# US 2020/0287900 A1 · Identity verification method and apparatus

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2020/0287900 A1
- Document ID: US20200287900A1
- Status: Published application
- Kind: A1
- Application number: 16/884,526
- Filed: 2020-05-27
- Published: 2020-09-10
- Earliest claimed: 2018-05-28
- Assignees: Alibaba Group Holding Limited
- Applicants: Alibaba Group Holding Limited
- Inventors: Wei WANG; Xiuli ZHAO
- Web page: https://patentgrep.com/patent/US20200287900A1

## CPC

- H04L 63/0876: {based on the identity of the terminal or configuration, e.g. MAC address, hardware or software configuration or device fingerprint}
- H04L 63/105: {Multiple levels of security}
- G06F 21/31: User authentication
- G06F 21/44: Program or device authentication
- H04L 63/205: {involving negotiation or determination of the one or more network security mechanisms to be used, e.g. by negotiation between the client and the server or between peers or by selection according to the capabilities of the entities involved (negotiation of communication capabilities H04L69/24)}
- G06F 2221/2113: Multi-level security, e.g. mandatory access control
- G06N 20/00: Machine learning
- H04L 63/083: {using passwords (cryptographic mechanisms or cryptographic arrangements for entity authentication using a predetermined code H04L9/3226)}
- H04L 63/0861: {using biometrical features, e.g. fingerprint, retina-scan (cryptographic mechanisms or cryptographic arrangements for entity authentication using biological data H04L9/3231)}
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems, methods, and storage media for identity verification are disclosed. An example of an identity verification method comprises: receiving from a computing device an identity verification request for identity verification of a user of the computing device for accessing a service system, wherein the identity verification request comprises device environment information of the computing device; determining a credibility of the computing device according to (i) the device environment information of the computing device and (ii) historical feature information comprising device environment information of a historical device used by the user, wherein the historical device previously logged into the service system or underwent a previous identity verification for accessing the service system; determining, according to the credibility of the computing device, an identity verification policy for performing identity verification on the user; and performing identity verification on the user according to the identity verification policy.

## Claims

### Claim 1

An identity verification method, comprising:
receiving from a computing device an identity verification request for identity verification of a user of the computing device for accessing a service system, wherein the identity verification request comprises device environment information of the computing device, wherein the device environment information comprises one or more of: identity ID information, and WI-FI information, LBS (location-based service) information, and IP (Internet Protocol IP) address;
determining a credibility of the computing device according to (i) the device environment information of the computing device and (ii) historical feature information comprising device environment information of a historical device used by the user, wherein the historical device previously logged into the service system or underwent a previous identity verification for accessing the service system;
determining, according to the credibility of the computing device, an identity verification policy for performing identity verification on the user, wherein the identity verification policy comprises one or more of: an identity verification information type, an identity verification information amount, and an identity verification threshold range; and
performing identity verification on the user according to the identity verification policy.

### Claim 2 (depends on 1)

The method according to claim 1, wherein determining the credibility of the computing device comprises:
determining a feature vector corresponding to the computing device according to the device environment information of the computing device when a type of information comprised in the device environment information matches a preset information type;
inputting the feature vector to a device verification model; and
outputting, from the device verification model based on device environment information of a device, a credibility label of the computing device and a probability that the identity verification request is a credible request.

### Claim 3 (depends on 2)

The method according to claim 2, wherein before inputting the feature vector to the device verification model, the method further comprises:
obtaining historical feature information corresponding to the historical device;
determining, according to a time interval corresponding to the historical feature information, a training set comprising first historical feature information and a verification set comprising second historical feature information; and
training the device verification model by using the first historical feature information, and verifying the device verification model by using the second historical feature information.

### Claim 4 (depends on 3)

The method according to claim 3, wherein determining, according to the time interval corresponding to the historical feature information, the training set comprising first historical feature information and the verification set comprising second historical feature information comprises:
according to the time interval corresponding to the historical feature information, assigning first historical feature information generated at a first time point to the training set, and assigning second historical feature information generated at a second time point to the verification set, wherein the first time point is earlier in time than the second time point.

### Claim 5 (depends on 3)

The method according to claim 3, wherein the device verification model comprises a LR (logistic regression) model and an XGBoost (extreme gradient boosting) model, wherein training the device verification model comprises:
determining, according to the first historical feature information, a first historical feature vector, a second historical feature vector, and a first credibility label of a historical device corresponding to the first historical feature information;
inputting the first historical feature vector to the LR model;
obtaining a trained LR model by adjusting a model parameter of the LR model according to the first credibility label and a first target label output by the LR model;
determining, according to an output of the trained LR model, a first probability that an identity verification request sent by the device corresponding to the first historical feature information is a credible request;
inputting the first probability and the second historical feature vector to the XGBoost model; and
obtaining a trained XGBoost model by adjusting a model parameter of the XGBoost model based on an output of the XGBoost model and the first credibility label, wherein the output of the XGBoost model comprises the first credibility label of the historical device corresponding to the first historical feature information and a probability that an identity verification request sent by the historical device corresponding to the first historical feature information is a credible request.

### Claim 6 (depends on 5)

The method according to claim 5, wherein verifying the device verification model by using the second historical feature information comprises:
determining, according to the second historical feature information, a third historical feature vector, a fourth historical feature vector, and a second credibility label of a historical device corresponding to the second historical feature information;
inputting the third historical feature vector to the trained LR model;
determining, according to the output of the trained LR model, a second probability that an identity verification request sent by the device corresponding to the second historical feature information is a credible request;
inputting the second probability and the fourth historical feature vector to the trained XGBoost model, and obtaining an output of the device verification model by using the trained XGBoost model, wherein the output of the device verification model comprises the second credibility label of the historical device corresponding to the second historical feature information and a third probability that an identity verification request sent by the historical device is a credible request;
obtaining evaluation data of the device verification model by using a binary classification method according to the output of the device verification model and the second credibility label; and
adjusting the model parameter of the LR model and/or the model parameter of the XGBoost model according to the evaluation data until the evaluation data of the device verification model satisfies a preset condition.

### Claim 7 (depends on 6)

The method according to claim 6, wherein the evaluation data of the device verification model comprises at least one of an F1 score and an AUC (area under the curve) value.

### Claim 8 (depends on 5)

The method according to claim 5, wherein the model parameter of the LR model comprises at least one of: a regularization term, a regularization coefficient, and a maximum iteration number.

### Claim 9 (depends on 5)

The method according to claim 5, wherein a number of components in the first historical feature vector is greater than a number of components in the second historical feature vector.

### Claim 10 (depends on 2)

The method according to claim 2, wherein determining the credibility of the computing device comprises:
determining that the computing device is incredible when the type of the information comprised in the device environment information does not match the preset information type.

### Claim 11 (depends on 10)

The method according to claim 10, wherein the preset information type comprises at least one of: an identity ID of a device, and a WI-FI connection, a LBS (location-based service), and an IP (Internet Protocol) address corresponding to the device.

### Claim 12 (depends on 1)

The method according to claim 1, wherein the historical device comprises at least one of: a historical login device and a historical identity verification device.

### Claim 13 (depends on 1)

The method according to claim 1, wherein determining, according to the credibility of the computing device, the identity verification policy comprises:
determining, according to the credibility of the computing device and a correspondence between the identity verification policy and a credibility interval, the identity verification policy, wherein:
a risk level matching the identity verification policy is in a negative relationship with the credibility of the computing device.

### Claim 14 (depends on 13)

The method according to claim 13, wherein:
when the identity verification policy comprises the identity verification information type, a security level of the identity verification information type in the identity verification policy is in a negative relationship with the credibility of the computing device;
when the identity verification policy comprises the identity verification information amount, the identity verification information amount in the identity verification policy is in a negative relationship with the credibility of the computing device; and
when the identity verification policy comprises the identity verification threshold range, the identity verification threshold range in the identity verification policy is in a positive relationship with the credibility of the computing device.

### Claim 15 (depends on 1)

The method according to claim 1, wherein performing the identity verification on the user according to the identity verification policy comprises:
determining, according to the identity verification policy, an identity verification process, information required for the identity verification, and an identity verification threshold range of the information required for the identity verification;
prompting, according to the identity verification process, the computing device to report the information required for the identity verification;
receiving the information required for the identity verification from the computing device; and
performing the identity verification according to the identity verification threshold range of the information required for the identity verification.

### Claim 16

A non-transitory computer-readable storage medium storing instructions executable by one or more processors, wherein execution of the instructions causes the one or more processors to perform operations comprising:
receiving from a computing device an identity verification request for identity verification of a user of the computing device for accessing a service system, wherein the identity verification request comprises device environment information of the computing device, wherein the device environment information comprises one or more of: identity ID information, and WI-FI information, LBS (location-based service) information, and IP (Internet Protocol IP) address;
determining a credibility of the computing device according to (i) the device environment information of the computing device and (ii) historical feature information comprising device environment information of a historical device used by the user, wherein the historical device previously logged into the service system or underwent a previous identity verification for accessing the service system;
determining, according to the credibility of the computing device, an identity verification policy for performing identity verification on the user, wherein the identity verification policy comprises one or more of: an identity verification information type, an identity verification information amount, and an identity verification threshold range; and
performing identity verification on the user according to the identity verification policy.

### Claim 17 (depends on 16)

The storage medium of claim 16, wherein determining the credibility of the computing device comprises:
determining a feature vector corresponding to the computing device according to the device environment information of the computing device when a type of information comprised in the device environment information matches a preset information type;
inputting the feature vector to a device verification model; and
outputting, from the device verification model based on device environment information of a device, a credibility label of the computing device and a probability that the identity verification request is a credible request.

### Claim 18 (depends on 17)

The storage medium of claim 17, wherein before inputting the feature vector to the device verification model, the operations further comprise:
obtaining historical feature information corresponding to the historical device;
determining, according to a time interval corresponding to the historical feature information, a training set comprising first historical feature information and a verification set comprising second historical feature information; and
training the device verification model by using the first historical feature information, and verifying the device verification model by using the second historical feature information.

### Claim 19 (depends on 18)

The storage medium of claim 18, wherein determining, according to the time interval corresponding to the historical feature information, the training set comprising first historical feature information and the verification set comprising second historical feature information comprises:
according to the time interval corresponding to the historical feature information, assigning first historical feature information generated at a first time point to the training set, and assigning second historical feature information generated at a second time point to the verification set, wherein the first time point is earlier in time than the second time point.

### Claim 20

A system, comprising one or more processors and one or more non-transitory computer-readable storage media storing instructions executable by the one or more processors, wherein execution of the instructions causes the one or more processors to perform operations comprising:
receiving from a computing device an identity verification request for identity verification of a user of the computing device for accessing the system, wherein the identity verification request comprises device environment information of the computing device, wherein the device environment information comprises one or more of: identity ID information, and WI-FI information, LBS (location-based service) information, and IP (Internet Protocol IP) address;
determining a credibility of the computing device according to (i) the device environment information of the computing device and (ii) historical feature information comprising device environment information of a historical device used by the user, wherein the historical device previously logged into the system or underwent a previous identity verification for accessing the system;
determining, according to the credibility of the computing device, an identity verification policy for performing identity verification on the user, wherein the identity verification policy comprises one or more of: an identity verification information type, an identity verification information amount, and an identity verification threshold range; and
performing identity verification on the user according to the identity verification policy.
