Automated path-based recommendation for risk mitigation

Patent metadata

Field	Value
Publication number	US-11900294-B2
Application number	US-202017753122-A
Country	US
Kind code	B2
Filing date	Aug 19, 2020
Priority date	Aug 22, 2019
Publication date	Feb 13, 2024
Grant date	Feb 13, 2024

Abstract

Systems and methods for automated path-based recommendation for risk mitigation are provided. An entity assessment server, responsive to a request for a recommendation for modifying a current risk assessment score of an entity to a target nsk assessment score, accesses an input attribute vector for the entity and clusters of entities defined by historical attribute vectors. The entity assessment server assigns the input attribute vector to a particular cluster and determines a requirement on movement from a first point to a second point in a multi-dimensional space based on the statistics computed from the particular cluster. The first point corresponds to the current risk assessment score and the second point corresponds to the target risk assessment score. The entity assessment server computes an attribute-change vector so that a path defined by the attribute-change vector complies with the requirement and generates the recommendation from the attribute-change vector.

First claim

The invention claimed is:

1. A method in which one or more processing devices of a server system perform operations comprising:
receiving, from a user device, a request for a recommendation for modifying a current value of a risk assessment score computed from input attribute values of an entity associated with the user device to a target value of the risk assessment score, the risk assessment score used to control access by a computing device associated with the entity to an interactive computing environment hosted by a host system, wherein the host system is configured to present a first type of interactive user experience to a first type of user group associated with the current value of the risk assessment score and present to a second type of interactive user experience to a second type of user group associated with the target value of the risk assessment score;
accessing (a) an input attribute vector having the input attribute values of the entity and (b) clusters of points in an n-dimensional space, wherein the clusters of points are respectively defined by historical attribute vectors;
assigning the input attribute vector to a particular cluster based on a similarity measure between the particular cluster and an input point defined by the input attribute vector;
computing an attribute-change vector indicating a path from (a) a first point that is located in an m-dimensional space and that corresponds to the current value of the risk assessment score to (b) a second point in the m-dimensional space and that corresponds to the target value of the risk assessment score, wherein applying a risk assessment function to a sum of the input attribute vector and the attribute-change vector outputs the target value of the risk assessment score, wherein computing the attribute-change vector comprises:
determining, based on statistics computed from the particular cluster to which the input attribute vector is assigned, a requirement on movement from the first point to the second point, and
selecting the attribute-change vector that complies with the requirement;

generating, using the computed attribute-change vector, the recommendation for modifying the current value of the risk assessment score to the target value of the risk assessment score; and
transmitting, to the user device, the recommendation in response to the request for the recommendation for use in improving the current value of the risk assessment score to the target value of the risk assessment score which causes the host system to modify a functionality of an online interface by rearranging a layout of the online interface to switch from presenting the first type of interactive user experience to presenting the second type of interactive user experience.

2. The method of claim 1 , wherein:
the risk assessment function defines a surface within the m-dimensional space, and
the path complying with the requirement comprises the path being a shortest path along the surface between the first point and the second point.

3. The method of claim 1 , wherein:
the statistics computed from the particular cluster comprise a precision matrix for the particular cluster and a mean vector, and
the requirement comprises minimizing an objective function subject to a score change constraint that corresponds to the target value of the risk assessment score, the objective function is computed from the precision matrix, the mean vector, and the attribute-change vector.

4. The method of claim 1 , wherein
the statistics computed from the particular cluster comprise a precision matrix for the particular cluster and a mean vector; and
the requirement comprises maximizing a score change objective function subject to a constraint that is computed from the precision matrix, the mean vector, and the attribute-change vector.

5. The method of claim 1 , wherein computing the attribute-change vector further comprises determining that the attribute-change vector further complies with at least one constraint, the at least one constraint comprising:
an integer constraint requiring a first element of the attribute-change vector to have an integer value,
an auto-increment constraint requiring a second element of the attribute-change vector to increase over a time period, the time period defined by a first time value associated with the first point and a second time value associated with the second point,
a time constraint preventing a third element of the attribute-change vector from changing over the time period, or
an autoencoder constraint applied to the sum of the input attribute vector and the attribute-change vector.

6. The method of claim 1 , wherein generating the recommendation comprises:
generating explanatory data indicating an impact of each value in the attribute-change vector on modifying the current value of the risk assessment score to the target value of the risk assessment score.

7. A non-transitory computer-readable storage medium having program code that is executable by a processor device to cause a computing device to perform operations, the operations comprising:
accessing (a) an input attribute vector having input attribute values of an entity and (b) clusters of points in an n-dimensional space, wherein the clusters of points are respectively defined by historical attribute vectors;
assigning the input attribute vector to a particular cluster based on a similarity measure between the particular cluster and an input point defined by the input attribute vector;
computing an attribute-change vector indicating a path from (a) a first point that is located in an m-dimensional space and that corresponds to a first value of a risk assessment score computed from the input attribute values of the entity to (b) a second point in the m-dimensional space and that corresponds to a second value of the risk assessment score, wherein applying a risk assessment function to a sum of the input attribute vector and the attribute-change vector outputs the second value of the risk assessment score, wherein computing the attribute-change vector comprises:
determining, based on statistics computed from the particular cluster to which the input attribute vector is assigned, a requirement on movement from the first point to the second point, and
selecting the attribute-change vector that complies with the requirement;

generating, using the computed attribute-change vector, a recommendation for modifying the first value of the risk assessment score to the second value of the risk assessment score, the risk assessment score used to control access by a computing device associated with an entity to an interactive computing environment hosted by a host system, wherein the host system is configured to present a first type of interactive user experience to a first type of user group associated with the first value of the risk assessment score and present to a second type of interactive user experience to a second type of user group associated with the second value of the risk assessment score; and
causing the recommendation to be transmitted in response to a request for the recommendation for use in improving the first value of the risk assessment score to the second value of the risk assessment score which causes the host system to modify a functionality of an online interface by rearranging a layout of the online interface to switch from presenting the first type of interactive user experience to presenting the second type of interactive user experience.

8. The non-transitory computer-readable storage medium of claim 7 , wherein the operations further comprise:
updating the input attribute vector by adding the attribute-change vector to the input attribute vector;
assigning the updated input attribute vector to a s

Assignees

• Equifax Inc

Inventors

• Miller Stephen
• Jordan Lewis
• Turner Matthew
• Day Mark
• Joshua Allan

Classifications

• G06N3/082

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

• G06N3/09

  Supervised learning · CPC title

• G06N3/0499

  Feedforward networks · CPC title

• G06N3/0455

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

• G06Q10/0635 Primary

  Risk analysis of enterprise or organisation activities · CPC title