Information processing method and server

Abstract

The invention provides an information processing method and a server. The method comprises the following steps: receiving an outgoing request sent by electronic equipment; in response to the outgoing request, identifying to-be-encrypted information in the to-be-processed information; encrypting the to-be-encrypted information to obtain encrypted to-be-processed information; sending the encrypted to-be-processed information to the second server, so as to process the encrypted to-be-processed information through the target model in the second server; and receiving a processing result sent by the second server, and decrypting the processing result to obtain the target information. Through the method provided by the invention, the to-be-encrypted information can be prevented from being leaked.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and in particular to an information processing method and server. Background Art

[0002] Large Model refers to a machine learning model with a large number of parameters and complex structure, which can process massive data and complete various complex tasks. For example, natural language processing, computer vision, speech recognition, etc. Among them, the Large Language Models (LLM) in natural language processing is a deep neural network model that is trained based on massive text data and learns and masters general language knowledge and abilities. It can not only generate natural language text, but also deeply understand the meaning of the text and handle various natural language tasks.

[0003] With the application and popularization of large language models, tasks such as text summarization, translation, and question-answering in daily work and life can all be provided by external large language models. While enjoying the convenience brought by technological progress, it also raises the issue of leakage of encrypted information, which will bring great risks to user privacy and security. Summary of the invention

[0004] The embodiment of the present application discloses an information processing method and a server, which solves the technical problem of leakage of information to be encrypted.

[0005] In a first aspect, an embodiment of the present application provides an information processing method, which is applied to a first server, wherein the first server is connected to a second server via an external network, and the method comprises: receiving an outgoing request sent by an electronic device, wherein the outgoing request is used to request that the information to be processed be sent to the target model for processing; responding to the outgoing request, identifying the information to be encrypted in the information to be processed; based on the type of the information to be encrypted, obtaining encrypted information corresponding to the type of the information to be encrypted; replacing the information to be encrypted in the information to be processed with the corresponding encrypted information to obtain encrypted information to be processed; sending the encrypted information to be processed to the second server so that the encrypted information to be processed is processed by the target model in the second server; and receiving the processing result sent by the second server, and replacing the encrypted information in the processing result based on the information to be encrypted to obtain the target information.

[0006] By adopting the information processing method provided in the embodiment of the present application, the information to be encrypted in the information to be processed can be intercepted, or the information to be encrypted in the information to be processed can be identified and then encrypted and concealed, thereby retaining the semantic environment of the context. This can prevent the leakage of the information to be encrypted and protect information security, and can also use the convenient services of the target model to obtain accurate target information.

[0007] In a possible implementation, the identifying of the information to be encrypted in the information to be processed includes: parsing the information to be processed and obtaining text information; processing the text information by pattern matching to obtain target text content; and, based on the target text content, determining that the information to be processed does not fall into the blacklist, calling the first model to identify the information to be encrypted in the information to be processed. Through the above technical solution, the information in the information to be processed can be preliminarily screened by pattern matching, and when the information in the information to be processed falls into the blacklist, the information to be processed can be intercepted to prevent the leakage of the encrypted information.

[0008] In a possible implementation, the identifying the information to be encrypted in the information to be processed includes: using the second model to identify the named entities in the text information; inputting the named entities into the first model to obtain the information to be encrypted. Through the above technical solution, the information to be encrypted in the text information can be accurately identified by the first model, thereby improving the accuracy of identifying the information to be encrypted.

[0009] In one possible implementation, the method of encrypting the information to be encrypted based on the type of the information to be processed to obtain the encrypted information to be processed includes: using a third model to identify the type of each named entity in the information to be encrypted and the corresponding accuracy rate; assigning a corresponding weight based on the frequency of occurrence of each named entity in the information to be processed; obtaining a corresponding score value through a weighted sum algorithm based on the accuracy rate and the weight; and encrypting the information to be encrypted based on the score value to obtain the encrypted information to be processed.

[0010] In a possible implementation, encrypting the information to be encrypted based on the score value to obtain the encrypted information to be processed includes: when the score value is greater than or equal to a preset value, determining the corresponding target entity from a pre-established password library based on the named entity; replacing the named entity based on the target entity to obtain the encrypted information to be processed. Through the above technical solution, the information to be encrypted in the information to be processed can be concealed without affecting the context content in the information to be processed, and without affecting the accuracy of the target model in processing the information to be processed.

[0011] In a possible implementation, the receiving of the processing result sent by the second server and replacing the encrypted information in the processing result based on the information to be encrypted to obtain the target information includes: receiving the processing result and identifying the processing result using the first model; and when it is determined by the first model that the processing result does not contain the information to be encrypted, decrypting the information to be encrypted in the processing result to obtain the target information. Through the above technical solution, after obtaining the processing result according to the target model, the information to be encrypted in the processing result can be identified, and when it is determined that the processing result does not contain the information to be encrypted, the processing result can be decrypted, thereby further avoiding the leakage of the information to be encrypted.

[0012] In a possible implementation, the method further includes: generating an information pair according to the information to be encrypted and the encrypted information to be encrypted, wherein the information pair includes a named entity and a target entity.

[0013] In a possible implementation, replacing the encrypted information in the processing result based on the information to be encrypted to obtain the target information includes: determining an information pair corresponding to the encrypted information to be encrypted in the processing result; obtaining the corresponding named entity based on the information pair; and replacing the target entity based on the named entity to obtain the target information. Through the above technical solution, the information to be encrypted can be decrypted to obtain accurate target information.

[0014] In a second aspect, the present application provides an information processing device, comprising: a receiving module for receiving an outgoing request sent by the electronic device, the outgoing request being used to request that the information to be processed be sent to a target model for processing; an identification module being used to respond to the outgoing request and identify the information to be encrypted in the information to be processed; an encryption module being used to obtain encryption information corresponding to the type of the information to be encrypted based on the type of the information to be encrypted; replacing the information to be encrypted in the information to be processed with the corresponding encryption information to obtain encrypted information to be processed; a processing module being used to send the encrypted information to be processed to the second server, processing the encrypted information to be processed by the target model in the second server, obtaining a processing result and sending the processing result to the first server; and the processing module being also used to receive the processing result and decrypt the information to be encrypted in the processing result to obtain the target information.

[0015] In a third aspect, the present application provides a server, comprising a processor and a memory, wherein the processor implements the above-mentioned information processing method when executing a computer program stored in the memory.

[0016] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor, the information processing method as described above is implemented.

[0017] In the information processing method provided by the present application, the information to be encrypted in the information to be processed is identified and encrypted to obtain the encrypted information to be processed, the encrypted information to be processed is processed using the target model to obtain the processing result, and the processing result is decrypted to obtain the target information. The information to be encrypted in the information to be processed can be intercepted, or the information to be encrypted in the information to be processed can be identified and encrypted and concealed to retain the semantic environment of the context. It can prevent the leakage of the information to be encrypted and protect information security, and can also use the convenient service of the target model to obtain accurate target information. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of an application scenario of the information processing method provided in an embodiment of the present application.

[0019] Figure 2 It is a flowchart of the information processing method provided in an embodiment of the present application.

[0020] Figure 3 It is a flowchart of a method for preprocessing information to be processed in an embodiment of the present application.

[0021] Figure 4 It is a flowchart of a method for identifying named entities using a second model in an embodiment of the present application.

[0022] Figure 5 It is a schematic diagram of the training process of training the first model provided in an embodiment of the present application.

[0023] Figure 6 It is a schematic diagram of dependency analysis provided in an embodiment of the present application.

[0024] Figure 7 It is an overall architecture diagram of the information processing method provided in the embodiment of the present application.

[0025] Figure 8 It is a structural diagram of the information processing device provided in an embodiment of the present application.

[0026] Fig. 9 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0027] To facilitate understanding, some illustrations of concepts related to the embodiments of the present application are given by way of example for reference.

[0028] It should be noted that in this application, "at least one" refers to one or more, and "more than one" refers to two or more than two. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: the existence of A alone, the existence of A and B at the same time, and the existence of B alone, where A and B can be singular or plural. The terms "first", "second", "third", "fourth", etc. (if any) in the specification, claims and drawings of this application are used to distinguish similar objects, rather than to describe a specific order or sequence. For ease of understanding, the technical terms involved in the embodiments of the present application will be introduced below.

[0029] (1) Machine Learning (ML)

[0030] Machine learning is an artificial intelligence technology that enables computers to learn from data and improve their performance. Machine learning is a branch of artificial intelligence that uses algorithms and statistical models to enable computer systems to use sample data to improve the performance of task execution. Unlike traditional programming, machine learning does not require hard-coded rules, but allows machines to make predictions or decisions about outputs by receiving input data.

[0031] (2) Natural Language Processing (NLP)

[0032] Natural language processing is a technology that enables computers to understand, interpret, and generate human language. Through natural language processing technology, a variety of applications such as machine translation, question-answering systems, sentiment analysis, and text summarization can be achieved.

[0033] (3) Pattern Matching (PM)

[0034] Pattern matching is a technique used to find specific patterns in data structures. It is used to find all substrings in a string that are identical to another substring. For example, find "hjh" in the string "hjh123abc".

[0035] (4) Unstructured Data (UD)

[0036] Unstructured data refers to data with irregular or incomplete data structure, no predefined data model, and data that is not convenient to be logically expressed and implemented using a two-dimensional logical table in a database. The format of unstructured data is diverse. As an example, unstructured data can include documents or texts in various formats.

[0037] (5) Information to be encrypted

[0038] The information to be encrypted includes personal identity information, mobile phone numbers, bank card information and other information involving personal privacy, as well as information involving internal corporate secrets that needs to be restricted.

[0039] In some business scenarios, user information to be processed needs to be sent to a large language model via the network for processing. User information to be processed (such as text, files, pictures, etc.) may contain confidential information, personal privacy information waiting to be encrypted. Uploading unprocessed encrypted information to an external model will pose a great threat to user privacy and security.

[0040] The embodiments of the present application utilize machine learning and other technologies to intelligently identify information to be encrypted in information to be processed (such as text, pictures, files, etc.), and conduct audits and classifications based on the reference data of the information to be encrypted to determine whether it needs to be intercepted or whether the information to be encrypted can be processed and then sent out. For the information to be processed that can be processed and sent out, different encryption strategies are implemented according to the specific scenario and the type of data entity in the information to be encrypted. After encryption, the state attribute information and context environment information of the encrypted content in the user's information to be processed can be highly retained. In this way, after the processed information to be processed is sent out to the target model, for example, an artificial intelligence (AI) model or other service, it not only effectively conceals the information to be encrypted, but also greatly reduces the interference with the AI ​​model reasoning, and better guarantees the reliability and accuracy of the results returned by the AI ​​model.

[0041] In order to better understand the information processing method provided by the embodiment of the present application, the application scenario of the information processing method provided by the embodiment of the present application is first described below.

[0042] Figure 1 Schematic diagram of the application environment of the information processing method provided in the embodiment of the present application. Figure 1 As shown, in some embodiments of the present application, the information processing method can be applied to the first server 1, and the first server 1 is respectively connected to the electronic device 2 (such as a computer) and the second server 3. Among them, the electronic device 2 is connected to the first server 1 through an internal network (such as a local area network, an intranet or an encrypted network, etc.), and the first server 1 communicates with the second server 3 through an external network (such as a wide area network, etc.). The information processing method provided in the embodiment of the present application can prevent the user from leaking the information to be encrypted when using an external target model (that is, during the communication between the first server 1 and the second server 3).

[0043] The electronic device 2 can be a computer, a server, a smart screen, a mobile phone, a tablet computer, a smart wearable device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, a netbook, etc. The embodiment of the present application does not impose any restrictions on the specific type of the electronic device.

[0044] The first server 1 is used to receive an outgoing request sent by the electronic device 2, and the outgoing request is used to request to send the information to be processed to the target model in the second server 3 for processing.

[0045] The second server 3 stores a target model for providing services to users of the electronic device 2. For example, the second server 3 stores a large language model. When a user needs to analyze text data, the text data can be input through the electronic device 2 and sent to the first server 1. After the text data is processed by the first server 1, the processed text data is sent to the second server 3. The second server 3 uses the large language model to analyze the text data and outputs the text result, sends the text result to the first server 1, and the first server 1 decrypts the text result and sends the decrypted result to the electronic device 2. The electronic device 2 displays the text result for the user to view. In other embodiments, the second server 3 also stores other target models, for example, a visual target model and a multimodal target model. The first server 1 or the second server 3 can be a server, or a server cluster composed of several servers, or a cloud computing service center, etc., which is not limited in the embodiments of the present application.

[0046] The technical solution of the present application is described in detail below with specific embodiments. The specific embodiments described below may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0047] Figure 2 Schematic diagram of the flow of the information processing method provided in the embodiment of the present application. The method of this embodiment can be executed by the first server. Figure 2 As shown, the method of this embodiment may include:

[0048] 201: Receive an outgoing request sent by an electronic device, where the outgoing request is used to request that information to be processed be sent to a target model for processing.

[0049] In some embodiments of the present application, when it is necessary to use the target model to solve related problems, the electronic device can send different outbound requests to the first server according to different application scenarios. The outbound request includes information to be processed, and the outbound request is used to request that the information to be processed be sent to the target model in the second server for processing. Among them, the target model can be a large model. For example, when a user needs to answer a question, the user enters the corresponding question and the question "How to answer this question". The electronic device can receive the user's target question and the corresponding question, and send the target question and the corresponding question to the first server. For another example, when the user needs to obtain a summary of the meeting minutes, the electronic device can receive an outbound request, which includes the target question "Please output the summary of the following meeting minutes" entered by the user and the corresponding meeting minutes, and send the outbound request to the first server.

[0050] In some embodiments of the present application, the information to be processed may be text content, image content, and other information. Text content includes documents related to products or technical solutions, contracts, meeting minutes, legal documents, industry research reports, market analysis, academic papers, legal opinions, judgments, policy documents, business emails, correspondence between individuals or companies, product instructions, software operating manuals, corporate business plans, marketing strategies, management proposals, etc. Text content may also be content obtained after voice content conversion, and voice content includes human-computer dialogue, telephone recordings, consultations between doctors and patients, and recordings of diagnostic processes.

[0051] In some embodiments of the present application, the information to be processed includes at least a target question and a prompt word. In order to enable the target model to better understand the information to be processed and output an answer that matches the target question in the information to be processed, the electronic device can add prompt words to the information to be processed after receiving the information to be processed, so that the target model in the second server can better understand the target question in the information to be processed. For example, if the target question in the information to be processed is about the breach of contract clause in the contract, the prompt word may include keywords such as "breach of contract" and "liability". In this way, the keywords included in the prompt word can help the target model quickly locate the relevant contract terms and give accurate answers.

[0052] In some embodiments of the present application, prompt words may include task description, background knowledge, search path, task instructions, etc. When using the target model to retrieve the answer corresponding to the target question, the prompt words can be used to guide the target model to output the expected answer by adjusting the prompt words without changing the target model structure or retraining the target model.

[0053] In some embodiments of the present application, the type of information to be processed can be determined according to business scenarios, including extracting summaries, human-computer dialogue, and voice-to-text conversion.

[0054] 202: Respond to the outgoing request and identify the information to be encrypted in the information to be processed.

[0055] In some embodiments of the present application, since the information to be processed may include information to be encrypted, the electronic device directly sends the information to be processed to a target model (such as a target model in the second server) deployed outside the enterprise or in the cloud for processing, which may cause privacy leakage. Therefore, the electronic device first sends an outbound request to the first server, and the first server responds to the outbound request to identify the information to be encrypted in the information to be processed, and then sends the processed information to be encrypted to the target model in the second server for processing. In an embodiment of the present application, the pre-trained first model can be used to identify the information to be encrypted in the information to be processed.

[0056] In the embodiment of the present application, before identifying the information to be encrypted in the information to be processed, the information processing method includes preprocessing the information to be processed. The specific processing flow can be referred to below. Figure 3 Description of the flowchart shown. Before identifying the information to be encrypted in the information to be processed, the conventional types of information in the information to be processed can be preliminarily screened by preprocessing the information to be processed. In order to improve the accuracy and success rate of identifying the information to be encrypted, the information to be processed can be further processed by the pre-trained first model, so as to identify the information to be encrypted that cannot be effectively identified by the aforementioned preprocessing process. For example, unconventional types of information to be encrypted such as custom types. And improve the accuracy of all the identified information to be encrypted according to the context.

[0057] In the embodiment of the present application, the first model is a pre-trained model, and using the first model to identify the information to be encrypted in the information to be processed includes: using the second model to identify the named entities in the text information; inputting the named entities into the first model to obtain the information to be encrypted. The specific process of using the second model to identify the named entities in the text information is as follows: Figure 4 The method for training the first model includes establishing a database of information to be encrypted; constructing a training sample data set based on the database of information to be encrypted; and training the first model based on the training sample data set. The specific process of the method for training the first model is described below. Figure 5 Description of the flowchart shown.

[0058] In an embodiment of the present application, the first model may be a natural language processing (NLP) model, or a neural network language model (NNLM).

[0059] In some embodiments, the database of information to be encrypted may include a rule base and a label base. The rule base may be determined based on the reference data of the information to be encrypted. The rule base includes the confidentiality level, type, business scenario, and processing rules of the information to be encrypted. The label base includes the confidentiality level, business scenario, judgment result, and processing rules of the information to be encrypted, as shown in Table 1.

[0060] Table 1

[0061]

[0062]

[0063] In some embodiments of the present application, the types of information to be encrypted can be divided, and the confidentiality level of the information to be encrypted can be determined according to the division results, and the corresponding processing method can be determined according to the confidentiality level. In some embodiments, the types of information to be encrypted include personal privacy information, company confidential information and other information to be encrypted.

[0064] In some embodiments, personal privacy information includes name, phone number, ID number, age, photo, address, email address, bank card number, etc. Business scenarios involving personal privacy information include extracting summaries, human-computer dialogue, and voice-to-text conversion.

[0065] In some embodiments, company confidential information includes internal account passwords, positions, salaries, employee email addresses, work numbers, business confidential contracts, intellectual property rights, confidential customer lists, and trade secrets. Business scenarios involving company confidential information include extracting summaries, human-computer dialogue, and voice-to-text conversion.

[0066] In some embodiments, other information to be encrypted includes information that violates social morality or public order and good customs, terrorist information, etc.

[0067] In some embodiments of the present application, the confidentiality level of the information to be encrypted may also be determined based on the reference data of the information to be encrypted. The confidentiality level of the information to be encrypted includes top secret, confidential and secret.

[0068] In an embodiment of the present application, a rule base stores a correspondence table of named entities, types of named entities, and corresponding information to be encrypted. The information to be encrypted can be determined based on the correspondence table. For example, when the named entities in the text information of the information to be processed include information such as AA Technology, Rendering Service, and Contract, it can be determined that AA Technology is the company name, Rendering Service is the product, there is a contract in the title of the text information, and there is an amount in the content, thereby determining that the information to be processed is a service agreement and the information to be encrypted is a service agreement.

[0069] In the embodiment of the present application, the type of the named entity can also be determined. For example, when the named entity includes "Zhang San", the type of the named entity is determined to be a name; when the named entity includes "18888888888", the type of the named entity is determined to be a phone number; when the named entity includes "CC University", and combined with the context, the type of the named entity is determined to be educational experience; when the named entity includes "XX Company", and combined with the context content of "XX Company", the type of the named entity is determined to be work experience. When the named entity includes "AA Technology", the type of the named entity is determined to be a company; when the named entity includes "Renderbus / Rendering Service", the type of the named entity is determined to be a product.

[0070] 203: Encrypt the information to be encrypted to obtain the encrypted information to be processed.

[0071] In some embodiments of the present application, since the information to be processed contains information to be encrypted, it is not suitable to be sent to the target model for processing to avoid leaking relevant information and causing losses. Therefore, it is necessary to encrypt the information to be encrypted in the information to be processed. However, in order to ensure that the encrypted information to be processed does not affect the processing result of the target model, it is necessary to encrypt the information to be encrypted according to the state attributes and context of the information to be encrypted.

[0072] In an embodiment of the present application, encrypting the information to be encrypted to obtain the encrypted information to be processed includes: based on the type of the information to be encrypted, obtaining encrypted information corresponding to the type of the information to be encrypted; replacing the encrypted information in the information to be processed with the corresponding encrypted information to obtain the encrypted information to be processed. This can avoid information loss and reduce the accuracy of the target model processing. Specifically, the third model is used to identify the type of each named entity in the information to be encrypted and the corresponding accuracy; the corresponding weight is assigned based on the frequency of each named entity in the information to be processed; the corresponding score value is obtained through a weighted sum algorithm based on the accuracy and weight; the information to be encrypted is encrypted based on the score value to obtain the encrypted information to be processed.

[0073] In the embodiment of the present application, the third model is a pre-trained model for identifying named entity types. For example, the accuracy of identifying the type of the named entity "AA Technology" in the information to be encrypted as a company by the third model is 0.95; the accuracy of identifying the type of the named entity "Renderbus / Rendering Service" in the information to be encrypted as a product by the third model is 0.9; the accuracy of identifying the type of the named entity "Contract" in the information to be encrypted as a contract by the third model is 0.8; the accuracy of identifying the type of the named entity "AA City BB Street CC Community" in the information to be encrypted as an address by the third model is 0.8; the accuracy of identifying the type of the named entity "132XXXXXXXX" in the information to be encrypted as a contact number by the third model is 0.7; and so on. The corresponding weight based on the frequency of the named entity "AA Technology" appearing in the information to be processed is 0.3; the corresponding weight based on the frequency of the named entity "Renderbus / Rendering Service" appearing in the information to be processed is 0.32; the corresponding weight based on the frequency of the named entity "Contract" appearing in the information to be processed is 0.33; the corresponding weight based on the frequency of the named entity "AA City BB Street CC Community" appearing in the information to be processed is 0.01; the corresponding weight based on the frequency of the named entity "132XXXXXXXX" appearing in the information to be processed is 0.01; based on the accuracy and weight of each of the above named entities, the corresponding score value is 0.852 obtained through the weighted sum algorithm; the information to be encrypted is encrypt based on the score value to obtain the encrypted information to be processed.

[0074] In an embodiment of the present application, when the score value is greater than or equal to a preset value, it indicates that the information to be encrypted needs to be encrypted, and the corresponding target entity is determined from a pre-established password library according to the named entity; the named entity is replaced based on the target entity to obtain the encrypted information to be processed. When the score value is less than the preset value, it indicates that the information to be encrypted does not need to be encrypted. The pre-established password library stores target entities that can replace named entities. Among them, the target entity is the content generated according to the named entity and the scene corresponding to the named entity. The password library includes at least a name library, an address library, and a mobile phone number rule library. For example, when the input content is determined to be a personal resume according to the named entity and the type of the named entity, and the named entity includes "Zhang San", "Zhang San" is encrypted. Since the state attributes and context of the encrypted information to be encrypted in the input content need to be retained, the corresponding target entity, such as "Li Si", can be determined from the password library according to the named entity; and "Li Si" is replaced in the input content to obtain the encrypted input content.

[0075] In the embodiment of the present application, an information pair may also be generated based on the information to be encrypted and the encrypted information to be encrypted. The information pair includes a named entity and a target entity. For example, an information pair {"Zhang San":"Li Si"} is generated based on the named entity "Zhang San" and the target entity "Li Si".

[0076] In the embodiment of the present application, all the information to be encrypted in the information to be processed can also be encrypted separately, and the target entity used to encrypt the information to be encrypted can be saved in a code book for subsequent decryption. In order to ensure that the information is not leaked, the code book can also be encrypted. For example, the code book can be encrypted by a symmetric encryption algorithm and an asymmetric encryption algorithm.

[0077] 204: Send the encrypted information to be processed to the second server, so that the encrypted information to be processed is processed by the target model in the second server.

[0078] In the embodiment of the present application, due to the encryption method provided in the above embodiment, the encrypted information to be processed does not affect the understanding of the target model in the second server. Therefore, after the information to be processed is processed by the target model, the result desired by the user can be output. For example, when a user needs to obtain a summary of the meeting minutes, the target question "Please output the summary of the following meeting minutes" input by the user and the corresponding meeting minutes can be received. If the first model identifies that there is company confidential information waiting to be encrypted in the meeting minutes. The information to be encrypted can be encrypted, and the encrypted information to be processed can be obtained, and the encrypted information to be processed can be sent to the second server, and the encrypted information to be processed can be processed by the target model in the second server to obtain a summary of the meeting minutes and send the summary of the meeting minutes to the first server.

[0079] 205: Receive the processing result sent by the second server, and decrypt the processing result to obtain the target information.

[0080] In an embodiment of the present application, the processing result obtained by the target model in the second server includes the encrypted information to be encrypted. In order to restore the information to be processed, the information to be encrypted in the processing result can be decrypted to obtain the target information. Specifically, the processing result sent by the second server is received, and the encrypted information in the processing result is replaced based on the information to be encrypted to obtain the target information. Specifically, it includes: determining the information pair corresponding to the encrypted information to be encrypted in the processing result; obtaining the corresponding named entity based on the information pair; replacing the target entity based on the named entity to obtain the target information. For example, replacing "Zhang San" in the information to be processed based on "Li Si" to obtain the encrypted information to be processed. Generate an information pair {"Zhang San": "Li Si"} based on the named entity "Zhang San" and the target entity "Li Si". When decrypting, determine the information pair {"Zhang San": "Li Si"} based on the encrypted information to be encrypted "Li Si", determine the named entity "Zhang San" based on the information pair, and then replace the target entity "Li Si" based on the named entity "Zhang San". After obtaining the processing result through the target model, "Li Si" can be decrypted into "Zhang San" to obtain the target information.

[0081] In an embodiment of the present application, before decrypting the information to be encrypted in the processing result and obtaining the target information, the processing result can be checked again to see if the processing result contains the information to be encrypted. For example, if the processing result contains high-level information or harmful content, the user may be involved in legal issues. Specifically, the first server receives the processing result sent by the second server and identifies the processing result using the first model. When the processing result is determined by the first model to contain no information to be encrypted, the information to be encrypted in the processing result is decrypted to obtain the target information. When the processing result is determined by the first model to contain information to be encrypted, a prompt message is output to prompt the user that the target information cannot be returned, and the information security specialist is notified to manually review the processing result through the first server. In an embodiment of the present application, the information to be encrypted in the information to be processed is identified by the first model, and the information to be encrypted is encrypted to obtain the encrypted information to be processed. The encrypted information to be processed is processed using the target model to obtain the processing result, and the processing result is decrypted to obtain the target information. The information to be encrypted in the information to be processed can be intercepted, or the information to be encrypted in the information to be processed can be encrypted and concealed after identification, retaining the semantic environment of the context. It can not only prevent the leakage of encrypted information and protect information security, but also use the convenient services of the target model to obtain accurate target information.

[0082] Figure 3 A flow chart of a method for preprocessing information to be processed provided in an embodiment of the present application, such as Figure 3 As shown, the method for preprocessing the information to be processed may include the following steps:

[0083] 301, parse the information to be processed and obtain text information.

[0084] In some embodiments of the present application, in different application scenarios, the information to be processed input by the user through the electronic device includes different types of files. For example, in an application scenario where the user uses the target model to obtain a portrait of a job seeker based on the job seeker's resume, the information to be processed is the job seeker's resume, and the format of the information to be processed can be PDF format; for another example, in an application scenario where the user uses the target model to obtain the summary corresponding to the meeting minutes based on the meeting minutes, the information to be processed is the meeting minutes, and the format of the information to be processed can be Word format. In order to enable the target model to understand the information to be processed, it is necessary to parse and process different types of files to generate text information, which is conducive to the recognition of the target model. Among them, parsing the information to be processed includes preprocessing the information to be processed, extracting content from the information to be processed, and structuring the information to be processed. Specifically, preprocessing the information to be processed includes format conversion and encoding conversion of the information to be processed; extracting content from the information to be processed includes text extraction, image recognition, and table recognition; structuring the information to be processed includes layout recovery and document tree construction.

[0085] 302, processing the text information through pattern matching to obtain target text content.

[0086] In some embodiments of the present application, after obtaining text information by parsing the information to be processed, the content in the text information can be preliminarily screened to identify the content with a fixed format in the text information. Specifically, the electronic device pre-stores the template content, and the target text content is obtained by pattern matching the text information with the template content. For example, when the information to be processed includes a service agreement, the electronic device stores the template content corresponding to the service agreement. Among them, the template content corresponding to the service agreement includes basic contract information, service terms, and other additional terms. The basic contract information includes the contract title, the names and addresses of the two parties to the contract, the signing date, the contract period, and the contract object and scope; the service terms include service content, service period, service fees and payment methods, confidentiality clauses, breach of contract liability, contract termination clauses, and dispute resolution methods. Other additional terms include service guarantee clauses, intellectual property clauses, force majeure clauses, and supplementary agreements. By pattern matching the text information with the template content, the target text content including basic contract information, service terms, and other additional terms can be obtained.

[0087] In some embodiments, the starting position of the target text content in the text information can also be obtained through pattern matching. For example, when the target text content includes "Party A", the starting position of "Party A" in the text information is obtained as the first character of the second line.

[0088] In some embodiments, the target text content can also be obtained by processing text information with regular expressions. Regular expressions can identify a variety of information in the text, including strings of specific patterns, data formats, dates, email addresses, etc. For example, a string of specific patterns can be obtained by processing text information with regular expressions. Specifically, a string containing a specific feature word can be found through a regular expression, and the position matching the feature word in the text can be found. By processing text information with regular expressions to obtain data formats, for example, information with specific formats such as telephone numbers, postal codes, and product numbers can be matched. Dates, email addresses, URLs, special characters and symbols (e.g., spaces, punctuation marks, or non-printing characters) can also be obtained by processing text information with regular expressions.

[0089] 303, determine whether the information to be processed falls into the blacklist according to the target text content. If it is determined that the information to be processed does not fall into the blacklist, execute step 202; if it is determined that the information to be processed falls into the blacklist, execute step 304.

[0090] In an embodiment of the present application, information with a relatively high degree of confidentiality is pre-stored in a blacklist. For example, confidential customer information, information on dishonest personnel, or a list of dishonest enterprises, etc. In some embodiments of the present application, it can be determined whether the information to be processed falls into the blacklist based on the target text content. When it is determined that the information to be processed does not fall into the blacklist, step 202 is executed to call the first model to identify the information to be encrypted in the information to be processed. For example, when the target text content includes a telephone number, it is confirmed that the information to be processed does not fall into the blacklist, but it is not possible to determine whether the information to be processed contains other information to be encrypted, and it is necessary to call the first model to continue to identify the information to be encrypted in the information to be processed. When it is determined that the information to be processed falls into the blacklist, step 304 is executed to output a prompt message. For example, when the target text content includes the names of Party A and Party B, if it is determined that Party A or Party B falls into the blacklist and is not suitable for being sent to the external target model for processing, step 304 is executed to output a prompt message to prompt the user that the information to be processed cannot be sent.

[0091] 304, output prompt information.

[0092] In some embodiments of the present application, when it is determined that the information to be processed falls into the blacklist, a prompt message is output to prompt the user that the information to be processed involves information in the blacklist and cannot be sent to the target model.

[0093] In the embodiment of the present application, the target text information in the information to be processed is first screened by pattern matching, and it is determined whether the information to be processed falls into the blacklist according to the target text information. The information in the information to be processed can be preliminarily screened by pattern matching, and if the information in the information to be processed falls into the blacklist, the information to be processed is intercepted to prevent the leakage of the encrypted information.

[0094] Figure 4 A flow chart of a method for identifying named entities using a second model provided in an embodiment of the present application is shown in FIG. Figure 4 As shown, the method for identifying named entities using the second model may include the following steps:

[0095] 401: Divide the text information into units to obtain a plurality of division fields.

[0096] In the embodiment of the present application, the segmentation field refers to the basic unit of sentence segmentation. For example, a single Chinese character can be used as the most basic segmentation field, or a word after word segmentation processing can be used as the most basic segmentation field. Among them, the word segmentation processing can use the word segmentation tool in the existing related technology, such as the Jieba word segmentation tool. The goal of the word segmentation processing is to divide the sentence into multiple word segments.

[0097] 402: Assign a part-of-speech tag to each partition field.

[0098] In the embodiment of the present application, after obtaining multiple segmentation fields, in order to analyze the contextual semantics of the text information, a part-of-speech tag may be first assigned to each segmentation field to mark the part of speech of each segmentation field, such as verb, noun, adverb, conjunction, etc.

[0099] 403: Perform dependency analysis on each partition field after the part-of-speech tag is assigned.

[0100] In the embodiment of the present application, after assigning the part-of-speech tag, dependency analysis can be performed on each divided field to determine the relationship between the divided fields. The relationship between the divided fields includes the subject-predicate structure (SBV), the core relationship (Head of the Sentence, HED), the attribute-modifier relationship (ATT), the adverbial structure (ADV), the coordinate relationship (COO), the right adjunct relationship (RAD) and the verb-object relationship (VOB). Among them, the subject-predicate relationship SBV represents the structural relationship between the subject and the predicate in the sentence. In Chinese syntactic analysis, SBV is used to describe the relationship between the subject (Subject) and its corresponding predicate verb (Predicate). The core relationship HED refers to the core component of the sentence, which is usually the central meaning of the entire sentence. In dependency syntactic analysis, HED is used to mark the core words or phrases that dominate the other components of the sentence. The attribute-modifier relationship ATT represents the relationship between the modifier and the modified word. In Chinese syntactic analysis, ATT is used to describe the situation where an adjective (Attribute) modifies a noun (Modifier). The adverbial structure ADV refers to the role and position of an adverb or adverbial in a sentence. In dependency syntactic analysis, ADV is used to mark words or phrases used as adverbials, usually modifying verbs, adjectives or other adverbs. The parallel relationship COO represents the parallel relationship between two or more identical grammatical components. In Chinese syntactic analysis, COO is used to describe the relationship between parallel words, phrases or clauses. The right attachment relationship RAD refers to the relationship between the component attached to the right and the central word. In dependency syntactic analysis, RAD is used to mark those components that are located to the right of the central word and play a supplementary role. The verb-object relationship VOB represents the structural relationship between a verb and its direct object. In Chinese syntactic analysis, VOB is used to describe the relationship between a verb (Verb) and its direct object (Object).

[0101] For example, in the text information "This is a very long sentence", the division fields that can be obtained after unit division include: "this", "is", "one", "a", "very", "long", "of", and "sentence". After part-of-speech tagging of each division field, it can be obtained that: "this" is a pronoun, "is" is a verb, "one" is a numeral, "a" is a quantifier, "very" is an adverb, "long" is an adjective, "of" is a particle, and "sentence" is a noun. After performing dependency analysis, the relationship between each division field is determined as follows: Figure 6shown.

[0102] In some embodiments, each segmented field may be restored to its basic form, thereby eliminating semantic ambiguity caused by deformation of the segmented field such as tense, person, singular or plural, and thus improving the accuracy of identifying information to be encrypted.

[0103] 404: Perform boundary detection on the text information to obtain multiple sentences.

[0104] In the embodiment of the present application, the boundary of each sentence in the text information is identified according to the punctuation marks, and the text information is divided according to the identified boundary of each sentence to obtain multiple complete sentences. The sentences can be divided by searching for punctuation marks such as commas and periods in the text information.

[0105] 405: Identify the named entities in each sentence through a second model based on the result of the dependency analysis.

[0106] In the embodiments of the present application, named entities (named entities, or entities for short) refer to entities with specific meanings or strong referentiality in text, usually including names of people, places, names of organizations, dates and times, proper nouns, etc., and more extensive entities also include numbers, currencies, addresses, etc. Since it is difficult to effectively identify information to be encrypted in non-fixed formats (for example, names and addresses, etc.) through pattern matching, in order to improve the efficiency of identifying information to be encrypted in information to be processed, the named entities in each sentence can be identified through the second model, and then it is determined whether the information to be processed includes the information to be encrypted according to the information to be encrypted standard and the named entities.

[0107] In an embodiment of the present application, the second model is a named entity recognition (NER) model, which can also be called a named entity recognition system. The model can extract the above-mentioned named entities from the text information, and can identify more categories of named entities according to business needs. The second model can adopt a model of implementation methods such as feature template-based methods and neural network-based methods in the existing related technology. In one embodiment, the second model can adopt a model of Word Embedding + LSTM (Long Short-Term Memory Network) / BiLSTM (Bidirectional Long Short-Term Memory Network) + CRF (Conditional Random Field).

[0108] In an embodiment of the present application, in order to more accurately identify the information to be encrypted in the information to be processed, the named entities in the information to be processed can be identified through the second model, and the information to be encrypted in the information to be processed can be further screened.

[0109] Figure 5A flow chart of a method for training a first model provided in an embodiment of the present application is shown in FIG. Figure 5 As shown, the method for training the first model may include the following steps:

[0110] 501, establish a database of information to be encrypted.

[0111] In an embodiment of the present application, in order to train the first model, it is necessary to prepare source data, and reference data of the information to be encrypted can be established based on the information and documents within the enterprise, and a database of information to be encrypted can be established based on the reference data. Specifically, the user can define the information to be encrypted and the level according to relevant needs, obtain the reference data of the information to be encrypted, establish the database of information to be encrypted based on the reference data, and classify the information to be encrypted. The reference data of the information to be encrypted can be formulated with reference to industry standards and the needs of the user himself. For example, an enterprise user can determine the reference data of the information to be encrypted in collaboration with multiple departments within the enterprise based on industry standards. The enterprise stores reference data for determining the information to be encrypted, and the reference data of the information to be encrypted can be determined by an information security officer, and the reference data can be updated based on the data received in real time. The information to be encrypted extracted from the internal file based on the reference data is stored in the annotation library.

[0112] In an embodiment of the present application, the database of information to be encrypted may include a rule base and a notation base. The rule base may be determined based on the reference data of the information to be encrypted. The rule base includes the confidentiality level, type, business scenario, and processing rules of the information to be encrypted. The notation base includes the confidentiality level, business scenario, judgment result, and processing rules of the information to be encrypted, as shown in Table 1 above.

[0113] In some embodiments of the present application, source data for training the first model can also be prepared based on the log library. The log library can record information such as the user's work number, information to be encrypted, time, and audit classification results. Source data for training the first model can also be prepared based on an internal data lake and an external data source. Among them, the internal data lake is a data source within the enterprise, which stores blacklist information, public information, company information, and personal information within the enterprise. In an embodiment of the present application, entity data extraction is performed on the data in the internal data lake through the Neo4j image database. Among them, public information includes IP and language, etc.; company information includes contracts, company names, trade secrets, and intellectual property rights, etc.; personal information includes name, telephone number, and address, etc. The external data source is information obtained from an external platform of the enterprise. In an embodiment of the present application, the data in the external data lake is text-blocked or vectorized through a Neo4j vector index. For example, Company A, Company B, and Company C are listed as blacklisted companies by an organization, and the company has business dealings with Company A, Company B, and Company C. In order to prevent the target model enterprise in the second server from having any related connection with Company A, Company B and Company C, it is necessary to obtain the information of Company A, Company B and Company C from an organization through crawler technology and store it in an external data source.

[0114] 502, construct a training sample data set based on the information database to be encrypted, and pre-process the sample data set.

[0115] In the embodiment of the present application, natural language processing technology is used to pre-process the text data in the information database to be encrypted, including word segmentation, removal of stop words, part-of-speech tagging, etc. At the same time, the information to be encrypted can be classified according to different categories. Then the text data in the sample data set is annotated; the annotated training sample data set is divided into a training set, a verification set, and a test set. The training set is used when training the model, the verification set is used when verifying the model, and the test set is used when evaluating the model performance.

[0116] 503, establish an initial model.

[0117] In an embodiment of the present application, a machine learning algorithm is trained using a training sample data set to obtain an initial model, which includes at least an input layer, a convolutional layer, a pooling layer, a fully connected layer, and an output layer. In one example, the input layer may be followed by three convolutional layers with preset activation functions and three pooling layers, wherein each convolutional layer is followed by a pooling layer. Features of the training samples may be calculated through the convolutional layer, the pooling layer, and the fully connected layer. Among them, the number of convolutional layers may be set according to actual needs, and the preset activation function may be a Maxout activation function, etc., which is not specifically limited in the present application.

[0118] In an embodiment of the present application, after the input layer, the initial module can also be connected to a hash embedding layer for training information such as attributes of newly added input words.

[0119] 504, training an initial model based on the preprocessed sample data set.

[0120] In an embodiment of the present application, the initial model is iteratively updated and trained using the current batch of data in the training set; the test set is used to determine whether the initial model updated each time meets the preset conditions. If the preset conditions are not met, the preset model is updated again using the next batch of data according to the optimization algorithm (such as the gradient descent algorithm); the above steps are repeated until the initial model meets the preset conditions. Among them, the preset conditions can be that the loss value corresponding to the loss function is less than the preset loss threshold, the model performance (such as accuracy, recall rate, etc.) reaches the preset performance threshold, the number of iterations of the model reaches the preset number threshold, etc., wherein the above-mentioned various thresholds can be set according to actual needs, and this application does not impose specific restrictions on this.

[0121] 505, verify the initial model after training.

[0122] In the embodiments of the present application, during the training process, the model is verified and adjusted through a verification set, and the model parameters are continuously adjusted through an optimization algorithm so that the model can better fit the training data.

[0123] 506, evaluate the initial model after validation.

[0124] In an embodiment of the present application, after the trained initial model is verified by the verification set, the verified initial model is evaluated according to the model performance (e.g., accuracy, success rate, etc.), and if it is determined that the model performance meets the requirements, 507 is executed to obtain the first model; if it is determined that the model performance does not meet the requirements, return to 502. Among them, whether the model performance meets the requirements can be determined by determining whether the accuracy is greater than the preset accuracy. For example, if the accuracy is greater than or equal to the preset accuracy, it is determined that the model performance meets the requirements; if the accuracy is less than the preset accuracy, it is determined that the model performance does not meet the requirements.

[0125] 507, obtain the first model.

[0126] In the embodiment of the present application, if it is determined that the model performance meets the requirements, a trained first model can be obtained. In some embodiments, the initial model can also be trained by BERT / RoBERTa model to improve the accuracy of the first model in identifying the information to be encrypted.

[0127] See also Figure 7 , briefly summarizes the overall architecture diagram of the information processing method provided in the embodiment of the present application.

[0128] In order to solve the problem that when the user's information to be processed is sent to the target model for processing through the network, the encrypted information in the information to be processed is leaked, which poses a great threat to the user's privacy and security. The information processing method provided in this application can identify the encrypted information in the information to be processed and hide the encrypted information, so that the target model can accurately understand the information to be processed and output accurate results.

[0129] 701, receiving information to be processed sent by an electronic device.

[0130] In some embodiments of the present application, after the user passes the information to be processed by the electronic device, the information to be processed is sent to the first server for review, so as to determine whether there is information to be encrypted in the information to be processed by calling the first model in the first server. If it is determined that there is information to be encrypted in the information to be processed, the encrypted information is processed and then sent to the second server, and the target information corresponding to the information to be processed is output through the target model in the second server. For the specific process, please refer to Figure 2 201 in the above will not be repeated here.

[0131] It should be noted that the information to be processed includes at least a target question and a prompt word. In order to enable the target model in the second server to better understand the information to be processed and output an answer that matches the target question in the information to be processed, the electronic device can add prompt words to the information to be processed after receiving the information to be processed, so that the target model in the second server can better understand the target question in the information to be processed. For example, if the target question in the information to be processed is about the breach of contract clause in the contract, the prompt word can include keywords such as "breach of contract" and "liability". In this way, the keywords included in the prompt word can help the target model quickly locate the relevant contract terms and give accurate answers.

[0132] 702, review and classify the information to be processed.

[0133] In the embodiment of the present application, auditing and classifying the information to be processed includes: using the first model to identify the information to be encrypted in the information to be processed; and auditing and classifying the information to be encrypted. Specifically, the process of using the first model to identify the information to be encrypted in the information to be processed refers to: Figure 2 202 in the above will not be described in detail here. After identifying the information to be encrypted in the information to be processed, the information to be encrypted is reviewed and classified according to its confidentiality level. If it is determined that the confidentiality level of the information to be encrypted is top secret, 703 is executed to intercept the information to be processed, refuse to answer, and no longer send the information to be processed to the second server; if it is determined that the confidentiality level of the information to be encrypted is not top secret, 704 is executed to encrypt the information to be encrypted in the information to be processed, and then send the encrypted information to be processed to the second server.

[0134] 703, refused to answer.

[0135] 704, the encrypted information to be processed is processed through the target model to obtain a processing result.

[0136] In the embodiment of the present application, the detailed content of the information to be processed after the encryption processing is processed by the target model is referred to in Figure 2 203 in the above will not be repeated here.

[0137] 705, review and classify the processing results.

[0138] In the embodiment of the present application, auditing and classifying the processing results includes: using the first model to identify the information to be encrypted in the processing results; and auditing and classifying the information to be encrypted. Specifically, the process of using the first model to identify the information to be encrypted in the processing results is as follows: Figure 2 202 in the process will not be described here. After identifying the information to be encrypted in the processing result, the information to be encrypted is reviewed and classified according to its confidentiality level. If it is determined that the confidentiality level of the information to be encrypted is top secret, 706 is executed to refuse to return the target information; if it is determined that the confidentiality level of the information to be encrypted is not top secret, 707 is executed to decrypt the encrypted information to be encrypted in the processing result to obtain the target information, and the target information is returned to the first server.

[0139] 706, Refuse to return target information.

[0140] 707, return target information.

[0141] In the embodiment of the present application, after the information to be processed after encryption processing is processed by the target model and the processing result is obtained, the processing result can be checked again to see if the processing result contains the information to be encrypted. For example, if the processing result contains high-level information or harmful content of the announcement, the user may be involved in legal problems. Specifically, the first server receives the processing result sent by the second server and identifies the processing result using the first model. If it is determined by the first model that the processing result does not contain the information to be encrypted, the information to be encrypted in the processing result is decrypted to obtain the target information. If it is determined by the first model that the processing result contains the information to be encrypted, a prompt message is output to prompt the user that the target information cannot be returned, and the information security specialist is notified to manually review the processing result through the first server.

[0142] Figure 8 800 may include a plurality of functional modules composed of computer program segments. The computer programs of the various program segments in the information processing device 800 may be stored in a memory of a computer device and executed by at least one processor to perform (see Figures 2 to 4 Description) Information processing function.

[0143] In this embodiment, the information processing device 800 can be divided into multiple functional modules according to the functions it performs. The functional modules may include: a receiving module 801, an identification module 802, an encryption module 803, and a processing module 804. The module referred to in this application refers to a series of computer program segments that can be executed by at least one processor and can perform fixed functions, which are stored in a memory. In this embodiment, the information processing device 800 can be used to implement the following Figure 2 The information processing method shown. Figure 8 As shown, the encrypted information processing device 800 is applied to an electronic device (such as Figure 1 In the first server shown in FIG. 1 , the to-be-encrypted information processing device 800 includes:

[0144] The receiving module 801 is used to receive an outgoing request sent by an electronic device, where the outgoing request is used to request that the information to be processed be sent to a target model for processing;

[0145] The identification module 802 is used to respond to the outgoing request and identify the information to be encrypted in the information to be processed;

[0146] The encryption module 803 is used to obtain encryption information corresponding to the type of the information to be encrypted based on the type of the information to be encrypted; replace the information to be encrypted in the information to be processed with the corresponding encryption information to obtain the encrypted information to be processed;

[0147] The processing module 804 is used to send the encrypted information to be processed to the second server, so that the encrypted information to be processed is processed by the target model in the second server; and

[0148] The processing module 804 is further configured to receive the processing result sent by the second server, and replace the encrypted information in the processing result based on the information to be encrypted to obtain the target information.

[0149] In an embodiment of the present application, the processing module 804 is used to parse the information to be processed and obtain text information; process the text information through pattern matching to obtain target text content; and when it is determined based on the target text content that the information to be processed does not fall into the blacklist, call the first model to identify the information to be encrypted in the information to be processed.

[0150] In the embodiment of the present application, the recognition module 802 is used to use the second model to recognize the named entities in the text information; and input the named entities into the first model to obtain the information to be encrypted.

[0151] In an embodiment of the present application, the processing module 804 is used to use a third model to identify the type of each named entity in the information to be encrypted and the corresponding accuracy rate; assign a corresponding weight based on the frequency of occurrence of each named entity in the information to be processed; obtain a corresponding score value through a weighted sum algorithm based on the accuracy rate and the weight; encrypt the information to be encrypted based on the score value to obtain the encrypted information to be processed.

[0152] In an embodiment of the present application, the encryption module 803 is used to determine the corresponding target entity from a pre-established password library based on the named entity when the scoring value is greater than or equal to a preset value; replace the named entity based on the target entity to obtain the encrypted information to be processed.

[0153] In an embodiment of the present application, the processing module 804 is used to receive the processing result and identify the processing result using the first model; when it is determined through the first model that the processing result does not contain information to be encrypted, the information to be encrypted in the processing result is decrypted to obtain the target information.

[0154] In the embodiment of the present application, the processing module 804 is further configured to output a prompt message when it is determined that the information to be processed falls into the blacklist.

[0155] Fig. 9 1 is a schematic diagram of the structure of the first server provided in the embodiment of the present application. As shown in the figure, the first server 1 can be a server, or a server cluster composed of several servers, or a cloud computing service center, etc., which is not limited in the embodiment of the present application. The first server may include a communication module 101, a memory 102, a processor 103, an input / output (I / O) interface 104 and a bus 105. The processor 103 is coupled to the communication interface 101, the memory 102, and the I / O interface 104 respectively through the bus 105.

[0156] The communication module 101 may include a wired communication module and / or a wireless communication module. The wired communication module may provide one or more wired communication solutions such as universal serial bus (USB), controller area network bus (CAN, Controller Area Network). The wireless communication module may provide one or more wireless communication solutions such as wireless fidelity (wireless fidelity, Wi-Fi), bluetooth (bluetooth, BT), mobile communication network, frequency modulation (frequency modulation, FM), near field communication technology (near field communication, NFC), infrared technology (infrared, IR), etc.

[0157] The memory 102 may include one or more random access memories (RAM) and one or more non-volatile memories (NVM). The random access memory can be directly read and written by the processor 103, and can be used to store executable programs (such as machine instructions) of the operating system or other running programs, and can also be used to store user and application data. The random access memory may include static random-access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), etc.

[0158] The non-volatile memory may also store executable programs and user and application data, etc., and may be loaded into the random access memory in advance for direct reading and writing by the processor 103. The non-volatile memory may include a disk storage device and a flash memory.

[0159] The memory 102 is used to store one or more computer programs. The one or more computer programs are configured to be executed by the processor 103. The one or more computer programs include multiple instructions, and when the multiple instructions are executed by the processor 103, the information processing method executed on the first server 1 can be implemented.

[0160] In other embodiments, the first server 1 further includes an external memory interface for connecting to an external memory to expand the storage capacity of the first server 1.

[0161] The processor 103 may include one or more processing units, for example, the processor 103 may include an application processor (AP), a modem processor, a graphics processor (GPU), an image signal processor (ISP), a controller, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Different processing units may be independent devices or integrated into one or more processors.

[0162] The processor 103 provides computing and control capabilities. For example, the processor 103 is used to execute computer programs stored in the memory 102 .

[0163] The I / O interface 104 is used to provide a channel for user input or output. For example, the I / O interface 104 can be used to connect various input and output devices, such as a mouse, keyboard, touch device, display screen, etc., so that the user can enter information or visualize information.

[0164] The bus 105 is at least used to provide a channel for mutual communication among the communication module 101 , the memory 102 , the processor 103 , and the I / O interface 104 in the first server 1 .

[0165] It is to be understood that the structure illustrated in the embodiment of the present application does not constitute a specific limitation on the first server 1. In other embodiments of the present application, the first server 1 may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0166] An embodiment of the present application further provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to the methods in the above-mentioned embodiments of the present application.

[0167] The computer-readable storage medium may be the internal memory of the first server described in the above embodiment, such as the hard disk or memory of the first server. The computer-readable storage medium may also be an external storage device of the electronic device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., equipped on the first server.

[0168] In some embodiments, the computer-readable storage medium may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the electronic device, etc.

[0169] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0170] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0171] In the embodiments provided in the present application, it should be understood that the disclosed devices / terminal equipment and methods can be implemented in other ways. For example, the device / terminal equipment embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0172] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0173] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An information processing method, applied to a first server, characterized in that: The first server is connected to the second server via an external network, and the method includes: Receiving an outgoing request sent by an electronic device, wherein the outgoing request is used to request that the information to be processed be sent to a target model for processing; In response to the outgoing request, identifying the information to be encrypted in the information to be processed; Based on the type of the information to be encrypted, obtaining encryption information corresponding to the type of the information to be encrypted; Replacing the information to be encrypted in the information to be processed with the corresponding encrypted information to obtain the encrypted information to be processed; sending the encrypted information to be processed to the second server so that the encrypted information to be processed is processed by the target model in the second server; and The processing result sent by the second server is received, and the encrypted information in the processing result is replaced based on the information to be encrypted to obtain the target information.

2. The information processing method according to claim 1, characterized in that: The identifying the information to be encrypted in the information to be processed includes: Parsing the information to be processed and obtaining text information; Processing the text information through pattern matching to obtain target text content; When it is determined based on the target text content that the information to be processed does not fall into the blacklist, the first model is called to identify the information to be encrypted in the information to be processed.

3. The information processing method according to claim 2, characterized in that: The calling of the first model to identify the information to be encrypted in the information to be processed includes: identifying named entities in the text information using a second model; The named entity is input into the first model to obtain the information to be encrypted.

4. The information processing method according to claim 1, characterized in that: Determining to encrypt the information to be processed based on the type of the information to be processed to obtain the encrypted information to be processed includes: Using a third model to identify the type of each named entity in the information to be encrypted and the corresponding accuracy rate; assigning a corresponding weight based on the frequency of occurrence of each named entity in the information to be processed; Obtaining a corresponding score value through a weighted sum algorithm based on the accuracy and the weight; The information to be encrypted is encrypted based on the score value to obtain encrypted information to be processed.

5. The information processing method according to claim 4, characterized in that: The step of encrypting the information to be encrypted based on the score value to obtain the encrypted information to be processed comprises: In the case where the score value is greater than or equal to a preset value, determining a corresponding target entity from a pre-established password library according to the named entity; The named entity is replaced based on the target entity to obtain the encrypted information to be processed.

6. The information processing method according to claim 1, characterized in that: The receiving the processing result sent by the second server, and replacing the encrypted information in the processing result based on the information to be encrypted to obtain the target information includes: receiving the processing result, and identifying the processing result using the first model; When it is determined through the first model that the processing result does not contain information to be encrypted, the information to be encrypted in the processing result is decrypted to obtain target information.

7. The information processing method according to claim 6, characterized in that: The method further comprises: An information pair is generated according to the information to be encrypted and the encrypted information to be encrypted, wherein the information pair includes a named entity and a target entity.

8. The information processing method according to claim 7, characterized in that: The replacing the encrypted information in the processing result based on the information to be encrypted to obtain the target information includes: Determine an information pair corresponding to the encrypted information to be encrypted in the processing result; Obtaining the corresponding named entity based on the information pair; The target entity is replaced based on the named entity to obtain the target information.

9. The information processing method according to claim 2, characterized in that: The method further comprises: When it is determined that the information to be processed falls into the blacklist, a prompt message is output.

10. A server, characterized in that: The server includes a processor and a memory, and the processor implements the information processing method according to any one of claims 1 to 9 when executing a computer program stored in the memory.

Images