Asset risk profiling for enterprise resources

Patent metadata

Field	Value
Publication number	US-12354042-B2
Application number	US-202217878522-A
Country	US
Kind code	B2
Filing date	Aug 1, 2022
Priority date	Aug 2, 2021
Publication date	Jul 8, 2025
Grant date	Jul 8, 2025

Abstract

Aspects of the disclosure relate to using machine learning for asset risk profiling. A computing platform may receive risk information from an enterprise system. The computing platform may determine a ranking criterion for a plurality of risk factors. The computing platform may use a machine learning classifier to determine a weight corresponding to each risk factor and determine an asset risk profile (ARP) score for the enterprise system. Based on the ARP score, the computing platform may determine a risk control scheme and provide the risk control scheme to an enterprise control server.

First claim

What is claimed is:

1. A computer-implemented method comprising:
receiving risk information from a plurality of computing devices associated with an enterprise system, the risk information associated with a plurality of risk factors associated with the enterprise system, wherein the risk information is indicative of a likelihood score associated with an expected impact of loss for each risk factor in an event of a compromise of the enterprise system, prior to an application of enterprise controls;
determining a ranking criterion for the plurality of risk factors associated with the enterprise system;
receiving training data associated with a plurality of enterprise systems;
training a first machine learning classifier via a machine learning model based on the training data associated with the plurality of enterprise systems to determine a weight corresponding to each risk factor;
adjusting each weight during each iteration of training of the first machine learning classifier;
using the first machine learning classifier that is trained via the machine learning model to generate as an output the weight corresponding to each risk factor based on an input of the ranking criterion and the risk information;
determining, based on the risk information and the weights generated by the first machine learning classifier as the output, an asset risk profile (ARP) score for the enterprise system;
determining, based on the ARP score, an ARP impact rank of a first level or a second level, the first level ranked and comprising corresponding weights higher than the second level;
determining, based on the ARP score and the ARP impact rank, a risk control scheme associated with the enterprise system, the risk control scheme for the first level being different from the risk control scheme for the second level; and
providing, to an enterprise control server, the risk control scheme.

2. The computer-implemented method of claim 1 , wherein the compromise of the enterprise system comprises a compromise of confidentiality, integrity, availability, or combinations thereof.

3. The computer-implemented method of claim 1 , wherein using the first machine learning classifier comprises:
receiving a collection of risk data and labels as the training data associated with the plurality of enterprise systems; and
training the first machine learning classifier based on the collection of risk data and the labels.

4. The computer-implemented method of claim 3 , wherein using the first machine learning classifier further comprises:
determining a confidence score indicative of whether a predicted label output by the first machine learning classifier matches a pre-defined label of the enterprise system.

5. The computer-implemented method of claim 4 , wherein using the first machine learning classifier further comprises:
tuning the first machine learning classifier until the confidence score exceeds a first predetermined threshold.

6. The computer-implemented method of claim 1 , wherein determining the ARP scores comprises:
using the weights generated by the first machine learning classifier as inputs for a second machine learning classifier; and
using the second machine learning classifier to determine the ARP score for the enterprise system.

7. The computer-implemented method of claim 6 , wherein the first machine learning classifier is an unsupervised machine learning classifier and the second machine learning classifier is a supervised machine learning classifier.

8. The computer-implemented method of claim 1 , further comprising:
receiving one or more files; and
extracting the risk information from the one or more files.

9. The computer-implemented method of claim 8 , wherein the one or more files comprise audio files, internet files, visual files, audiovisual files, text files, image files, multimedia files, or combinations thereof.

10. The computer-implemented method of claim 1 , wherein at least one of the risk factors is related to cybersecurity threats or regulatory implications.

11. The computer-implemented method of claim 1 , wherein the weight corresponding to each risk factor comprises one or more components including at least a score multiplier assigned to the corresponding risk factor.

12. The computer-implemented method of claim 1 , wherein the first level of the ARP impact rank comprises the ARP score of a first threshold or above and is indicative of a critical impact system, and the second level of the ARP impact rank comprises the ARP score of below the first threshold and is indicative of a less critical impact system than the critical impact system of the first level such that the risk control scheme for the first level applies an enhanced critical control scheme compared to the risk control scheme for the second level.

13. The computer-implemented method of claim 1 , further comprising:
classifying the risk factors into a plurality of data categories based on a pervasiveness and a business value of the enterprise system, and an impact if the enterprise system is compromised.

14. The computer-implemented method of claim 13 , further comprising:
classifying the risk factors into one of a business drivers category, an attack surface category, a data access category, and a data types category.

15. An enterprise asset risk profiling platform comprising a processor configured to:
receive risk information from a plurality of computing devices associated with an enterprise system, the risk information associated with a plurality of risk factors associated with the enterprise system, wherein the risk information is indicative of a likelihood score associated with an expected impact of loss for each risk factor in an event of a compromise of confidentiality, integrity, availability, or combinations thereof, prior to an application of enterprise controls;
determine a ranking criterion for the plurality of risk factors associated with the enterprise system;
receive training data associated with a plurality of enterprise systems;
train a first machine learning classifier via a machine learning model based on the training data associated with the plurality of enterprise systems to determine a weight corresponding to each risk factor;
adjust each weight during each iteration of training of the first machine learning classifier;
use the first machine learning classifier that is trained via the machine learning model to generate as an output the weight corresponding to each risk factor based on an input of the ranking criterion and the risk information;
determine, based on the risk information and the weights generated by the first machine learning classifier as the output, an asset risk profile (ARP) score for the enterprise system;
determine, based on the ARP score, an ARP impact rank of a first level or a second level, the first level ranked higher than the second level;
determine, based on the ARP score and the ARP impact rank, a risk control scheme associated with the enterprise system, the risk control scheme for the first level being different from the risk control scheme for the second level; and
provide, to an enterprise control server, the risk control scheme.

16. The enterprise asset risk profiling platform of claim 15 , wherein using the first machine learning classifier comprises:
receiving a collection of risk data and labels as the training data associated with the plurality of enterprise systems; and
training the first machine learning classifier based on the collection of risk data and the labels.

17. The enterprise asset risk profiling platform of cl

Assignees

• Allstate Insurance Co

Inventors

• Jarvis Daniel

Classifications

• G06Q10/06375 Primary

  Prediction of business process outcome or impact based on a proposed change · CPC title

• G06Q10/0635 Primary

  Risk analysis of enterprise or organisation activities · CPC title