Encryption transmission method and device of data information, electronic equipment and storage medium

Abstract

The invention provides a data information encryption transmission method and device and electronic equipment, and the method comprises the steps: obtaining first data information, and inputting the first data information into a secret report language model obtained through pre-training; wherein the confidential language model comprises a base model obtained by pre-training according to an original confidential data set and a plug-in model obtained by updating according to an updated confidential data set; obtaining second data information output after the secret report language model performs first-level encryption and decryption processing on the first data information; and performing second-level encryption and decryption processing on the second data information to obtain third data information, and transmitting the third data information from the first transmission end to the second transmission end. According to the mode, the base model and the plug-in model are mutually combined, the technical effect of convenient updating can be realized, and the plug-in model can be flexibly updated in a scene in which a confidential mode is frequently changed, so that flexible conversion of encryption and decryption modes is realized to flexibly adapt to the confidential scene in which the confidential mode is frequently changed.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and in particular to a method, device, electronic device and storage medium for encrypting and transmitting data information. Background Art

[0002] Information encryption transmission is an important means to ensure data security. In related technologies, information encryption transmission can use encryption algorithms and keys to convert plaintext data into ciphertext data to ensure the confidentiality and integrity of data during transmission. Even if the data is intercepted, the original data cannot be decrypted and read without the key. It can be seen that information encryption transmission ensures the security of data during transmission through encryption algorithms and keys, which is an important measure to protect data confidentiality and integrity.

[0003] However, since there may be special circumstances where keys may be cracked during the communication process, how to further improve the security of encrypted information transmission has become a technical problem that needs to be solved urgently. Summary of the invention

[0004] The present invention provides a method and device for encrypting and transmitting data information, an electronic device, and a computer-readable storage medium.

[0005] In a first aspect, the present disclosure provides a method for encrypting and transmitting data information, which is applied to a first transmission end, and the method includes:

[0006] Acquire first data information, and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model includes: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set;

[0007] Acquire second data information outputted by the secret language model after performing first-level encryption and decryption processing on the first data information;

[0008] The second data information is encrypted and decrypted at a second level to obtain third data information, and the third data information is transmitted from the first transmission end to the second transmission end.

[0009] In a second aspect, the present disclosure provides an encrypted transmission device for data information, comprising:

[0010] An input module is adapted to obtain first data information and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model comprises: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set;

[0011] A processing module, adapted to obtain second data information outputted after the secret language model performs a first-level encryption and decryption process on the first data information;

[0012] The transmission module is adapted to perform a second-level encryption and decryption process on the second data information to obtain third data information, and transmit the third data information from the first transmission end to the second transmission end.

[0013] In a third aspect, the present disclosure provides an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores one or more computer programs executable by the at least one processor, and one or more of the computer programs are executed by the at least one processor so that the at least one processor can execute the above-mentioned method for encrypting and transmitting data information.

[0014] In a fourth aspect, the present disclosure provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the above-mentioned method for encrypting and transmitting data information when executed by a processor.

[0015] In a fifth aspect, the present disclosure provides a computer program product, including a computer-readable code, or a non-volatile computer-readable storage medium carrying a computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above-mentioned method for encrypting and transmitting data information.

[0016] In the encryption transmission method of data information provided in the embodiment of the present disclosure, the first data information can be encrypted and decrypted at the first level by means of the pre-trained secret language model to obtain the second data information, and the second data information can be further encrypted and decrypted at the second level to obtain the third data information. Through at least two levels of encryption and decryption processing, the security of the data transmission process can be significantly improved. Even if an illegal user maliciously cracks the key information in the second level, it is difficult to determine the model parameters of the secret language model one by one, and thus it is impossible to crack the encryption and decryption method of the first level. In addition, the secret language model further includes a base model pre-trained according to the original secret data set, and a plug-in model updated according to the updated secret data set. By combining the base model and the plug-in model, a technical effect of easy updating can be achieved, and the plug-in model can be flexibly updated in scenarios where the secret reporting method is frequently changed, thereby realizing flexible changes in the encryption and decryption method, so as to flexibly adapt to the secret reporting scenario where the secret reporting method is frequently changed.

[0017] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the embodiments of the present disclosure, they are used to explain the present disclosure and do not constitute a limitation of the present disclosure. The above and other features and advantages will become more apparent to those skilled in the art by describing detailed example embodiments with reference to the accompanying drawings, in which:

[0019] Figure 1 A flowchart of a method for encrypting and transmitting data information provided by an embodiment of the present disclosure;

[0020] Figure 2 A flow chart showing the information transfer party transmitting data information;

[0021] Figure 3 A flow chart showing a receiving party receiving data information is shown;

[0022] Figure 4 A block diagram of a data information encryption transmission device provided by an embodiment of the present disclosure;

[0023] Figure 5 A block diagram of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, the exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, those of ordinary skill in the art should recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0025] In the absence of conflict, the various embodiments of the present disclosure and the various features therein may be combined with each other.

[0026] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0027] The terms used herein are only used to describe specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are also intended to include plural forms, unless the context clearly indicates otherwise. It will also be understood that when the terms "including" and / or "made of" are used in this specification, the presence of the features, wholes, steps, operations, elements and / or components is specified, but the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof is not excluded. "Connected" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect.

[0028] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art. It will also be understood that terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant art and the present disclosure, and will not be interpreted as having an idealized or overly formal meaning unless explicitly defined as such herein.

[0029] In the technical solution disclosed in this disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals. The use of user data in this technical solution complies with relevant national laws and regulations (for example, the "Information Security Technology Personal Information Security Specification", etc.). For example: corresponding prescribed measures are taken for personal information access control; prescribed restrictions are imposed on the display of personal information; the purpose of using personal information does not exceed the scope of direct or reasonable association; when using personal information, clear identity orientation is eliminated to avoid precise positioning of specific individuals.

[0030] In the related art, information encryption transmission can use encryption algorithms and keys to convert plaintext data into ciphertext data to ensure the confidentiality and integrity of the data during transmission. Even if the data is intercepted, the original data cannot be decrypted and read without the key. It can be seen that information encryption transmission ensures the security of data during transmission through encryption algorithms and keys, which is an important measure to protect the confidentiality and integrity of data. However, due to the special situation that the key may be cracked during the communication process, how to further improve the security of information encryption transmission has become a technical problem that needs to be solved urgently. In order to solve the above problems, the present application provides an encryption transmission method for data information, which can significantly improve the security of data transmission through at least two levels of encryption and decryption processing, and can achieve the technical effect of easy updating by combining the base model with the plug-in model, and can flexibly update the plug-in model in the scenario where the secret reporting method is frequently changed, thereby realizing flexible transformation of the encryption and decryption method.

[0031] The encryption transmission method of data information according to the embodiment of the present disclosure can be performed by an electronic device such as a terminal device or a server, and the terminal device can be a user equipment (User Equipment, UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (Personal Digital Assistant, PDA), a handheld device, a computing device, a vehicle-mounted device, a wearable device, etc. The server can be an independent physical server, a server cluster composed of multiple physical servers, or a cloud server capable of cloud computing. The method can be implemented by a processor calling a computer-readable program instruction stored in a memory.

[0032] Figure 1 A flowchart of a method for encrypting and transmitting data information provided by an embodiment of the present disclosure. The method is applied to a first transmission end, wherein the first transmission end can be either a data transmitting end or a data receiving end, which is not limited in the present application. Figure 1 , the method comprises the following steps:

[0033] Step S110: Acquire first data information, and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model includes: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set.

[0034] The first data information refers to: original data information that needs to be encrypted or decrypted. For example, when the first transmission end is a data sending end, the first data information may be original plaintext data to be encrypted and transmitted generated by the first transmission end; when the first transmission end is a data receiving end, the first data information may be original encrypted ciphertext data received by the first transmission end from the second transmission end.

[0035] Accordingly, the pre-trained secret language model is used to perform encryption or decryption processing on the input data. For example, when the first data information is plain text data, the secret language model is used to perform encryption processing; when the first data information is ciphertext data, the secret language model is used to perform decryption processing.

[0036] The secret reporting language model is composed of a base model and a plug-in model. The base model is pre-trained based on the original secret reporting data set and is used to implement the basic functions of encryption and decryption processing. The plug-in model is updated based on the updated secret reporting data set and is used to assist and supplement the base model. The updated secret reporting data set is different from the original secret reporting data set, and the updated secret reporting data set can be determined based on the dynamic update results of the secret reporting method.

[0037] Step S120: obtaining second data information outputted after the secret language model performs first-level encryption and decryption processing on the first data information.

[0038] The first level of encryption and decryption processing is the encryption and decryption processing implemented by the secret language model. Therefore, the second data information refers to: information obtained by performing encryption or decryption processing on the first data information by the secret language model. For example, when the first data information is plain text data, the second data information is obtained by encrypting the first data information; when the first data information is ciphertext data, the second data information is obtained by decrypting the first data information.

[0039] Step S130: performing a second-level encryption and decryption process on the second data information to obtain third data information, and transmitting the third data information from the first transmission end to the second transmission end.

[0040] The second-level encryption and decryption processing can be implemented by various methods such as key method, and this application does not limit this. When the second data information is obtained by encrypting the first data information, the third data information is obtained by further encrypting the second data information; when the second data information is obtained by decrypting the first data information, the third data information is obtained by further decrypting the second data information.

[0041] Among them, the second transmission end corresponds to the first transmission end. When the first transmission end is a data sending end, the second transmission end is a data receiving end; when the first transmission end is a data receiving end, the second transmission end is a data sending end.

[0042] It can be seen that in the encryption transmission method of data information provided by the embodiment of the present disclosure, the first data information can be encrypted and decrypted at the first level by means of the pre-trained secret language model to obtain the second data information, and the second data information can be further encrypted and decrypted at the second level to obtain the third data information. Through at least two levels of encryption and decryption processing, the security of the data transmission process can be significantly improved. Even if an illegal user maliciously cracks the key information in the second level, it is difficult to determine the model parameters of the secret language model one by one, and thus it is impossible to crack the encryption and decryption method of the first level. In addition, the secret language model further includes a base model pre-trained according to the original secret data set, and a plug-in model updated according to the updated secret data set. By combining the base model and the plug-in model, a technical effect of easy updating can be achieved, and the plug-in model can be flexibly updated in the scenario where the secret reporting method is frequently changed, thereby realizing flexible transformation of the encryption and decryption method, so as to flexibly adapt to the secret reporting scenario where the secret reporting method is frequently changed.

[0043] In addition, those skilled in the art may also make various modifications and variations to the method in this application:

[0044] In some possible implementations, the secret reporting language model can be updated in the following manner: obtaining the i-th updated secret reporting data set corresponding to the i-th update process, generating the i-th update instruction set corresponding to the i-th updated secret reporting data set; training the updated plug-in model according to the i-th updated secret reporting data set and the i-th update instruction set; wherein the updated plug-in model is used to expand the pre-trained base model; wherein i is a natural number. It can be seen that in the process of updating the secret reporting language model, the base model can be kept unchanged, and the plug-in model can be dynamically updated, so as to supplement the encryption and decryption functions of the base model through the dynamically updated plug-in model. Specifically, in the process of updating the plug-in model each time, the corresponding update instruction set can be generated according to the updated secret reporting data set, so as to jointly train the updated plug-in model according to the updated secret reporting data set and the update instruction set. wherein the updated secret reporting data set is used to store the updated secret reporting method (such as the mapping rules between plaintext and ciphertext, etc.).

[0045] In addition, in the process of updating the plug-in model, it can be updated flexibly in a variety of ways: for example, the plug-in model after each update can cover the plug-in model before the update, so as to achieve iterative update of the plug-in model without increasing the number of plug-in models. For another example, a new plug-in model can be generated after each update, while the plug-in model after the last update is still retained, so that when the number of updates gradually increases, the number of plug-in models also gradually increases. In actual situations, the appropriate plug-in model update method can be flexibly selected according to the correlation between the current update process and the previous update process.

[0046] In some possible implementations, when it is determined that there is a data association relationship between the i-th update secret report data set corresponding to the i-th update process (i.e., this update process) and the i-1th update secret report data set corresponding to the i-1th update process (i.e., the previous update process), the updated plug-in model can be trained in the following manner: according to the i-th update secret report data set and the i-th update instruction set, the i-1th updated plug-in model is updated to obtain the i-th updated plug-in model; wherein, the i-th updated plug-in model is used to replace the i-1th updated plug-in model.

[0047] Among them, the update secret report data set corresponding to the current update process and the update secret report data set corresponding to the previous update process have a data association relationship, which means that the update secret report data set corresponding to the current update process is implemented on the basis of the update secret report data set corresponding to the previous update process. For example, the update secret report data set corresponding to the current update process iterates the update secret report data set corresponding to the previous update process, that is, the update secret report data set corresponding to the current update process can directly replace the update secret report data set corresponding to the previous update process. In this case, in order to avoid generating redundant plug-in models, the plug-in model after the i-1th update can be directly updated so that the plug-in model after the i-th update replaces the plug-in model after the i-1th update.

[0048] In some possible implementations, when it is determined that there is no data association relationship between the i-th update secret report data set corresponding to the i-th update process and the i-1-th update secret report data set corresponding to the i-1-th update process, the updated plug-in model can be trained in the following manner: based on the i-th update secret report data set and the i-th update instruction set, the plug-in model corresponding to the i-th update process is trained; wherein, the plug-in model corresponding to the i-th update process is different from the plug-in model corresponding to the i-1-th update process.

[0049] Among them, the update secret report data set corresponding to this update process does not have a data association relationship with the update secret report data set corresponding to the previous update process, which means that the update secret report data set corresponding to this update process is not implemented on the basis of the update secret report data set corresponding to the previous update process, but is independent of the update secret report data set corresponding to the previous update process. For example, the update secret report data set corresponding to this update process corresponds to the first part of the content in the secret report, and the update secret report data set corresponding to the previous update process corresponds to the second part of the content in the secret report. Among them, the first part of the content and the second part of the content are used together to form the secret report information. For example, the first part of the content and the second part of the content can respectively correspond to different areas of the same set of secret report schemes, or the first part of the content and the second part of the content can also respectively correspond to different implementation methods of the same set of secret report schemes. Therefore, the update secret report data set corresponding to this update process cannot directly replace the update secret report data set corresponding to the previous update process. In this case, the plug-in model corresponding to the i-th update process is trained separately, so that the plug-in model corresponding to the i-th update process is independent of the plug-in model corresponding to the i-1-th update process, so that the parameter matrices in the two models are used together as the supplementary content of the base model.

[0050] In some possible implementations, before training the updated plug-in model based on the ith update secret report data set and the ith update instruction set, the following operations are further performed: determining the data features of the ith update secret report data set; determining the parameter position information corresponding to the data features of the ith update secret report data set based on a pre-generated feature parameter mapping table; wherein the feature parameter mapping table is used to store the mapping relationship between multiple data features and multiple parameter position information. Accordingly, the model matrix included in the updated plug-in model corresponds to the local model parameters corresponding to the parameter position information in the base model.

[0051] Among them, the data features of the i-th update secret report data set can be data type features, data length features, data language features, etc. For example, data type features can be divided according to various situations such as data paragraph features, applicable scenario types, etc.; data length features are used to characterize the length of data in the update secret report data set to reflect the complexity of the data; data language features are used to characterize the language used by the data in the update secret report data set, such as numbers, characters, Chinese, English, etc. In short, those skilled in the art can flexibly set the specific types and connotations of data features as long as they can distinguish different data sets.

[0052] In addition, the base model contains a large number of model parameters, which can be characterized by parameter position information. Among them, the parameter position information is used to characterize the position and range of the parameters in the base model. For example, the parameter position information can be characterized according to the model level of the base model. Assuming that the base model contains dozens of model levels, the parameter position information is used to characterize which model levels the corresponding parameters are in the base model. For example, the parameter position information can be characterized according to the matrix area in the model matrix of the base model. For the model matrix of any model level, it has multiple matrix areas, and the corresponding parameters can be characterized by parameter position information. Which matrix areas in the model matrix of the base model are in. The number of matrix areas can be flexibly divided according to the matrix size. When the matrix size is small, it can be divided into four matrix areas, and when the matrix size is large, it can be divided into more matrix areas. In short, the corresponding parameter range in the base model can be quickly located by the parameter position information. In the process of implementing the present invention, the inventor found that there is a specific mapping relationship between a variety of data features and a plurality of parameter position information, and the mapping relationship can be determined in advance by testing or training the sample data.

[0053] Correspondingly, the feature parameter mapping table is used to store the mapping relationship between multiple data features and multiple parameter position information. Optionally, the number of parameter position information is determined according to the number of types of data features. By querying the feature parameter mapping table, the parameter position information corresponding to the data feature of the i-th update secret report data set can be determined, so that the model matrix contained in the updated plug-in model corresponds to the local model parameter corresponding to the parameter position information in the base model. For example, in the feature parameter mapping table, for any data feature, one or more parameter position information corresponding to the data feature can be stored. It can be seen that the mapping relationship between data features and parameter position information can be a one-to-one or one-to-many mapping relationship. With the help of the feature parameter mapping table, the model parameter area with the greatest correlation with the data feature of the i-th update secret report data set in the base model can be determined (which can be represented by the model hierarchy or the matrix area of ​​the model parameter matrix), so that the updated plug-in model is dedicated to the expansion processing of the model parameter area with the greatest correlation, thereby improving the accuracy and efficiency of the plug-in model.

[0054] In some possible implementations, there are multiple second transmission terminals, and the method can also transmit the pre-trained secret language model to the blockchain network so that multiple second transmission terminals can obtain the secret language model from the blockchain network. With the help of the blockchain network, the secret language model can be effectively prevented from being tampered with during the transmission process, thereby improving the security and accuracy of the model. Accordingly, the second transmission terminal is used to perform the second-level encryption and decryption processing on the received third data information to obtain the second data information, and perform the first-level encryption and decryption processing on the second data information through the secret language model to obtain the first data information.

[0055] For ease of understanding, the technical implementation details of the above embodiment are described in detail below using a specific example. The example mainly involves the fields of information security and artificial intelligence, and is particularly used to achieve high-security information encryption, transmission and decryption in a complex communication environment.

[0056] With the development of artificial intelligence technology, large language models have shown great potential in natural language processing, reasoning, information generation and other fields. However, how to use these large language models for secure information transmission, especially in high-confidentiality scenarios, still faces many challenges. Current encrypted information transmission technology mainly relies on traditional cryptographic methods, such as symmetric encryption or asymmetric encryption. Although these methods are effective in protecting information content, there is still room for improvement in terms of information integrity, tampering protection, and consistency verification during distribution.

[0057] Accordingly, this example aims to solve the following technical problems:

[0058] (1) Security of information transmission: Traditional information transmission methods may not be able to effectively prevent information from being tampered with, replaced, or stolen in complex scenarios. To solve this problem, this example combines a large language model with cryptography technology to ensure the security of information transmission.

[0059] (2) Transparency and immutability of model distribution: In the prior art, the distribution process of large language models lacks transparency and verification mechanisms, which can easily lead to tampering or replacement of the model during the distribution process, affecting the accurate transmission of confidential information. To solve this problem, this example uses blockchain technology to achieve transparent management and immutability of model distribution, thereby improving the security of confidential information transmission.

[0060] (3) Complexity of model version management: With the continuous update and iteration of large language models, how to effectively manage the versions of base models and plug-in models and ensure the correctness of model versions when transmitting information is one of the problems to be solved by the present invention. This example achieves the effect of facilitating update and iteration by updating the plug-in model in a targeted manner.

[0061] In this example, the secret report language model can also be referred to as a secret report large language model. The secret report large language model can be trained and distributed by an information management node. The information management node can be implemented by any of the first transmission terminal or the second transmission terminal mentioned above, or it can be a management node independent of the first transmission terminal or the second transmission terminal mentioned above. For example, the information management node can be implemented by the management server of the intelligence center department, and the intelligence center department pre-trains and regularly updates the secret report large language model, and distributes the trained and updated secret report large language model to each transmission terminal.

[0062] Specifically, the intelligence center is responsible for training the secret information language model and distributing it to various information transmission points (i.e. the transmission end mentioned above). The secret information language model can be trained in the following ways:

[0063] First, collect confidential data in a specific scenario. Then, based on the collected confidential data in the scenario, build a pre-training data set and instruction set. Next, pre-train the open source large language model (LLM) based on the pre-training data set to enhance its domain knowledge. Then, perform full-instruction supervised fine-tuning (SFT) on the pre-trained LLM based on the instruction set to enhance its instruction following capability so that it has the ability to output corresponding reasoning results for specific prompts. The trained model is the confidential large model, which can be regarded as the base model mentioned above. Among them, SFT (Supervised Fine-Tuning) is a deep learning strategy used in machine learning and natural language processing. It is used to further train the pre-trained model by providing labeled data to make the model perform better in specific tasks or fields.

[0064] The base model can be distributed and put into use. For example, the base model can be distributed to various information delivery points to ensure model consistency. However, in order to facilitate model updates, the following update mechanism can be further introduced: on the basis of the base model, according to the instruction set constructed by the new confidential data (i.e., the updated confidential data set), use LoRA technology to train an external model matrix (i.e., the external model), and update the model version in the form of base model + external model. Therefore, the model version can be further subdivided into a base model version and a corresponding external model version. By dividing the model version into two parts, the base model version and the corresponding external model version, the base model version and the external model version can be updated independently, thereby improving the flexibility of the update process. Among them, the full name of LoRA is Low-Rank Adaptation, which is a deep learning model optimization technology used to efficiently adjust the pre-trained model.

[0065] In addition, in order to further improve security, when the intelligence center distributes the secret report large language model to each information transmission point, the blockchain can be introduced for distribution management. Specifically, the model version information, timestamp, checksum and other information of each distribution can be recorded on the blockchain to ensure the transparency and immutability of the model distribution process. Accordingly, after each information transmission point receives the secret report large language model, the integrity and legitimacy of the secret report large language model can be verified through the blockchain to prevent the secret report large language model from being tampered with or replaced during the distribution process.

[0066] Figure 2 FIG. 1 shows a flow chart of the information transmission party transmitting data information. Figure 2 As shown, the data transmission method of the information transmitting party includes the following steps:

[0067] Step S201: Obtain the information to be transmitted.

[0068] Specifically, it is necessary to clarify the information to be transmitted, which may be information to be encrypted or information to be decrypted.

[0069] Step S202: Formulate prompt content according to the information to be transmitted.

[0070] Specifically, a prompt is formulated according to the information to be transmitted. The prompt is used to prompt and guide the secret reporting language model.

[0071] Step S203: verify the inference result of the secret large language model, and record the model-related data if the verification passes.

[0072] Among them, when verifying the inference results of the confidential large language model, it is necessary to combine different versions of the confidential large language model and the large model hyperparameters (such as temperature value, random sampling probability, repeated penalty coefficient, etc.), and continuously repeat the experiment to ensure that the inference result of the model is the information that needs to be transmitted. It can be seen that the model-related data recorded when the verification is passed refers to: when the inference result of the confidential large language model is correct, the version information and hyperparameter information of the model. Among them, the hyperparameter information can include: the specific value of the temperature value, the specific probability value of the random sampling probability, the specific coefficient value of the repeated penalty coefficient, etc.

[0073] Step S204: encrypt the inference results of the confidential large language model and model-related data.

[0074] The reasoning result (i.e., the encrypted result) of the secret large language model can be understood as the first-level encryption processing result. Therefore, the content obtained after encrypting the reasoning result of the secret large language model and the model-related data in step S204 can be understood as the second-level encryption processing result.

[0075] Specifically, the model version (base model version, plug-in model version), model hyperparameters, the formulated prompt and other related data can be encrypted according to traditional cryptographic encryption algorithms (such as symmetric or asymmetric encryption, etc.) to convert them into ciphertext.

[0076] Step S205: Transmit the ciphertext in the communication system.

[0077] Figure 3 FIG. 2 shows a flow chart of the information receiving party receiving data information. Figure 3 As shown, the data receiving method of the information receiving party includes the following steps:

[0078] Step S301: receiving ciphertext.

[0079] Step S302: Decrypt the encrypted data using a traditional encryption algorithm based on the encrypted data features to obtain the inference results of the encrypted large language model and model-related data mentioned above, including, for example, model version, hyperparameters, prompt, etc.

[0080] Step S303: Based on the inference result (ie, encryption result) of the secret large language model and model-related data, the secret large language model is used to obtain the plain text.

[0081] Specifically, based on the model version and hyperparameter information, the centrally distributed secret large language model is used for reasoning (the reasoning here is Figure 2 The reverse process of the reasoning in the above example can be the decryption process) to obtain the transmitted information. In this step, it is necessary to ensure that the template version and hyperparameter information of the secret large language model are consistent with Figure 2 The relevant information used in is exactly the same.

[0082] It can be seen that this example can achieve multi-level encryption processing through cryptography combined with the secret large language model to improve the security of information transmission. In addition, with the help of blockchain technology, the distribution reliability of multi-terminal models can be improved. This example provides a solution for implementing the model training and distribution mechanism of the intelligence center in the encrypted information transmission scenario. This example can apply the large language model to encrypted information transmission, thereby further improving information security through the large language model on the basis of traditional encryption technology. In addition, during the information transmission and reception process, the information is first input into the LLM model, and the model version, hyperparameters, prompt and other contents are used to construct the encrypted ciphertext, and the encrypted ciphertext is transmitted to the receiving end. Correspondingly, the decrypted ciphertext can realize the processing flow of reasoning information through the LLM model.

[0083] It can be understood that the above-mentioned various method embodiments mentioned in the present disclosure can be combined with each other to form a combined embodiment without violating the principle logic. Due to space limitations, the present disclosure will not repeat them. It can be understood by those skilled in the art that in the above-mentioned method of the specific implementation method, the specific execution order of each step should be determined according to its function and possible internal logic.

[0084] In addition, the present disclosure also provides an encrypted transmission device, an electronic device, and a computer-readable storage medium for data information, all of which can be used to implement any encrypted transmission method for data information provided by the present disclosure. The corresponding technical solutions and descriptions are referred to in the corresponding records of the method part and will not be repeated here.

[0085] Figure 4 A block diagram of a data information encryption transmission device provided in an embodiment of the present disclosure.

[0086] Reference Figure 4The embodiment of the present disclosure provides a data information encryption transmission device 40, the data information encryption transmission device comprising:

[0087] The input module 41 is adapted to obtain first data information and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model comprises: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set;

[0088] A processing module 42, adapted to obtain second data information outputted after the secret language model performs a first-level encryption and decryption process on the first data information;

[0089] The transmission module 43 is adapted to perform a second-level encryption and decryption process on the second data information to obtain third data information, and transmit the third data information from the first transmission end to the second transmission end.

[0090] In an optional implementation, the secret reporting language model is updated in the following manner:

[0091] Acquire an i-th update secret report data set corresponding to an i-th update process, and generate an i-th update instruction set corresponding to the i-th update secret report data set;

[0092] According to the i-th update secret report data set and the i-th update instruction set, an updated plug-in model is trained; wherein the updated plug-in model is used to expand the base model obtained by pre-training; wherein i is a natural number.

[0093] In an optional implementation, when it is determined that there is a data association relationship between the i-th update secret report data set corresponding to the i-th update process and the i-1-th update secret report data set corresponding to the i-1-th update process, the training of the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set includes:

[0094] According to the i-th update secret report data set and the i-th update instruction set, the plug-in model after the i-1th update is updated to obtain the i-th updated plug-in model; wherein the i-th updated plug-in model is used to replace the i-1th updated plug-in model.

[0095] In an optional implementation, when it is determined that there is no data association relationship between the i-th update secret report data set corresponding to the i-th update process and the i-1-th update secret report data set corresponding to the i-1-th update process, the training of the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set includes:

[0096] The plug-in model corresponding to the i-th update process is trained based on the i-th update secret report data set and the i-th update instruction set; wherein the plug-in model corresponding to the i-th update process is different from the plug-in model corresponding to the i-1-th update process.

[0097] In an optional implementation, before training the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set, the method further includes:

[0098] Determining data characteristics of the i-th update secret report data set;

[0099] Determine the parameter position information corresponding to the data feature of the i-th update secret report data set according to a pre-generated feature parameter mapping table; wherein the feature parameter mapping table is used to store the mapping relationship between multiple data features and multiple parameter position information;

[0100] The model matrix included in the updated plug-in model corresponds to the local model parameters corresponding to the parameter position information in the base model.

[0101] In an optional implementation, the number of the second transmission terminals is multiple, and the transmission module is further used to: transmit the pre-trained secret language model to the blockchain network, so that the multiple second transmission terminals can obtain the secret language model from the blockchain network;

[0102] Furthermore, the second transmission end is used to perform second-level encryption and decryption processing on the received third data information to obtain the second data information, and to perform first-level encryption and decryption processing on the second data information through the secret language model to obtain the first data information.

[0103] Each module in the above-mentioned data information encryption transmission device can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the corresponding operations of each of the above modules.

[0104] Figure 5 A block diagram of an electronic device provided in an embodiment of the present disclosure.

[0105] Reference Figure 5An embodiment of the present disclosure provides an electronic device, which includes: at least one processor 701; at least one memory 702, and one or more I / O interfaces 703, connected between the processor 701 and the memory 702; wherein the memory 702 stores one or more computer programs that can be executed by the at least one processor 701, and the one or more computer programs are executed by the at least one processor 701 so that the at least one processor 701 can execute the above-mentioned encryption transmission method of data information.

[0106] Each module in the above electronic device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0107] The present disclosure also provides a computer-readable storage medium on which a computer program is stored, wherein the computer program implements the above-mentioned method for encrypting and transmitting data information when executed by a processor / processing core. The computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium.

[0108] The embodiment of the present disclosure also provides a computer program product, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the above-mentioned method for encrypting and transmitting data information.

[0109] It will be appreciated by those skilled in the art that all or some of the steps, systems, and functional modules / units in the methods disclosed above may be implemented as software, firmware, hardware, and appropriate combinations thereof. In a hardware implementation, the division between the functional modules / units mentioned in the above description does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed by several physical components in cooperation. Some or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application-specific integrated circuit. Such software may be distributed on a computer-readable storage medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium).

[0110] As is known to those of ordinary skill in the art, the term computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable program instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technology, portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. In addition, it is known to those of ordinary skill in the art that communication media typically contain computer-readable program instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information delivery medium.

[0111] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0112] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0113] The computer program product described herein may be implemented in hardware, software, or a combination thereof. In one optional embodiment, the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK), etc.

[0114] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0115] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0116] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0117] The flow chart and block diagram in the accompanying drawings show the possible architecture, function and operation of the system, method and computer program product according to multiple embodiments of the present disclosure. In this regard, each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of the module, program segment or instruction includes one or more executable instructions for realizing the specified logical function. In some alternative implementations, the function marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of special hardware and computer instructions.

[0118] Example embodiments have been disclosed herein, and although specific terms are employed, they are used and should be interpreted only in a general illustrative sense and not for limiting purposes. In some instances, it will be apparent to those skilled in the art that, unless otherwise expressly noted, features, characteristics, and / or elements described in conjunction with a particular embodiment may be used alone or in combination with features, characteristics, and / or elements described in conjunction with other embodiments. Therefore, those skilled in the art will appreciate that various changes in form and detail may be made without departing from the scope of the present disclosure as set forth in the appended claims.

Claims

1. A method for encrypting and transmitting data information, characterized in that: Applied to a first transmission end, the method includes: Acquire first data information, and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model includes: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set; Acquire second data information outputted by the secret language model after performing first-level encryption and decryption processing on the first data information; The second data information is encrypted and decrypted at a second level to obtain third data information, and the third data information is transmitted from the first transmission end to the second transmission end.

2. The method according to claim 1, characterized in that The secret language model is updated in the following way: Acquire an i-th update secret report data set corresponding to an i-th update process, and generate an i-th update instruction set corresponding to the i-th update secret report data set; According to the i-th update secret report data set and the i-th update instruction set, an updated plug-in model is trained; wherein the updated plug-in model is used to expand the base model obtained by pre-training; wherein i is a natural number.

3. The method according to claim 2, characterized in that In the case where it is determined that there is a data association relationship between the i-th update secret report data set corresponding to the i-th update process and the i-1-th update secret report data set corresponding to the i-1-th update process, the training of the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set includes: According to the i-th update secret report data set and the i-th update instruction set, the plug-in model after the i-1th update is updated to obtain the i-th updated plug-in model; wherein the i-th updated plug-in model is used to replace the i-1th updated plug-in model.

4. The method according to claim 2, characterized in that: In the case where it is determined that there is no data association relationship between the i-th update secret report data set corresponding to the i-th update process and the i-1-th update secret report data set corresponding to the i-1-th update process, the training of the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set includes: According to the i-th update secret report data set and the i-th update instruction set, a plug-in model corresponding to the i-th update process is trained; wherein the plug-in model corresponding to the i-th update process is different from the plug-in model corresponding to the i-1-th update process.

5. The method according to any one of claims 2 to 4, characterized in that: Before training the updated plug-in model according to the i-th update secret report data set and the i-th update instruction set, the method further includes: Determining data characteristics of the i-th update secret report data set; Determine the parameter position information corresponding to the data feature of the i-th update secret report data set according to a pre-generated feature parameter mapping table; wherein the feature parameter mapping table is used to store the mapping relationship between multiple data features and multiple parameter position information; The model matrix included in the updated plug-in model corresponds to the local model parameters corresponding to the parameter position information in the base model.

6. The method according to any one of claims 1 to 4, characterized in that: The number of the second transmission terminals is multiple, and the method further includes: transmitting the pre-trained secret language model to the blockchain network, so that the multiple second transmission terminals can obtain the secret language model from the blockchain network; Furthermore, the second transmission end is used to perform second-level encryption and decryption processing on the received third data information to obtain the second data information, and to perform first-level encryption and decryption processing on the second data information through the secret language model to obtain the first data information.

7. A data information encryption transmission device, characterized in that: include: An input module, adapted to obtain first data information, and input the first data information into a pre-trained secret reporting language model; wherein the secret reporting language model comprises: a base model pre-trained according to an original secret reporting data set, and an external model updated according to an updated secret reporting data set; A processing module, adapted to obtain second data information outputted after the secret language model performs a first-level encryption and decryption process on the first data information; The transmission module is adapted to perform a second-level encryption and decryption process on the second data information to obtain third data information, and transmit the third data information from the first transmission end to the second transmission end.

8. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores one or more computer programs that can be executed by the at least one processor, and the one or more computer programs are executed by the at least one processor so that the at least one processor can execute the encryption transmission method for data information as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the computer program implements the method for encrypting and transmitting data information according to any one of claims 1 to 6.

10. A computer program product, characterized in that It includes a computer-readable code, or a non-volatile computer-readable storage medium carrying a computer-readable code. When the computer-readable code runs in a processor of an electronic device, the processor in the electronic device executes the encryption transmission method of data information as described in any one of claims 1 to 6.

Images