# US 2025/0094595 A1 · Artificial intelligence-based automated method for restoring mask rom firmware binary and apparatus for the same

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

- Number: US 2025/0094595 A1
- Document ID: US20250094595A1
- Status: Published application
- Kind: A1
- Application number: 18/418,441
- Filed: 2024-01-22
- Published: 2025-03-20
- Earliest claimed: 2023-09-18
- Assignees: Electronics and Telecommunications Research Institute
- Applicants: Electronics And Telecommunications Research Institute
- Inventors: Dae-Won KIM; Sang-Su Lee; Yong-Je Choi
- Web page: https://patentgrep.com/patent/US20250094595A1

## CPC

- G06F 21/572: {Secure firmware programming, e.g. of basic input output system [BIOS]}
- G06F 21/64: Protecting data integrity, e.g. using checksums, certificates or signatures
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Disclosed herein are an Artificial Intelligence (AI)-based automated method for restoring a mask Read-Only Memory (ROM) firmware binary and an apparatus for the same. The AI-based automated method is performed by an AI-based automated apparatus for restoring a mask ROM firmware binary, and includes generating a bit-order list by inputting bit information and a parameter that are input for restoration to a pre-trained generation model, and restoring a firmware binary based on the bit-order list.

## Claims

### Claim 1

An Artificial Intelligence (AI)-based automated method for restoring a mask Read-Only Memory (ROM) firmware binary, the AI-based automated method being performed by an AI-based automated apparatus for restoring a mask ROM firmware binary, the AI-based automated method comprising:
generating a bit-order list by inputting bit information and a parameter that are input for restoration to a pre-trained generation model; and
restoring a firmware binary based on the bit-order list.

### Claim 2 (depends on 1)

The AI-based automated method of claim 1, wherein the pre-trained generation model corresponds to a model trained in consideration of an independent variable and a dependent variable of chip data and a bit-order list generation policy.

### Claim 3 (depends on 2)

The AI-based automated method of claim 2, wherein the bit-order list generation policy is established based on a characteristic of the parameter associated with generation of the bit-order list.

### Claim 4 (depends on 1)

The AI-based automated method of claim 1, further comprising:
performing system verification on the firmware binary based on string detection, executable code detection, and loop code detection.

### Claim 5 (depends on 4)

The AI-based automated method of claim 4, further comprising:
updating the pre-trained generation model by utilizing a system verification result based on the system verification and a result of user verification on the firmware binary as feedback information.

### Claim 6

An Artificial Intelligence (AI)-based automated apparatus for restoring a mask Read-Only Memory (ROM) firmware binary, comprising:
a processor configured to generate a bit-order list by inputting bit information and a parameter that are input for restoration to a pre-trained generation model, and restore a firmware binary based on the bit-order list; and
a memory configured to store the pre-trained generation model.

### Claim 7 (depends on 6)

The AI-based automated apparatus of claim 6, wherein the pre-trained generation model corresponds to a model trained in consideration of an independent variable and a dependent variable of chip data and a bit-order list generation policy.

### Claim 8 (depends on 7)

The AI-based automated apparatus of claim 7, wherein the bit-order list generation policy is established based on a characteristic of the parameter associated with generation of the bit-order list.

### Claim 9 (depends on 6)

The AI-based automated apparatus of claim 6, wherein the processor is configured to perform system verification on the firmware binary based on string detection, executable code detection, and loop code detection.

### Claim 10 (depends on 9)

The AI-based automated apparatus of claim 9, wherein the processor is configured to update the pre-trained generation model by utilizing a system verification result based on the system verification and a result of user verification on the firmware binary as feedback information.
