# US 2024/0406182 A1 · Utilizing patterns in operation of user input devices for user authenticity verification

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

- Number: US 2024/0406182 A1
- Document ID: US20240406182A1
- Status: Published application
- Kind: A1
- Application number: 18/329,185
- Filed: 2023-06-05
- Published: 2024-12-05
- Earliest claimed: 2023-06-05
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: PAUL LLAMAS VIRGEN; JONATHAN OROZCO CERVANTES; ANDRES ALEJANDRE JIMENEZ; ALAN VERDUGO MUÑOZ
- Web page: https://patentgrep.com/patent/US20240406182A1

## CPC

- H04L 63/102: {Entity profiles}
- G06F 21/316: {by observing the pattern of computer usage, e.g. typical user behaviour}
- H04L 63/1425: {Traffic logging, e.g. anomaly detection}
- G06F 2221/033: Test or assess software
- G06F 2221/2133: Verifying human interaction, e.g., Captcha
- G06N 20/00: Machine learning
- H04L 63/101: {Access control lists [ACL]}
- 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

A computer system includes: a processor; a memory in communication with the processor; multiple user input devices; and a security tool to capture data defining usage patterns of an authorized user operating the multiple user input devices to prepare an authorized user profile and to implement security measures when usage patterns of a current user differ from the usage patterns of the authorized user from the profile. The security tool calls a Machine Learning Model (MLM) that is trained on the usage patterns of the authorized user from the profile, the MLM predicting for the security tool whether usage patterns of the current user indicate that the current user is not the authorized user based on the training of the MLM.

## Claims

### Claim 1

A computer system comprising:
a processor;
a memory in communication with the processor;
multiple user input devices; and
a security tool to capture data defining usage patterns of an authorized user operating the multiple user input devices to prepare an authorized user profile and to implement security measures when usage patterns of a current user differ from the usage patterns of the authorized user from the profile;
wherein the security tool calls a Machine Learning Model (MLM) that is trained on the usage patterns of the authorized user from the profile, the MLM predicting for the security tool whether usage patterns of the current user indicate that the current user is not the authorized user based on the training of the MLM.

### Claim 2 (depends on 1)

The computer system of claim 1, wherein the multiple user input devices comprise both a keyboard and a pointing device of the computer system.

### Claim 3 (depends on 1)

The computer system of claim 1, further comprising a network interface, wherein the security tool calls the MLM via the network interface with the MLM hosted on the network.

### Claim 4 (depends on 1)

The computer system of claim 1, further comprising a network interface for connecting to Internet of Things (IoT) devices, the security tool recording in the authorized user profile a list of IoT devices associated with the authorized user.

### Claim 5 (depends on 4)

The computer system of claim 4, wherein a confidence level of the prediction of the MLM that the current user is or is not an authorized user is adjusted based on current detection of the IoT devices listed in the authorized user profile.

### Claim 6 (depends on 1)

The computer system of claim 1, wherein:
the MLM reports to the security tool an accuracy threshold quantifying how closely the usage patterns of the current user match the usage patterns of the authorized user from the profile; and
the security tool implements different security measures based on the accuracy threshold reported.

### Claim 7 (depends on 6)

The computer system of claim 6, wherein, based on the accuracy threshold indicating that the current user is not confirmed as the authorized user, the security tool presents a puzzle or challenge to the current user, captures additional usage pattern data for the current user responding to the challenge, and provides the additional usage pattern data to the MLM to update the MLM prediction whether the current user is authorized.

### Claim 8 (depends on 6)

The computer system of claim 6, further comprising a network interface for connecting to Internet of Things (IoT) devices, the security tool recording in the authorized user profile a list of IoT devices associated with the authorized user, wherein a confidence level of the prediction of the MLM that the current user is or is not an authorized user is adjusted based on current detection of the IoT devices listed in the authorized user profile.

### Claim 9

A method of providing security for a computer system that includes
a processor,
a memory in communication with the processor,
multiple user input devices, and
a security tool, the method comprising:
with the security tool, capturing data defining usage patterns of an authorized user operating the multiple user input devices to prepare an authorized user profile;
with the security tool, calling a Machine Learning Model (MLM) that is trained on the usage patterns of the authorized user from the profile, the MLM predicting for the security tool whether usage patterns of the current user indicate that the current user is not the authorized user based on the training of the MLM; and
with the security tool, implementing security measures when usage patterns of a current user differ from the usage patterns of the authorized user from the profile causing the MLM to predict that the current user is not the authorized user.

### Claim 10 (depends on 9)

The method of claim 9, wherein the multiple user input devices comprise both a keyboard and a pointing device of the computer system.

### Claim 11 (depends on 9)

The method of claim 9, further comprising calling the MLM via a network interface with the MLM hosted on a network remote from the computer system.

### Claim 12 (depends on 9)

The method of claim 9, the computer system further comprising a network interface for connecting to Internet of Things (IoT) devices, the method further comprising, with the security tool, recording in the authorized user profile a list of IoT devices associated with the authorized user.

### Claim 13 (depends on 12)

The method of claim 12, wherein a confidence level of the prediction of the MLM that the current user is or is not an authorized user is adjusted based on current detection of the IoT devices listed in the authorized user profile.

### Claim 14 (depends on 9)

The method of claim 9, wherein:
the MLM reports to the security tool an accuracy threshold quantifying how closely the usage patterns of the current user match the usage patterns of the authorized user from the profile; and
the security tool implements different security measures based on the accuracy threshold reported.

### Claim 15 (depends on 14)

The method of claim 14, wherein, based on the accuracy threshold indicating that the current user is not confirmed as the authorized user, the security tool presents a puzzle or challenge to the current user, captures additional usage pattern data for the current user responding to the challenge, and provides the additional usage pattern data to the MLM to update the MLM prediction whether the current user is authorized.

### Claim 16 (depends on 14)

The method of claim 14, the computer system comprising a network interface for connecting to Internet of Things (IoT) devices, the security tool recording in the authorized user profile a list of IoT devices associated with the authorized user, wherein a confidence level of the prediction of the MLM that the current user is or is not an authorized user is adjusted based on current detection of the IoT devices listed in the authorized user profile.

### Claim 17

A computer program product comprising a non-transitory computer readable medium storing instructions for a processor, the instructions, when implemented by the processor alone or in combination with other processors, cause the processor to implement a security tool for a computer, wherein:
the security tool captures data defining usage patterns of an authorized user operating multiple user input devices to prepare an authorized user profile;
the security tool calls a Machine Learning Model (MLM) that is trained on the usage patterns of the authorized user from the profile, the MLM predicting for the security tool whether usage patterns of the current user indicate that the current user is not the authorized user based on the training of the MLM; and
the security tool implements security measures when usage patterns of a current user differ from the usage patterns of the authorized user from the profile causing the MLM to predict that the current user is not the authorized user.

### Claim 18 (depends on 17)

The computer program product of claim 17, wherein the security tool calls the MLM via a network interface with the MLM hosted on a network remote from the computer.

### Claim 19 (depends on 17)

The computer program product of claim 17, wherein:
the computer further comprises a network interface for connecting to Internet of Things (IoT) devices,
the security tool records in the authorized user profile a list of IoT devices associated with the authorized user, and
the security tool adjusts a confidence level of the prediction of the MLM that the current user is or is not an authorized user based on current detection of the IoT devices listed in the authorized user profile.

### Claim 20 (depends on 17)

The computer program product of claim 17, wherein the security tool presents a puzzle or challenge to the current user, captures additional usage pattern data for the current user responding to the challenge, and provides the additional usage pattern data to the MLM to update the MLM prediction whether the current user is authorized.
