Unsupervised machine learning system to automate functions on a graph structure

Patent metadata

Field	Value
Publication number	US-11710033-B2
Application number	US-201816006559-A
Country	US
Kind code	B2
Filing date	Jun 12, 2018
Priority date	Jun 12, 2018
Publication date	Jul 25, 2023
Grant date	Jul 25, 2023

Abstract

Machine learning models, semantic networks, adaptive systems, artificial neural networks, convolutional neural networks, and other forms of knowledge processing systems are disclosed. An ensemble machine learning system is coupled to a graph module storing a graph structure, wherein a collection of entities and the relationships between those entities forms nodes and connection arcs between the various nodes. A hotfile module and hotfile propagation engine coordinate with the graph module or may be subsumed within the graph module, and implement the various hot file functionality generated by the machine learning systems.

First claim

We claim:

1. A method comprising:
determining data corresponding to one or more graph representations of a first plurality of entities, wherein the one or more graph representations indicate a plurality of relationships between at least two of the first plurality of entities, and wherein the one or more graph representations are unlabeled;
training, using the data corresponding to the one or more graph representations, an artificial neural network for machine learning executing on one or more computing devices, wherein the artificial neural network comprises a plurality of nodes, wherein the nodes are configured to process an input, and wherein the plurality of nodes are configured based on the one or more graph representations;
determining a first graph representation comprising a second plurality of entities;
determining a plurality of definitional functions corresponding to one or more of the second plurality of entities;
receiving, from the artificial neural network and based on the first graph representation and the plurality of definitional functions, output indicating a modification to a hotfile; and
modifying the hotfile associated with one or more entities of the first plurality of entities, wherein the hotfile is a dynamic graph representing risk associated with transaction data.

2. The method of claim 1 , further comprising:
determining the modification to the hotfile based on the first graph representation, the plurality of definitional functions, and historical hotfile data.

3. The method of claim 1 , wherein a first definitional function of the plurality of definitional functions indicates a degree of relationship between a first entity of the second plurality of entities and a second entity of the second plurality of entities.

4. The method of claim 1 , wherein training the artificial neural network comprises providing, to the artificial neural network, data comprising the one or more graph representations, and wherein the data is unlabeled.

5. The method of claim 1 , wherein the one or more graph representations are associated with one or more transactions between at least two of the plurality of entities.

6. The method of claim 1 , wherein the modification to the hotfile causes the method to:
add or remove a first entity of the first plurality of entities to or from the hotfile; or
add or remove a first relationship between two entities of the first plurality of entities to or from the hotfile.

7. The method of claim 1 , further comprising:
determining a transaction between at least two entities of the first plurality of entities; and
causing, based on the hotfile, rejection of the transaction.

8. An apparatus comprising:
one or more processors; and
memory storing instructions that, when executed by the one or more processors, cause the apparatus to:
determine data corresponding to one or more graph representations of a first plurality of entities, wherein the one or more graph representations indicate a plurality of relationships between at least two of the first plurality of entities, and wherein the one or more graph representations are unlabeled;
train, using the data corresponding to the one or more graph representations, an artificial neural network for machine learning executing on one or more computing devices, wherein the artificial neural network comprises a plurality of nodes, wherein the nodes are configured to process an input, and wherein the plurality of nodes are configured based on the one or more graph representations;
determine a first graph representation comprising a second plurality of entities;
determine a plurality of definitional functions corresponding to one or more of the second plurality of entities;
receive, from the artificial neural network and based on the first graph representation and the plurality of definitional functions, output indicating a modification to a hotfile; and
modifying the hotfile associated with one or more entities of the first plurality of entities, wherein the hotfile is a dynamic graph representing risk associated with transaction data.

9. The apparatus of claim 8 , wherein the instructions, when executed by the one or more processors, cause the apparatus to:
determining the modification to the hotfile based on the first graph representation, the plurality of definitional functions, and historical hotfile data.

10. The apparatus of claim 8 , wherein a first definitional function of the plurality of definitional functions indicates a degree of relationship between a first entity of the second plurality of entities and a second entity of the second plurality of entities.

11. The apparatus of claim 8 , wherein training the artificial neural network comprises providing, to the artificial neural network, data comprising the one or more graph representations, and wherein the data is unlabeled.

12. The apparatus of claim 8 , wherein the one or more graph representations are associated with one or more transactions between at least two of the plurality of entities.

13. The apparatus of claim 8 , wherein the modification to the hotfile causes the apparatus to:
add or remove a first entity of the first plurality of entities to or from the hotfile;
add or remove a first relationship between two entities of the first plurality of entities to or from the hotfile; or
modify the hotfile associated with one or more entities of the first plurality of entities.

14. The apparatus of claim 8 , wherein the instructions, when executed by the one or more processors, cause the apparatus to:
determining a transaction between at least two entities of the first plurality of entities; and
causing, based on the hotfile, rejection of the transaction.

Assignees

• Bank Of America

Inventors

• Morris Ronnie J
• Pusey-Conlin Dana M
• Edkin Lorraine C
• Sims Scott A
• Filliben Joel
• Payne Margaret A
• Widmann Craig Douglas
• Kursun Eren

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/082

  modifying the architecture, e.g. adding, deleting or silencing nodes or connections · CPC title

• G06N3/0455

  Auto-encoder networks; Encoder-decoder networks · CPC title

• G06N3/0895

  Weakly supervised learning, e.g. semi-supervised or self-supervised learning · CPC title