View source for 18488497. APPARATUS AND METHOD WITH ENCRYPTED DATA NEURAL NETWORK OPERATION simplified abstract (Samsung Electronics Co., Ltd.)

=APPARATUS AND METHOD WITH ENCRYPTED DATA NEURAL NETWORK OPERATION=


==Organization Name==

[[:Category:Samsung Electronics Co., Ltd.|Samsung Electronics Co., Ltd.]]


[[Category:Samsung Electronics Co., Ltd.]]

==Inventor(s)==

[[:Category:Jong-Seon No of Seoul (KR)|Jong-Seon No of Seoul (KR)]][[Category:Jong-Seon No of Seoul (KR)]]

[[:Category:Junghyun Lee of Seoul (KR)|Junghyun Lee of Seoul (KR)]][[Category:Junghyun Lee of Seoul (KR)]]

[[:Category:Yongjune Kim of Daegu (KR)|Yongjune Kim of Daegu (KR)]][[Category:Yongjune Kim of Daegu (KR)]]

[[:Category:Joon-Woo Lee of Seoul (KR)|Joon-Woo Lee of Seoul (KR)]][[Category:Joon-Woo Lee of Seoul (KR)]]

[[:Category:Young Sik Kim of Gwangju (KR)|Young Sik Kim of Gwangju (KR)]][[Category:Young Sik Kim of Gwangju (KR)]]

[[:Category:Eunsang Lee of Seoul (KR)|Eunsang Lee of Seoul (KR)]][[Category:Eunsang Lee of Seoul (KR)]]

==APPARATUS AND METHOD WITH ENCRYPTED DATA NEURAL NETWORK OPERATION - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18488497 titled 'APPARATUS AND METHOD WITH ENCRYPTED DATA NEURAL NETWORK OPERATION

==Simplified Explanation==
The patent application describes an apparatus and method for operating a neural network with encrypted data. The apparatus generates a target approximate polynomial to approximate a neural network operation using a determined target approximation region.

* One or more processors execute instructions to generate a target approximate polynomial approximating a portion of a neural network model.
* The target approximate polynomial is based on a first approximate polynomial, parameters, maximum and minimum values of input data.
* The apparatus generates a neural network operation result using the target approximate polynomial and input data.

==Key Features and Innovation==
* Encryption of data for neural network operation.
* Generation of target approximate polynomial to approximate neural network operation.
* Use of determined target approximation region for accurate approximation.

==Potential Applications==
The technology can be applied in secure data processing, machine learning, and artificial intelligence systems.

==Problems Solved==
* Secure operation of neural networks with encrypted data.
* Accurate approximation of neural network operations.

==Benefits==
* Enhanced data security.
* Improved accuracy in neural network operations.

==Commercial Applications==
Title: Secure Neural Network Operation with Encrypted Data
This technology can be used in industries such as cybersecurity, finance, healthcare, and autonomous vehicles for secure and accurate data processing.

==Prior Art==
Readers can explore prior research on encrypted data processing in neural networks and approximation techniques in machine learning.

==Frequently Updated Research==
Stay updated on advancements in encrypted data processing in neural networks and neural network approximation methods.

==Questions about Secure Neural Network Operation with Encrypted Data==
===How does encrypted data processing impact the performance of neural networks?===
Encrypted data processing can enhance data security but may introduce computational overhead.

===What are the potential challenges in implementing encrypted data processing in neural networks?===
Challenges may include key management, performance trade-offs, and compatibility with existing systems.


==Original Abstract Submitted==

An apparatus and method with encrypted data neural network operation is provided. The apparatus includes one or more processors configured to execute instructions and one or more memories storing the instructions, wherein the execution of the instructions by the one or more processors configures the one or more processors to generate a target approximate polynomial, approximating a neural network operation, of a portion of a neural network model, using a determined target approximation region, for the target approximate polynomial, based on a first approximate polynomial generated based on parameters corresponding to a generation of the first approximate polynomial, a maximum value of input data to the portion of the neural network layer, and a minimum value of the input data, and generate a neural network operation result using the target approximate polynomial and the input data.



[[Category:G06N3/048]]
[[Category:G06N3/08]]
[[Category:CPC_G06N3/048]]