METHOD AND APPARATUS FOR ENCRYPTING A NEURAL NETWORK MODEL AND A DATA MEDIUM Russian patent published in 2020 - IPC G06F9/00 G06F21/60 G06F21/62 

FIELD: data processing.

SUBSTANCE: present invention relates to a method and an apparatus for encrypting a neural network model. Method includes following steps. Encrypted model file is obtained by encrypting at least part of the model information in the source model file. Initial model file describes the target neural network model. Program code of the model is generated in accordance with the encoded model file. Program code of the model describes the target neural network model. Installation package for application installation (APP) is sent to user device (UE) based on request sent by specified user device (UE). Installation package includes a program code of the model.

EFFECT: technical result consists in improvement of efficiency of obtaining information in a neural network, thereby improving efficiency of the application, due to the fact that it is not necessary to call the original model file and save it on the user device (UE), and in improving the information security internal for the model, due to the fact that the server sends the encrypted model code.

12 cl, 5 dwg