# US 2019/0319978 A1 · Determining Exploit Prevention using Machine Learning

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

- Number: US 2019/0319978 A1
- Document ID: US20190319978A1
- Status: Published application
- Kind: A1
- Application number: 15/952,656
- Filed: 2018-04-13
- Published: 2019-10-17
- Earliest claimed: 2018-04-13
- Assignees: Webroot Inc.
- Applicants: Webroot Inc.
- Inventors: Jonathan Giffard
- Web page: https://patentgrep.com/patent/US20190319978A1

## CPC

- H04L 63/1433: {Vulnerability analysis}
- G06F 16/2255: {Hash tables}
- 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

Examples of the present disclosure describe systems and methods for determining exploit prevention software settings using machine learning. In aspects, exploit prevention software may be used to identify processes executing on a computing device. Metadata for the identified processes may be determined and transmitted to a machine learning system. The machine learning system may use an exploit prevention model to determine exploit prevention configuration settings for each of the processes, and may transmit the configuration setting to the computing device. The computing device may implement the configuration settings to protect the processes and monitor the stability of the protected processes as they execute. The computing device may transmit the stability data to the machine-learning system. The machine-learning system may then modify the exploit prevention model based on the stability data.

## Claims

### Claim 1

A system comprising:
at least one processor; and
memory coupled to the at least one processor, the memory comprising computer executable instructions that, when executed by the at least one processor, performs a method comprising:
obtaining process information for a list of processes running on a computing device;
transmitting the process information to a machine learning system for exploit prevention;
applying, to the computing device, received process configuration settings from the machine learning system;
monitoring stability data for one or more processes corresponding to the applied process configuration settings; and
transmitting the stability data to the machine learning system.

### Claim 2 (depends on 1)

The system of claim 1, wherein the process information comprises metadata for the one or more processes, wherein the metadata corresponds to at least one of a process identifier, methods or functions used by the one or more processes, and objects used during execution of the one or more processes.

### Claim 3 (depends on 1)

The system of claim 1, wherein obtaining the process information comprises generating one or more hash values for the list of processes, and wherein the one or more hash values are transmitted to the machine learning system.

### Claim 4 (depends on 1)

The system of claim 1, wherein the process configuration settings were generated using the process information, and comprise at least one configuration setting for each process identified in the process information.

### Claim 5 (depends on 1)

The system of claim 1, wherein applying the received process configuration settings comprises at least one of: restarting the computing device, restarting one or more processes, modifying previously-configured configuration settings, and recompiling a file.

### Claim 6 (depends on 1)

The system of claim 1, wherein monitoring the stability data comprises at least one of:
collecting system health statistics for the computing device, collecting performance statistics for the one or more processes, evaluating process status, evaluating event files, and evaluating checkpoints in executing code.

### Claim 7 (depends on 1)

The system of claim 1, wherein the method further comprises:
evaluating the stability data; and
when the stability data indicates a decrease in the stability of one or more processes, reducing, by the computing device, exploit protection provided by the process configuration settings.

### Claim 8

A system comprising:
at least one processor; and
memory coupled to the at least one processor, the memory comprising computer executable instructions that, when executed by the at least one processor, performs a method comprising:
receiving, at a machine learning system, process information for a list of processes running on a computing device;
based on the process information, determining process configuration settings for one or more processes in the list of processes;
transmitting the process configuration settings to the computing device;
receiving stability data from the computing device, wherein the stability data represent process stability of the one or more processes in the list of processes; and
adjusting a machine learning model using the stability data.

### Claim 9 (depends on 8)

The system of claim 8, wherein the process information comprises one or more hash values and metadata for one or more processes in the list of processes.

### Claim 10 (depends on 9)

The system of claim 9, wherein determining the process configuration settings comprises comparing the hash values to one or more data sources, wherein the one or more data sources comprise a set of hash values and corresponding process configuration settings.

### Claim 11 (depends on 10)

The system of claim 10, wherein, when the comparison determines a match between the one or more hash values and the set of hash values, the process configuration settings corresponding to the matched hash value in the set of hash values is assigned to the respective hash value in the one or more hash values.

### Claim 12 (depends on 8)

The system of claim 8, wherein determining the process configuration settings comprises providing the process information to a machine learning model configured to generate the process configuration settings.

### Claim 13 (depends on 12)

The system of claim 12, wherein generating the process configuration settings comprises:
determining a first hash value in the process information;
using an approximation algorithm to determine at least a second hash value similar to the first hash value; and
assigning a process configuration setting associated with the second hash value to the first hash value.

### Claim 14 (depends on 8)

The system of claim 8, wherein the determined process configuration settings are presented on a user interface, wherein the user interface is configured to receive a selection of one or more of the determined process configuration settings.

### Claim 15 (depends on 8)

The system of claim 8, wherein the stability data corresponds to at least one of: process uptime, resource utilization, task execution speed, task completion satisfaction, and security events.

### Claim 16 (depends on 8)

The system of claim 8, wherein the method further comprises evaluating the stability data, wherein evaluating the stability data comprises at least one of: comparing the stability data to previously received stability data, comparing the stability data to benchmark stability data, and performing trend analysis on the stability data.

### Claim 17 (depends on 16)

The system of claim 16, wherein evaluating the stability data further comprises:
generating one or more metrics for the stability data; and
and applying the one or more metrics to the stability data, wherein the one or more metrics indicate at least one of a level of stability, a priority, and a recency for a process configuration setting.

### Claim 18 (depends on 8)

The system of claim 8, wherein adjusting the machine learning model comprises training the machine learning model to provide improved process configuration settings.

### Claim 19 (depends on 18)

The system of claim 18, wherein improved process configuration settings reduce at least one of: a number of process crashes and a number of process errors generated.

### Claim 20

A system comprising:
at least one processor; and
memory coupled to the at least one processor, the memory comprising computer executable instructions that, when executed by the at least one processor, performs a method comprising:
obtaining process information for a list of processes running on a computing device;
transmitting the process information to a machine learning system for exploit prevention;
based on the process information, determining process configuration settings for one or more processes in the list of processes;
transmitting the process configuration settings to the computing device;
applying, to the computing device, the process configuration settings;
monitoring stability data for the one or more processes corresponding to the applied process configuration settings, wherein the stability data represent process stability of the one or more processes;
transmitting the stability data to the machine learning system; and
adjusting a machine learning model using the stability data.
