# US 2026/0291516 A1 · Electronic device, method, and non-transitory computer-readable recording medium for protecting artificial intelligence model

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

- Number: US 2026/0291516 A1
- Document ID: US20260291516A1
- Status: Published application
- Kind: A1
- Application number: 19/676,287
- Filed: 2026-05-13
- Published: 2026-09-24
- Earliest claimed: 2024-02-19
- Assignees: Samsung Electronics Co., Ltd.
- Applicants: Samsung Electronics Co., Ltd.
- Inventors: Jungbae KIM; Mooyoung KIM; Euntaik LEE; Dohyoung KIM; Seungjin KIM; Hyeonsu LEE; Gyusung CHO
- Web page: https://patentgrep.com/patent/US20260291516A1

## CPC

- G06N 3/0495: Quantised networks; Sparse networks; Compressed networks
- H03M 7/3077: {Sorting}
- G06F 12/023: {Free address space management}
- G06F 12/1433: {for a module or a part of a module}
- G06F 21/6218: {to a system of files or objects, e.g. local or distributed file system or database}
- G06F 21/78: to assure secure storage of data (address-based protection against unauthorised use of memory G06F12/14; record carriers for use with machines and with at least a part designed to carry digital markings G06K19/00)
- G06N 3/063: using electronic means
- H03M 7/6005: {Decoder aspects}
- H03M 7/6011: {Encoder aspects}
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/047: Probabilistic or stochastic networks
- G06N 3/088: Non-supervised learning, e.g. competitive learning
- G06N 3/09: Supervised learning
- G06N 3/092: Reinforcement learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H03M: CODING; DECODING; CODE CONVERSION IN GENERAL (using fluidic means F15C4/00; optical analogue/digital converters G02F7/00; coding, decoding or code conversion, specially adapted for particular applications, see the relevant subclasses, e.g. G01D, G01R, G06F, G06T, G09G, G10L, G11B, G11C, H04B, H04L, H04M, H04N; ciphering or deciphering for cryptography or other purposes involving the need for secrecy G09C)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Disclosed is an electronic device. The electronic device may store weights related to an artificial intelligence (AI) model in a non-volatile memory. The electronic device may compress, by means of a compressor, some weights related to a designated operation among the weights stored in the non-volatile memory on the basis of a load request for the designated operation, and load the compressed weights to a first volatile memory. The electronic device may decompress the some weights compressed and loaded to the first volatile memory by means of a decompressor and load the decompressed weights to a second volatile memory. The electronic device may perform the designated operation on the basis of the some weights decompressed and loaded to the second volatile memory.

## Claims

### Claim 1

An electronic device comprising:
a compressor compressing data based on a compression algorithm;
a decompressor decompressing data based on a decompression algorithm corresponding to the compression algorithm;
at least one processor;
non-volatile memory storing instructions;
a first volatile memory; and
a second volatile memory,
wherein the instructions, are loaded into the first volatile memory, and, when executed by the at least one processor, individually or collectively, cause the electronic device to:
store, in the non-volatile memory, weights related to an artificial intelligence, artificial intelligence (AI), model,
based on a load request for a specified operation, compress, through the compressor, partial weights related to the specified operation, wherein the partial weights correspond to some of the weights among the weights stored in the non-volatile memory, and load the compressed partial weights into the first volatile memory,
decompress, through the decompressor, the partial weights with a compressed state, loaded into the first volatile memory, and load the decompressed partial weights into the second volatile memory, and
perform the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.

### Claim 2 (depends on 1)

The electronic device of claim 1,
wherein the compression algorithm includes an algorithm that changes an order of at least portion of the data or encodes the at least portion of the data while maintaining size of the data, and
wherein the decompression algorithm includes an algorithm that restores the order of at least portion of the data or decodes the at least portion of the data while maintaining size of the data.

### Claim 3 (depends on 1)

The electronic device of claim 1, wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
as at least a portion of an operation of compression through the compressor, change an order of the partial weights or encode the partial weights; and
as at least a portion of an operation of decompression through the decompressor, restore the order of the partial weights or decode the partial weights.

### Claim 4 (depends on 1)

The electronic device of claim 1,
wherein the at least one processor comprises a neural processing unit (NPU) and
wherein the NPU comprises the second volatile memory.

### Claim 5 (depends on 1)

The electronic device of claim 1, wherein the decompressor is included in the at least one processor.

### Claim 6 (depends on 1)

The electronic device of claim 1,
wherein the at least one processor comprises a central processing unit (CPU), a neural processing unit, NPU, and a graphic processing unit (GPU),
wherein the instructions, when executed by the CPU, cause the electronic device to:
store the weights in the non-volatile memory,
compress the partial weights, and load the compressed partial weights into the first volatile memory, and
decompress the partial weights, and load the decompressed partial weights into the second volatile memory, and
wherein the instructions, when executed by the NPU or the GPU, cause the electronic device to:
perform the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.

### Claim 7 (depends on 1)

The electronic device of claim 1, comprising:
communication circuitry,
wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
obtain, from a server through the communication circuitry, the AI model including the weights, and
store the obtained AI model in the non-volatile memory.

### Claim 8 (depends on 1)

The electronic device of claim 1,
wherein the non-volatile memory comprises a general area and a protected area, and
wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
determine an area storing the partial weights among the general area and the protected area, based on the load request for the specified operation,
in case that the partial weights are stored in the general area, compress, through the compressor, the partial weights, and load the compressed partial weights into the first volatile memory, and
in case that the partial weights are stored in the protected area, bypass the compressor, and load the partial weights into the first volatile memory, such that the partial weights are not compressed through the compressor.

### Claim 9 (depends on 8)

The electronic device of claim 8, wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
in case that the partial weights are stored in the general area, compress, through the compressor, the partial weights, and load the compressed partial weights into the first volatile memory; and
in case that the partial weights are stored in the protected area, bypass the compressor, and load the partial weights into the first volatile memory, such that the partial weights are not compressed through the compressor.

### Claim 10

An electronic device comprising:
communication circuitry;
a compressor compressing data based on a compression algorithm;
a decompressor decompressing data based on a decompression algorithm corresponding to the compression algorithm;
at least one processor;
non-volatile memory storing instructions;
a first volatile memory; and
a second volatile memory,
wherein the instructions, are loaded into the first volatile memory, and, when executed by the at least one processor, individually or collectively, cause the electronic device to:
obtain, from a server through the communication circuitry, an artificial intelligence, artificial intelligence (AI), model including weights,
based on a store request for the weights, compress, through the compressor, the weights included in the AI model, and store the compressed weights in the non-volatile memory,
load, into the first volatile memory, partial weights with a compressed state related to a specified operation among the weights, wherein the partial weights correspond to some of the weights among the weights stored in the non-volatile memory, the partial weights in the compressed state being stored in the non-volatile memory,
decompress, through the decompressor, the partial weights with the compressed state loaded into the first volatile memory, and load the decompressed partial weights into the second volatile memory, and
perform the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.

### Claim 11 (depends on 10)

The electronic device of claim 10,
wherein the compression algorithm includes an algorithm that changes an order of at least portion of the data or encodes the at least portion of the data while maintaining size of the data, and
wherein the decompression algorithm includes an algorithm that restores the order of the at least portion of the data or decodes the at least portion of the data while maintaining size of the data.

### Claim 12 (depends on 10)

The electronic device of claim 10, wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
as at least a portion of an operation of compression through the compressor, changes the order of the partial weights or encodes the partial weights; and
as at least a portion of an operation of decompression through the decompressor, restore the order of the partial weights or decodes the partial weights.

### Claim 13 (depends on 10)

The electronic device of claim 10,
wherein the at least one processor comprises a neural processing unit (NPU), and
wherein the NPU comprises the second volatile memory.

### Claim 14 (depends on 10)

The electronic device of claim 10, wherein the decompressor is included in the at least one processor.

### Claim 15 (depends on 10)

The electronic device of claim 10,
wherein the at least one processor comprises a central processing unit (CPU), a neural processing unit, NPU, and a graphic processing unit (GPU),
wherein the instructions, when executed by the CPU, cause the electronic device to:
obtain the AI model,
compress the weights and store the compressed weights in the non-volatile memory,
load the partial weights into the first volatile memory, and
decompress the partial weights, and load the decompressed partial weights into the second volatile memory, and
wherein the instructions, when executed by the NPU or the GPU, cause the electronic device to:
perform the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.

### Claim 16 (depends on 10)

The electronic device of claim 10,
wherein the non-volatile memory comprises a general area and a protected area, and
wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
determine an area storing the partial weights among the general area and the protected area, based on the load request for the specified operation,
in case that the partial weights are stored in the general area, compress, through the compressor, the partial weights, and load the compressed partial weights into the first volatile memory, and
in case that the partial weights are stored in the protected area, bypass the compressor, and load the partial weights into the first volatile memory, such that the partial weights are not compressed by the compressor.

### Claim 17 (depends on 16)

The electronic device of claim 16, wherein the instructions, when executed by the at least one processor, individually or collectively, cause the electronic device to:
in case that the partial weights are stored in the general area, compress, through the compressor, the partial weights, and load the compressed partial weights into the first volatile memory; and
in case that the partial weights are stored in the protected area, bypass the compressor, and load the partial weights into the first volatile memory, such that the partial weights are not compressed by the compressor.

### Claim 18

A method of an electronic device comprising a compressor compressing data based on a compression algorithm, a decompressor decompressing data based on a decompression algorithm corresponding to the compression algorithm, non-volatile memory storing instructions, a first volatile memory, and a second volatile memory, the method comprising:
storing, in the non-volatile memory, weights related to an artificial intelligence, artificial intelligence (AI), model;
based on a load request for a specified operation, compressing, through the compressor, partial weights related to the specified operation, wherein the partial weights correspond to some of the weights among the weights stored in the non-volatile memory, and load the compressed partial weights into the first volatile memory;
decompressing, through the decompressor, the partial weights with a compressed state, loaded into the first volatile memory, and load the decompressed partial weights into the second volatile memory; and
performing the specified operation based on the partial weights with a decompressed state, loaded into the second volatile memory.

### Claim 19 (depends on 18)

The method of claim 18,
wherein the compression algorithm includes an algorithm that changes an order of at least portion of the data or encodes the at least portion of the data while maintaining size of the data, and
wherein the decompression algorithm includes an algorithm that restores the order of at least portion of the data or decodes the at least portion of the data while maintaining size of the data.

### Claim 20 (depends on 18)

The method of claim 18, comprising:
as at least a portion of an operation of compression through the compressor, changing an order of the partial weights or encode the partial weights; and
as at least a portion of an operation of decompression through the decompressor, restoring the order of the partial weights or decode the partial weights.
