Natural language semantic key generation method based on deep neural network coding

Abstract

The invention discloses a semantic key generation method based on natural language semantic similarity. According to the method, a text with a certain elastic range is used as a password, a text semantic vector is extracted by using a deep neural network model, and the semantic vector is recoded through a deep neural network to generate a key. In the authentication process, the input natural language text does not need to be matched with the original text one by one, and the same key can be generated to complete the authentication process as long as the semantics of the input natural language text is similar to that of the original text. In the process, a user does not need to memorize an original text in a one-word-free manner, and only needs to basically memorize short text semantics and formats, the memories can be described through a natural language, a secret key is generated, and authentication is completed. According to the method, the risk of privacy leakage of biological feature template authentication does not exist, and meanwhile, a user can generate a high-security key (>512 bit length) and the security and the flexibility of the authentication process are improved.

Description

Technical field [0001] The present invention belongs to the field of natural language processing and cryptography, involving Chinese semantics extracting tasks and key generation tasks, involving a semantic key generation method based on natural language semantic simplicity, specifically, based on deep neural network Semantic feature extraction and method of generating a key based on semantic feature. Background technique [0002] With the rapid development of the Internet, users often need to authenticate the user's identity before communicating with the outside world, and can only link user operations with previous operations in the network, so that users can view history. Or for new operations, such as online shopping, funds, etc. [0003] In the current identity authentication technology, common methods can be divided into two categories: one is the user's authentication by the user to complete the personal identity by saving, the commonly used password is a combination of no...

AI Technical Summary

Problems solved by technology

But the disadvantages are also obvious, that is, the biometric authentication method needs to retain the user's biometric template for comparison and authentication, and it is difficult to guarantee user privacy and security.

Images