device to perform operations comprising:
generating the data recommendation comprises generating a data visual recommendation comprising the target data visualization.

6. The non-transitory computer readable medium of claim 1 , wherein generating the data recommendation comprises generating a data attribute recommendation by:
determining a target data attribute for the target user and a target data attribute embedding from the data attribute embeddings;
combining the target data attribute embedding and the target user embedding to determine a compatibility score; and
generating the data attribute recommendation utilizing the compatibility score.

7. The non-transitory computer readable medium of claim 6 , wherein generating the data attribute recommendation comprises:
determining, for an additional target data attribute, an additional target data attribute embedding from the data attribute embeddings;
generating a first edge score by combining the target data attribute embedding and the target user embedding;
generating a second edge score by combining the target data attribute embedding and the target user embedding; and
determining the compatibility score from the first edge score and the second edge score.

8. The non-transitory computer readable medium of claim 1 , wherein the data attribute graph represents data attributes in the one or more data sets, the visual configuration parameters graph represent visual configuration parameters of data visualizations, and the combined data attribute visual configuration parameters graph represents one or more attributes used with specific visual configuration parameters.

9. A system comprising:
one or more memory devices; and
one or more processing devices coupled to the one or more memory devices, the one or more processing devices configured to perform operations comprising:
generating, from one or more data sets associated with users, a data attribute graph from data attributes utilized by the users, a visual configuration parameters graph from visual configuration parameters of data visualizations portraying the data attributes, and a combined data attribute visual configuration parameters graph from the data attributes and the visual configuration parameters;
generating, based on the data attribute graph, the visual configuration parameters graph, and the combined data attribute visual configuration parameters graph, a digital graph representation by:
generating visual configuration nodes reflecting the visual configuration parameters of the data visualizations,
generating user nodes corresponding to the users,
generating data attribute nodes corresponding to the data attributes, and
generating edges connecting the visual configuration nodes, the user nodes, and the data attribute nodes within the digital graph representation, wherein the edges reflect historical interactions between the users, the data attributes, and the visual configuration parameters of the data visualizations;

generating, utilizing a graph neural network, user embeddings for the user nodes, data attribute embeddings for the data attribute nodes, and visual configuration embeddings for the visual configuration parameters from the digital graph representation; and
determining a data recommendation comprising:
a target data attribute for a target user utilizing the data attribute embeddings and a target user embedding corresponding to the target user from the user embeddings, or
a target data visualization comprising a set of target visual configuration parameters for the target user utilizing the visual configuration embeddings and the target user embedding corresponding to the target user.

10. The system of claim 9 , wherein the one or more processing devices are further configured to perform operations comprising:
generating a compatibility score utilizing a compatibility neural network from the target user embedding and a target data attribute embedding; and
generating the data recommendation based on the compatibility score.

11. The system of claim 9 , wherein the one or more processing devices are further configured to perform operations comprising:
determining a target data att

Assignees

• Adobe Inc

Inventors

• Ojo Fayokemi
• Rossi Ryan
• Hoffswell Jane
• Guo Shunan
• Du Fan
• Kim Sungchul
• Xiao Chang
• Koh Eunyee

Classifications

• G06N3/02

  Neural networks · CPC title

• G06N3/045

  Combinations of networks · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title

• G06N3/0475

  Generative networks · CPC title