Systems and methods for streaming classification of distributed ledger-based activities

Patent metadata

Field	Value
Publication number	US-12050573-B2
Application number	US-202217806211-A
Country	US
Kind code	B2
Filing date	Jun 9, 2022
Priority date	Jun 10, 2021
Publication date	Jul 30, 2024
Grant date	Jul 30, 2024

Abstract

Systems and methods for streaming classification of distributed ledger-based activities are disclosed. In one embodiment, a method for streaming classification of distributed ledger-based activities may include: receiving, at a detection computer program executed by a server and from a distributed ledger network, information for a plurality of transactions; receiving, from a database and at the detection computer program, node information that identifies a plurality of nodes as involved in illicit transactions, as being involved in licit transactions, or unknown; applying, by the detection computer program, exponential time sampling to sample labels and associated features; training, by the detection computer program, a classifier; receiving, by the detection computer program, a streaming transaction from the distributed ledger network; predicting, by the detection computer program, a classification for the streaming transaction using the trained classifier; and outputting, by the detection computer program, an alert based on the prediction.

First claim

What is claimed is:

1. A method for streaming classification of distributed ledger-based activities, comprising:
receiving, at a detection computer program executed by a server and from a distributed ledger network, information for a plurality of transactions;
receiving, from a database and at the detection computer program, node information in which a plurality of nodes are previously labeled as involved in illicit transactions, are previously labeled as being involved in licit transactions, or are previously labeled as unknown;
applying, by the detection computer program, exponential time sampling to sample the labels and associated features;
training, by the detection computer program, a classifier using the sampled labels and features;
receiving, by the detection computer program, a streaming transaction from the distributed ledger network;
predicting, by the detection computer program, a classification for the streaming transaction using the trained classifier; and
taking, by the detection computer program, an action based on the prediction, wherein the action comprises approving the streaming transaction to proceed in response to the classification being a licit transaction, and stopping the streaming transaction and outputting an alert in response to the classification being an illicit transaction.

2. The method of claim 1 , wherein, for each transaction, the information comprises a plurality of nodes involved in the transaction, an amount of the transaction, and a number of parties to the transaction.

3. The method of claim 2 , wherein the information further comprises a block number, and an index and a value of a currency for the transaction.

4. The method of claim 1 , wherein the classifier comprises a Graph Convolutional Network.

5. The method of claim 4 , wherein the Graph Convolutional Network comprises an architecture having three hidden layers.

6. The method of claim 1 , wherein the classifier comprises a Random Forest classifier.

7. The method of claim 1 , wherein the trained classifier outputs a prediction as a binary value.

8. The method of claim 1 , wherein the trained classifier outputs a prediction of a probability.

9. The method of claim 1 , wherein the action further comprises adding addresses associated with the streaming transaction to a database in response to the classification being illicit.

10. A system, comprising:
a distributed ledger network comprising a plurality of nodes and storing information for a plurality of transactions; and
a server executing a detection computer program that receives the information for a plurality of transactions, receives node information in which a plurality of nodes are previously labeled as involved in illicit transactions, are previously labeled as being involved in licit transactions, or are previously labeled as unknown, applies exponential time sampling to sample the labels and associated features, trains a classifier using the sampled labels and features, receives a streaming transaction from the distributed ledger network, predicts a classification for the streaming transaction using the trained classifier, and takes an action based on the prediction, wherein the action comprises approving the streaming transaction to proceed in response to the classification being a licit transaction, and stopping the streaming transaction and outputting an alert in response to the classification being an illicit transaction.

11. The system of claim 10 , wherein, for each transaction, the information comprises a plurality of nodes involved in the transaction, an amount of the transaction, and a number of parties to the transaction.

12. The system of claim 11 , wherein the information further comprises a block number, and an index and a value of a currency for the transaction.

13. The system of claim 10 , wherein the classifier comprises a Graph Convolutional Network comprising an architecture having three hidden layers.

14. The system of claim 10 , wherein the classifier comprises a Random Forest classifier.

15. The system of claim 10 , wherein the trained classifier outputs a prediction as a binary value or a prediction of a probability.

16. The system of claim 10 , wherein the action comprises adding addresses associated with the streaming transaction to a database in response to the classification being illicit.

17. A non-transitory computer readable storage medium, including instructions stored thereon, which when read and executed by one or more computer processors, cause the one or more computer processors to perform steps comprising:
receiving, from a distributed ledger network, information for a plurality of transactions, wherein, for each transaction, the information comprises a plurality of nodes involved in the transaction, an amount of the transaction, and a number of parties to the transaction;
receiving, from a database, node information in which a plurality of nodes are previously labeled as involved in illicit transactions, are previously labeled as being involved in licit transactions, or are previously labeled as unknown;
applying exponential time sampling to sample the labels and associated features;
training a classifier using the sampled labels and features, wherein the classifier comprises a Graph Convolutional Network comprising an architecture having three hidden layers or a Random Forest classifier;
receiving a streaming transaction from the distributed ledger network;
predicting a classification for the streaming transaction using the trained classifier, wherein the trained classifier outputs a prediction as a binary value or a prediction of a probability; and
taking an action based on the prediction, wherein the streaming transaction is approved to proceed in response to the classification being a licit transaction, and wherein an alert is output in response to the classification being an illicit transaction.

18. The non-transitory computer readable storage medium of claim 17 , wherein the action further comprises adding addresses associated with the streaming transaction to a database in response to the classification being an illicit transaction.

Assignees

• Jpmorgan Chase Bank Na

Inventors

• Eloul Shaltiel
• Moran Sean
• Mendel Jacob
• Balitsky Shafir
• Dadia Tal

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0895

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

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G06F16/285 Primary

  Clustering or classification · CPC title