# US 2022/0046048 A1 · Method and apparatus for predicting attack vulnerability of computer network

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

- Number: US 2022/0046048 A1
- Document ID: US20220046048A1
- Status: Published application
- Kind: A1
- Application number: 17/394,989
- Filed: 2021-08-05
- Published: 2022-02-10
- Earliest claimed: 2020-08-06
- Assignees: Electronics and Telecommunications Research Institute
- Applicants: Electronics and Telecommunications Research Institute
- Inventors: Ki Jong KOO; Dae Sung MOON; Jooyoung LEE; Ik Kyun KIM; Kyungmin PARK; Ho HWANG
- Web page: https://patentgrep.com/patent/US20220046048A1

## CPC

- H04L 63/1433: {Vulnerability analysis}
- G06F 9/455: Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06N 3/0499: Feedforward networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06N 5/046: Forward inferencing; Production systems
- 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

Collecting the topology and asset information of the virtual generated computer network, converting the topology and asset information into a training data set for training the neural network model, training the neural network model based on the training data set, and training A method and apparatus for predicting an attack vulnerability of a computer network through the step of inferring an attack vulnerability of a target computer network using a neural network model are provided.

## Claims

### Claim 1

A method for predicting an attack vulnerability of a target computer network using a neural network model, the method comprising:
collecting topology and asset information of a computer network which is created in a virtual environment;
converting the topology and asset information into a training data set for training the neural network model;
training the neural network model based on the training data set; and
inferring the attack vulnerability of the target computer network using the trained neural network model.

### Claim 2 (depends on 1)

The method of claim 1, wherein
the collecting topology and asset information of the computer network which is created in a virtual environment comprises:
constructing a virtual topology of the computer network in the virtual environment;
calculating a vulnerability score of each of a plurality of hosts included in the constructed virtual topology;
generating a host connection matrix between the plurality of hosts; and
generating an attack path table corresponding to the topology based on the host connection matrix and the vulnerability score of each of the plurality of hosts.

### Claim 3 (depends on 2)

The method of claim 2, wherein
the converting the topology and asset information into a training data set for training the neural network model comprises
generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table.

### Claim 4 (depends on 3)

The method of claim 3, wherein
the generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table comprises
generating host connection information based on a topology table on the topology and the host connection matrix.

### Claim 5 (depends on 4)

The method of claim 4, wherein
the generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table comprises
generating asset connection information based on the asset information and the host connection matrix.

### Claim 6 (depends on 5)

The method of claim 5, wherein
the generating the training data set by preprocessing the connection matrix, the asset information, and the attack path table comprises
generating attack path information corresponding to the topology based on the attack table.

### Claim 7 (depends on 6)

The method of claim 6, further comprising
generating the training data set based on the host connection information, the asset connection information, and the attack path information.

### Claim 8 (depends on 1)

The method of claim 1, wherein
the inferring the attack vulnerability of the target computer network using the trained neural network model comprises:
collecting topology and asset information of the target computer network;
generating an input data set by pre-processing the topology and asset information of the target computer network; and
inferring the attack vulnerability of the target computer network by using the neural network model trained based on the input data set.

### Claim 9

An apparatus for predicting an attack vulnerability of a target computer network using a neural network model, the apparatus comprising:
a processor and a memory, wherein the processor executes a program stored in the memory to perform:
collecting topology and asset information of a computer network which is created in a virtual environment;
converting the topology and asset information into a training data set for training the neural network model;
training the neural network model based on the training data set; and
inferring the attack vulnerability of the target computer network using the trained neural network model.

### Claim 10 (depends on 9)

The apparatus of claim 9, wherein
when the processor performs the collecting topology and asset information of the computer network which is created in a virtual environment, the processor performs:
constructing a virtual topology of the computer network in the virtual environment;
calculating a vulnerability score of each of a plurality of hosts included in the constructed virtual topology;
generating a host connection matrix between the plurality of hosts; and
generating an attack path table corresponding to the topology based on the host connection matrix and the vulnerability score of each of the plurality of hosts.

### Claim 11 (depends on 10)

The apparatus of claim 10, wherein
when the processor the converting the topology and asset information into a training data set for training the neural network model, the processor performs:
generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table.

### Claim 12 (depends on 11)

The apparatus of claim 11, wherein
when the processor performs the generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table, the processor performs
generating host connection information based on a topology table on the topology and the host connection matrix.

### Claim 13 (depends on 12)

The apparatus of claim 12, wherein
when the processor performs the generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table, the processor performs
generating asset connection information based on the asset information and the host connection matrix.

### Claim 14 (depends on 13)

The apparatus of claim 13, wherein
when the processor performs the generating the training data set by preprocessing the host connection matrix, the asset information, and the attack path table, the processor performs
generating attack path information corresponding to the topology based on the attack table.

### Claim 15 (depends on 14)

The apparatus of claim 14, wherein
the processor executes the program to further perform
generating the training data set based on the host connection information, the asset connection information, and the attack path information.

### Claim 16 (depends on 9)

The apparatus of claim 9, wherein
when the processor performs the inferring the attack vulnerability of the target computer network using the trained neural network model, the processor performs:
collecting topology and asset information of the target computer network;
generating an input data set by pre-processing the topology and asset information of the target computer network; and
inferring the attack vulnerability of the target computer network by using the neural network model trained based on the input data set.

### Claim 17 (depends on 9)

The apparatus of claim 9, further comprising:
a user interface to visualize the inferred attack vulnerability.
