Information processing method, computing device, storage medium, and program product

Abstract

Embodiments of the present disclosure provide an information processing method, a computing device, a storage medium, and a program product. The information processing method comprises: acquiring content input by a user; in light of user feature information and knowledge base feature information, determining a target knowledge base that the user can access; searching the target knowledge base for target knowledge content matching the input content; and on the basis of the target knowledge content and the input content, using a large language model to generate feedback information. The technical solution provided by the embodiments of the present disclosure improves the accuracy of large language models, and ensures the data security.

Description

Information processing method, computing device, storage medium and program product

[0001] This disclosure claims priority to the Chinese patent application filed with the China Patent Office on December 20, 2023, with application number 202311770727.3 and application name “Information Processing Method and Computing Device,” the entire contents of which are incorporated by reference into this disclosure. Technical Field

[0002] The embodiments of the present disclosure relate to the field of artificial intelligence technology, and in particular to an information processing method, a computing device, a storage medium, and a program product. Background Art

[0003] With the development of technologies such as deep learning, big data, and cloud computing, various large-scale pre-trained language models have made great progress. Question-answering systems based on large language models have been widely used in customer service, medical health, education and training, and other fields.

[0004] Large language models are typically pre-trained using large amounts of data and adjusted using a variety of common instructions to facilitate human interaction. This allows them to handle a wide range of tasks and offers significant advantages in general domain tasks requiring extensive knowledge. However, for tasks in specific specialized fields, general large language models still have limitations due to a lack of specialized domain knowledge.

[0005] Summary of the Invention

[0006] The embodiments of the present disclosure provide an information processing method, a computing device, a storage medium, and a program product to solve the problem of low accuracy of large language models and ensure data security.

[0007] In a first aspect, an embodiment of the present disclosure provides an information processing method, including:

[0008] Get user input;

[0009] Determine the target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information;

[0010] Searching the target knowledge base for target knowledge content that matches the input content;

[0011] Feedback information is generated using a large language model based on the target knowledge content and the input content.

[0012] In a second aspect, an embodiment of the present disclosure provides a computing device, including a processing component and a storage component;

[0013] The storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the information processing method as described in the first aspect above.

[0014] In a third aspect, an embodiment of the present disclosure provides a computer-readable storage medium storing a computer program, which, when executed by a computer, implements the information processing method as described in the first aspect above.

[0015] In a fourth aspect, an embodiment of the present disclosure further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the information processing method as described in the first aspect above.

[0016] In the disclosed embodiment, for user input, the user's characteristic information and knowledge base characteristic information are first combined to determine a target knowledge base accessible to the user; target knowledge content matching the input content is searched from the target knowledge base; and feedback information is generated using a large language model based on the target knowledge content and the input content. Because knowledge bases have specific characteristics such as breadth, expertise, and timeliness, the technical solution of the disclosed embodiment enhances user input content, adding more prior knowledge to the user input content, making the feedback information obtained more professional and accurate, thereby improving the adaptability and accuracy of the large language model. The target knowledge content is determined only in the target knowledge base accessible to the user, thereby improving data security.

[0017] These and other aspects of the present disclosure will become more apparent from the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the embodiments of the present disclosure or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present disclosure. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] FIG1 shows a flow chart of an embodiment of an information processing method provided by the present disclosure;

[0020] FIG2 shows a flow chart of another embodiment of an information processing method provided by the present disclosure;

[0021] FIG3 shows a schematic diagram of scene interaction in a practical application of the technical solution of an embodiment of the present disclosure;

[0022] FIG4 shows a schematic structural diagram of an information processing device according to an embodiment of the present disclosure;

[0023] FIG5 shows a schematic structural diagram of an embodiment of a computing device provided by the present disclosure. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present disclosure.

[0025] In some of the processes described in the specification and claims of the present disclosure and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any order of execution. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit "first" and "second" to be different types.

[0026] The technical solutions of the embodiments of the present disclosure can be applied to application scenarios of large language models.

[0027] A large language model (LLM) is a natural language processing model with a very high number of parameters and computing power. It can be used for various natural language processing tasks such as text generation, machine translation, and question-answering systems. It is an AI (Artificial Intelligence) model.

[0028] Combined with the description of background technology, it can be seen that large language models perform well in general-purpose tasks, but they still have some shortcomings in terms of credibility, accuracy, and professionalism in some specific professional fields.

[0029] During the development of this disclosure, the inventors discovered that, given that large language models cannot effectively handle natural language tasks in specialized domains, it is possible to fine-tune large language models by integrating specific domain knowledge, thereby increasing their adaptability to those domains. However, due to the need to collect high-quality, annotated corpus data within a specific domain, the difficulty and high cost of data collection, coupled with the high level of expertise required for annotated data, make fine-tuning training difficult and costly.

[0030] In order to improve the adaptability and accuracy of the large language model and improve the applicability of the large language model, the inventor has proposed the technical solution of the embodiment of the present disclosure after a series of studies. In the embodiment of the present disclosure, there is no need to use professional field corpus to fine-tune the large language model. Instead, the knowledge base is used to determine the target knowledge content that matches the user input content, and the target knowledge content is used as the context information of the user input content to assist. Since the knowledge base has the characteristics of breadth, professionalism and timeliness, the technical solution of the embodiment of the present disclosure enhances the user input content, and can add more prior knowledge to the user input content, so that the feedback information obtained is more professional and accurate, thereby improving the large language model. Adaptability and accuracy. In addition, the inventors further discovered that in order to ensure the quality and scale of the knowledge base, the large language model in actual applications will integrate knowledge bases from different knowledge sources. Faced with multi-source heterogeneous knowledge, these knowledge bases face security issues. Therefore, in order to improve the adaptability and accuracy of the large language model while improving data security, in the embodiment of the present disclosure, the target knowledge base that the user can access can be first determined based on the user feature information and the knowledge base feature information, so that the target knowledge content is only determined in the target knowledge base, which meets the knowledge base's security requirements, improves data security, and can also distinguish knowledge usage permissions based on user identity to ensure safe sharing of the knowledge base.

[0031] The following will be combined with the accompanying drawings in the embodiments of the present disclosure to clearly and completely describe the technical solutions in the embodiments of the present disclosure. Obviously, the embodiments described are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments of the present disclosure, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present disclosure.

[0032] It should be noted that the embodiments of the present disclosure may involve the use of user data. In actual applications, user-specific personal data may be used in the scenarios described herein within the scope permitted by applicable laws and regulations, subject to compliance with applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0033] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0034] It should be noted that the technical solution of the embodiment of the present disclosure is applicable to a network virtual environment. The users described generally refer to "virtual users". Real users can register user accounts in the server through registration to obtain user identities in the network environment.

[0035] Figure 1 is a flowchart of an embodiment of an information processing method provided by an embodiment of the present disclosure, wherein the technical solution of the embodiment of the present disclosure can be applied to an information processing system composed of a user end and a server end in a practical application, wherein the user end can perceive the user input content, and the server end is used to call a large language model for processing.

[0036] The connection between the client and the server is established via a network. The network provides the medium for the communication link between the client and the server. The network can include various connection types, such as wired or wireless communication links or fiber optic cables. The client can interact with the server via the network to receive or send information.

[0037] In practical applications, the user end can be a browser, an APP (Application), or a web application such as an H5 (HyperText Markup Language 5) application, a light application (also known as a mini-program, a lightweight application), or a cloud application. The user end can be deployed in an electronic device and needs to rely on the device to run or certain apps in the device to run. For example, the electronic device can have a display and support information browsing, such as a personal mobile terminal such as a mobile phone, tablet computer, personal computer, desktop computer, smart speaker, smart watch, etc. Various other types of applications can also be configured in the electronic device, such as human-computer dialogue applications, model training applications, text processing applications, web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc. Electronic devices can refer to devices used by users that have the required computing, Internet access, communication, etc., such as mobile phones, tablet computers, personal computers, wearable devices, etc. Electronic devices generally include at least one processing component and at least one storage component. Electronic devices may also include basic configurations such as network card chips, IO buses, audio and video components, which are not limited in this disclosure. Optionally, depending on the implementation form of the electronic device, some peripheral devices may also be included, such as a keyboard, a mouse, an input pen, a printer, etc., which is not limited in this disclosure.

[0038] The server side may include servers that provide various services. It should be noted that the server side can be implemented as a distributed server cluster consisting of multiple servers or as a single server. The server can also be a server in a distributed system or a server integrated with blockchain. The server can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms, or an intelligent cloud computing server or intelligent cloud host equipped with artificial intelligence technology.

[0039] The technical solution of this embodiment can be specifically executed by the server. Of course, in other implementations, it can also be executed by the user. The large language model trained by the server can be deployed on the user, and the user provides more timely processing. This disclosure is not limited to this. The information processing method may include the following steps:

[0040] 101: Get user input.

[0041] 102: Determine the target knowledge base that the user can access by combining the user feature information and the knowledge base feature information.

[0042] Among them, the knowledge base is a special database used for knowledge management, which facilitates the collection, organization and extraction of relevant field knowledge. It is a data storage and organization method used to store, manage and retrieve structured and unstructured data. The knowledge base usually includes documents, pictures, audio, video, web pages and other content, which can be retrieved, filtered, sorted and classified by users.

[0043] In the technical solution of the embodiment of the present disclosure, the system may include multiple knowledge bases, which may be provided by one or more knowledge base providers. The multiple knowledge bases correspond to different knowledge sources. Therefore, it is possible to combine user feature information and knowledge base feature information to determine the target knowledge base that the user can access from multiple knowledge bases of different knowledge sources.

[0044] That is, the embodiments of the present disclosure can integrate knowledge bases from different knowledge sources, for example, aggregating professional data from multiple departments in an organization; or integrating external knowledge from different enterprises, institutions or individuals; or integrating domain knowledge from different fields such as medical health, education and training, etc.; this will face the situation where the sources of knowledge are dispersed and have different authority attributes. If these knowledge bases are combined into an overall knowledge base, there is a risk of data leakage. The inventors have found that, for example, in the same organization, the data of Department A can only be shared within the organization, while the data of Department B can be open to the outside world. In addition, different employees in the same organization should also be given different access rights according to their levels. Therefore, in order to improve data security, in the embodiments of the present disclosure, the user feature information and the knowledge base feature information can be first combined to determine the target knowledge base that the user can access and determine the target knowledge base with access rights. Among them, there can be at least one target knowledge base.

[0045] There are multiple implementations for determining the target knowledge base that a user can access by combining user feature information and knowledge base feature information, which will be described in detail in the following embodiments.

[0046] 103: Searching for target knowledge content matching the input content from the target knowledge base.

[0047] 104: Generate feedback information using a large language model based on the target knowledge content and input content.

[0048] In the embodiment of the present disclosure, target knowledge content that matches the input content is searched from the target knowledge base. The target knowledge content can serve as context information of the input content, thereby providing auxiliary information of the input content, making the feedback information generated by the large language model more accurate.

[0049] Optionally, first prompt information may be generated based on the target knowledge content and the input content; the first prompt information may be input into the large language model to obtain feedback information.

[0050] Prompts are a form of input used to prompt or guide the large speech model to produce expected output. They are used to instruct the large language model on what actions to take or what output to generate when performing a specific task. Prompts are natural language inputs, similar to commands or instructions, that let the large language model know what to do.

[0051] Optionally, the target knowledge content and input content can be entered into a prompt template to generate the first prompt message. For example, the prompt target template could be, "Please answer the user's question 'XX' based on the following background knowledge 'AA'." The 'AA' portion is used to enter the target knowledge content, and the 'XX' portion is used to enter the input content, thereby generating the first prompt message. The target knowledge content can serve as background knowledge for the input content, providing the large language model with the necessary knowledge, reducing factual errors in the large language model's reasoning, fully leveraging the large language model's powerful understanding and reasoning capabilities, and improving the accuracy of the large language model, thereby enhancing the user experience.

[0052] The large language model in the disclosed embodiment can be, for example, GPT-3 (Generative Pre-trained Transformer-3, a third-generation generative pre-trained model), GPT-4 (Generative General Pre-trained Transformer-4, a fourth-generation generative pre-trained model), BERT (Bidirectional Encoder Representation from Transformers, a bidirectional encoder model based on Transformers), Turing NLG (Turing Natural language Generation), etc. Large language models can be widely used in customer service, medical health, education and training and other fields. For example, an intelligent customer service robot built based on a large language model can automatically handle various service consultation issues of users; an intelligent question-and-answer system in the medical health field built based on a large language model can assist doctors in answering patients' questions; and an intelligent education software built based on a large language model can answer students' learning doubts based on a knowledge base. These applications utilize the powerful natural language processing capabilities of pre-trained large language models, greatly improving the question-and-answer experience and efficiency. In actual applications, for example, patient knowledge involved in the medical and health field often has a certain degree of privacy. Therefore, the technical solution of the embodiment of the present disclosure can grant different users access rights, so that the knowledge range that users can obtain only includes the knowledge base that they have access rights to. The large language model can provide answers based on the summary of the knowledge content that users can access, which can not only ensure the accuracy of the large language model, but also ensure that users cannot access knowledge to which they do not have access rights, thereby avoiding data leakage and ensuring data security.

[0053] As an optional implementation, the user characteristic information may include a user identifier, and the knowledge base characteristic information may include a knowledge base identifier. The above-mentioned combination of the user characteristic information and the knowledge base characteristic information to determine the target knowledge base that the user can access may include:

[0054] According to the user identifier and the knowledge base identifier, a first permission mapping relationship is searched to determine the target knowledge base to which the user has access rights; the first permission mapping relationship is used to indicate whether the user has access rights to the knowledge base.

[0055] The first permission mapping relationship may be preset to indicate whether the user has knowledge base access permission.

[0056] In some embodiments, first permission mapping relationships between the user and different knowledge bases may be determined based on a first determination result of whether the user meets first access requirements configured for different knowledge bases, and the first permission mapping relationships may be saved.

[0057] That is, a knowledge base provider can provide a knowledge base and perform configuration operations, and the first access requirements can be configured by the knowledge base provider for the knowledge base. If a user meets the first access requirements of any knowledge base, the user can be deemed to have access to the knowledge base; otherwise, the user can be deemed to have no access to the knowledge base. A first permission mapping relationship can be stored corresponding to the user identifier and the knowledge base identifier.

[0058] The user characteristic information may include not only the user identification but also at least one attribute characteristic, such as user name, organization, level, position, registration time, age, and gender, etc.

[0059] Therefore, it may be possible to combine the user identification and / or at least one attribute feature to determine whether the user meets the first access requirements respectively configured for different knowledge bases.

[0060] In some embodiments, a first permission mapping relationship between the user and different knowledge base providers may be defined based on the user registration information, and the first permission mapping relationship may be saved.

[0061] The user registration information may be generated when the user submits a registration request. In one implementation, the user characteristic information may include the user registration information. The user registration information may include the user identifier and one or more attribute characteristics in the above user characteristic information.

[0062] The first permission mapping relationship between the user and the different knowledge bases can be defined and saved by searching for the correspondence between different user registration information and different knowledge bases pre-configured by the system.

[0063] In addition, the user registration information may further include a desired knowledge base identifier, so that the knowledge base indicated by the desired knowledge base identifier can be determined to be a knowledge base that the user has permission to access, etc., in combination with the desired knowledge base identifier.

[0064] Among them, the first permission mapping relationship can be saved through a relational database, for example, it can be represented by a permission table, which can include fields such as user ID, knowledge base ID, and whether there is permission, so as to map the first rights of different users with different knowledge bases and save them in the permission table.

[0065] Of course, other information can also be stored in relational databases through other data tables. For example, the user table stores user-related information, which may include fields such as user ID, user name, user information addition time, and user information update time. The knowledge base table is used to store knowledge base-related information, which may include fields such as knowledge base ID, knowledge base name, knowledge base information addition time, knowledge base information update time, and first access requirements.

[0066] In addition, the first permission mapping relationship may also be marked through other types of databases such as a key-value database, etc., which is not limited in the present disclosure.

[0067] As another optional method, the user characteristic information may include a user identifier, and the knowledge base characteristic information may include a knowledge base identifier. The above-mentioned combination of the user characteristic information and the knowledge base characteristic information to determine the target knowledge base that the user can access may include:

[0068] According to the user identifier and the knowledge base provider identifier, a second permission mapping relationship is searched to determine the target knowledge base provider to which the user has access rights; and at least one knowledge base provided by the target knowledge base provider is used as the target knowledge base.

[0069] The second permission mapping relationship is used to indicate whether the user has access rights to the knowledge base provider.

[0070] In actual applications, a knowledge base provider may provide one or more knowledge bases. In the disclosed embodiment, a permission mapping relationship between a user and a knowledge base provider may be pre-set. If a user has permission to access a certain knowledge base provider, the user may access all knowledge bases provided by the knowledge base provider.

[0071] In some embodiments, a second permission mapping relationship between the user and the knowledge base provider may be determined based on a second determination result of whether the user meets the second access requirement corresponding to any knowledge base provider, and the second permission mapping relationship may be saved.

[0072] A knowledge base provider can provide knowledge bases and perform configuration operations. The second access requirements can be configured by the knowledge base provider. If a user meets the second access requirements of any knowledge base provider, the user can be deemed to have access to all knowledge bases provided by the knowledge base provider. Otherwise, the user can be deemed to have no access to any knowledge base provided by the knowledge base provider. A second permission mapping relationship can be stored corresponding to the user identifier and the knowledge base provider identifier. This second permission mapping relationship is used to indicate whether the user has access to the knowledge base provider, etc.

[0073] The user characteristic information may include not only the user identification but also at least one attribute characteristic, such as user name, organization, level, position, registration time, age, and gender, etc.

[0074] Therefore, it may be possible to determine whether the user meets the second access requirements respectively configured by different knowledge base providers by combining the user identification and / or at least one attribute feature.

[0075] In addition, in some embodiments, a second permission mapping relationship between the user and different knowledge base providers may be defined based on the user registration information, and the second permission mapping relationship may be saved.

[0076] Among them, user registration information can be created when the user submits a registration request. In one implementation, the user feature information can include at least part of the information in the user registration information. For example, the user registration information can include the user identifier and one or more attribute features in the above-mentioned user feature information.

[0077] The second permission mapping relationship between the user and different knowledge base providers can be defined and saved by searching for the correspondence between different user registration information and different knowledge base providers pre-configured by the system.

[0078] Among them, the second permission mapping relationship can be saved through a relational database, for example, it can be represented by a knowledge base provider table. The knowledge base provider table may include fields such as user ID, knowledge provider ID, and whether there is permission, so as to save the second permission mapping relationship between different users and different knowledge base providers in the knowledge base provider table.

[0079] In addition, the second permission mapping relationship may also be stored through other types of databases such as a key-value database, etc., which is not limited in this disclosure.

[0080] As another optional method, user feature information may include user identification and / or at least one attribute feature; knowledge base feature information may include knowledge base identification, knowledge base provider identification, first access requirements corresponding to the knowledge base, and second access requirements corresponding to the knowledge base provider to which the knowledge base belongs, etc.

[0081] The at least one attribute feature may be, for example, one or more of user name, organization, user level, position, age, gender, etc.

[0082] The above combination of user characteristic information and knowledge base characteristic information to determine the target knowledge base that the user can access may include:

[0083] Determining, based on the user characteristic information, whether the user meets at least one of the first access requirements corresponding to any knowledge base and the second access requirements corresponding to any knowledge base provider;

[0084] At least one knowledge base corresponding to the first access requirement and / or at least one knowledge base provided by a knowledge base provider whose user characteristic information meets the second access requirement are respectively used as target knowledge bases.

[0085] That is, in the embodiment of the present disclosure, the first permission mapping relationship or the second permission mapping relationship can be determined in advance based on the first access requirement or the second access requirement, and the first permission mapping relationship or the second permission mapping relationship can be searched based on the user identifier and the knowledge base identifier to determine whether the user has access rights to a certain knowledge base or a certain knowledge base provider.

[0086] Of course, the target knowledge base that the user can access may also be determined in real time based on the first access requirement or the second access requirement.

[0087] As can be seen from the above optional methods, the first access requirement or the second access requirement can be configured by the knowledge base provider. Therefore, in some embodiments, the apparatus may further include:

[0088] Sending a first registration prompt message to the knowledge base provider; generating knowledge base registration information according to the first registration request sent by the knowledge base provider.

[0089] The knowledge base registration information may include a knowledge base identifier, a knowledge base name, an identifier of a knowledge base provider to which the knowledge base belongs, and the first access requirement and / or the second access requirement, etc. The knowledge base characteristic information may include knowledge base registration information.

[0090] In some embodiments, the method may further include:

[0091] Send a second registration prompt message to the user terminal; create user registration information according to the registration request sent by the user terminal.

[0092] User registration information may include user identification, user ID, etc. User characteristic information may include user registration information.

[0093] In some embodiments, the first access requirement or the second access requirement may include one or more of the following access conditions:

[0094] having a configured predetermined user identification;

[0095] having one or more configured predetermined attribute characteristics;

[0096] as well as,

[0097] Within the validity period of the permission set by one or more of the above access conditions.

[0098] That is, the knowledge base provider can configure the scope of knowledge base development through user identification or attribute characteristics. For example, if a user has user identification X set by the knowledge base provider, the user can access all knowledge bases under the knowledge base provider; for example, if a user has attribute characteristics 1 and attribute characteristics 2 set by the knowledge base provider for knowledge base A, such as the user's position is general manager and belongs to the product department, then the user can access knowledge base A; for example, if a user has attribute characteristics 1 and attribute characteristics 2 set by the knowledge base provider for knowledge base A, and the access conditions are within the validity period of the permission, then it can be considered that the user can access knowledge base A.

[0099] In addition, as another optional method, the above-mentioned combination of user characteristic information and knowledge base characteristic information to determine the target knowledge base that the user can access may include:

[0100] The knowledge base whose matching degree between the knowledge base feature information and the user feature information meets the matching requirements is used as the target knowledge base that the user can access.

[0101] The knowledge base feature information and the user feature information can be vectorized and converted into vector features respectively, and then the degree of matching is determined by calculating the vector distance. For example, the degree of matching can be greater than a predetermined value.

[0102] Of course, the matching degree between the knowledge base feature information and the user feature information can also be calculated through a matching model. The matching model can be pre-trained based on sample knowledge base feature information and sample user feature information that matches it.

[0103] In the case where there is a knowledge base that the user can access, the input content can be directly input into the large language model to obtain feedback information.

[0104] In addition, in the absence of a user-accessible knowledge base, in order to further improve the accuracy of the large language model, in some embodiments, the method may further include:

[0105] When the target knowledge base is determined to be empty, at least one search engine is called to perform a network search based on the input content; the target search content is extracted from the search results; and feedback information of the input content is obtained using a large language model based on the target search content and the input content.

[0106] That is, the search can be performed by calling at least one search engine based on the input content through a network search method. The search results may include multiple web links, and the target search content can be extracted from the web content corresponding to at least one of the web links.

[0107] The second prompt information may be generated according to the target search content and the input content; and the second prompt information may be processed using a large language model to obtain feedback information of the input content.

[0108] The target search content and the input content may be input into a prompt template to generate the second prompt information.

[0109] The disclosed embodiment may also adopt a method of searching the entire network data, thereby combining the target search content as auxiliary information of the input content, thereby improving the processing accuracy of the large language model.

[0110] In addition, in the absence of a user-accessible knowledge base, in order to further improve the accuracy of the large language model, in some embodiments, the method may further include:

[0111] When the determination result of the target knowledge base is empty, the intention recognition model is used to identify the target intention corresponding to the input content; based on the target intention and the input content, the large language model is used to obtain feedback information of the input content.

[0112] That is, the embodiment of the present disclosure can improve the processing accuracy of the large language model by identifying the target intent corresponding to the input content and using the target intent as the context information of the input content.

[0113] Among them, the third prompt information can be generated according to the target intention and the input content; and the third prompt information can be processed using the large language model to obtain feedback information of the input content.

[0114] The target intention and input content can be filled into a pre-set prompt template to generate the third prompt information.

[0115] In addition, in some embodiments, the method may further include:

[0116] When the determination result of the target knowledge base is empty, the intention recognition model is used to identify the target intention corresponding to the input content;

[0117] In the case that the target knowledge base is determined to be empty, calling at least one search engine to perform a network search based on the input content; extracting the target search content from the search results;

[0118] Based on the target search content, target intent, and input content, a large language model is used to generate feedback information.

[0119] Here, the target search content, target intent and input content may be used to generate fourth prompt information, and the fourth prompt information may be processed using a large language model to obtain feedback information of the input content.

[0120] The target search content and target intent can serve as contextual information for the input content at the same time, adding more prior knowledge to the input content and making the feedback information obtained more accurate, thereby improving the accuracy of the large language model to assist.

[0121] Furthermore, in some embodiments, after obtaining target knowledge content that matches the input content from the target knowledge base, the target search content and target intent can be combined to enhance the input content. Therefore, feedback information can be generated using a large language model based on the target knowledge content, target search content, target intent, and input content.

[0122] In some embodiments, searching the target knowledge base for target knowledge content that matches the input content may include:

[0123] The input content is converted into a feature vector; the feature vector is searched for similarity in the vector data corresponding to the knowledge base content of the target knowledge base to determine the target knowledge content corresponding to the target vector data that meets the similarity requirements with the feature vector.

[0124] That is, the target knowledge content can be determined from the target knowledge base through vector retrieval technology.

[0125] Among them, the knowledge content in the knowledge base can be converted into vector data in advance and saved to the vector database, so that the target vector data matching the feature vector can be found from the vector database, and then the target knowledge content corresponding to the vector data can be determined from the knowledge base.

[0126] The similarity between the feature vector and different vector data may be determined by calculating vector distance or the like, and the similarity requirement may be, for example, that the similarity is greater than a specified threshold or the like.

[0127] Among them, vector data can be obtained by first converting the knowledge content into text format, switching it into multiple text blocks through word segmentation technology, and then vectorizing each text block to obtain the vector form of each text block. The text block, the corresponding digital vector, and text block related information such as the text to which the text block belongs and the vector library can be stored in a vector database. The corresponding target knowledge content is determined based on the text to which the text block corresponding to the target vector data that meets the similarity requirements with the feature vector belongs.

[0128] In an actual application, the technical solution of the embodiment of the present disclosure can be applied to an intelligent question-answering scenario, where the user input content is the user's question content; the feedback information is the answer content produced by the large language model; FIG2 shows a processing diagram of the technical solution of the embodiment of the present disclosure in an actual application. Based on the knowledge base provided by the knowledge base provider, the permission mapping relationship (first permission mapping relationship and / or second permission mapping relationship) with the user can be first determined 201;

[0129] Afterwards, the knowledge content in the knowledge base can be converted into text format such as TXT (Text) format 202; the text can be segmented into file blocks 203, and each text block can include, for example, 50-1000 words or symbols; then each text block can be vectorized, and the content of each file block can be converted into vector data in vector form 204; the text block, the corresponding vector data and the text block related information are stored in the vector database 205.

[0130] Before asking questions, the user may first register to generate user registration information, so that the above-mentioned permission mapping relationship can be determined in combination with the user registration information and knowledge base feature information.

[0131] Afterwards, the user's question content is obtained; the user's question content can be vectorized and converted into a corresponding feature vector 206.

[0132] Afterwards, based on the permission mapping relationship, the target knowledge base to which the user has access rights can be first determined, and the vector data corresponding to the target knowledge base can be obtained from the vector database. Then, the feature vector and the vector data corresponding to the target knowledge base can be searched for similarity 207 to find the target vector data with high semantic similarity, and then the target knowledge content corresponding to the target vector data can be determined 208.

[0133] The target knowledge base content and the user question content can be input into the template to generate the first prompt information 209 that is finally input into the large language model.

[0134] The first prompt information is input into the large language model 210, so that the response content can be obtained.

[0135] Among them, the above-mentioned method of vectorizing the input content or knowledge content can be implemented, for example, using a one-hot model, a bag of words model, a term frequency-inverse document frequency (TF-IDF), an N-gram model (N-Gram), a word-vector model (Word2vec), a document-vector model (Doc2vec), etc., and this disclosure does not specifically limit this.

[0136] In some embodiments, the obtaining of user input content may be obtaining user input content sent by a user terminal;

[0137] After the feedback information is generated using the large language model, the method may further include: sending the feedback information to the user terminal so that the user terminal displays the feedback information.

[0138] For ease of understanding, as shown in FIG3 , a schematic diagram of scene interaction in an actual application of an embodiment of the present disclosure is shown.

[0139] The server 301 can send a first registration prompt message to the client 302. The client 302 can face the knowledge base provider. The knowledge base provider can perform registration operations and configuration operations based on the first registration prompt message, and send a first registration request to the server 301. The server 301 can generate knowledge base registration information based on this. The knowledge base feature information can include at least part of the information in the knowledge base registration information.

[0140] The server 301 can send a second registration prompt message to the user 303. The user can perform registration operations and configuration operations based on the second registration prompt message, and send a second registration request to the server 301; the server 301 can generate user registration information based on this, and the user feature information can include at least part of the information in the user registration information.

[0141] The client terminal 303 may receive user input content and send the user input content to the server terminal 301 .

[0142] The server 301 can combine user feature information and knowledge base feature information to determine a target knowledge base accessible to the user, search the target knowledge base for target knowledge content that matches the input content, and then invoke the large language model to generate feedback information based on the target knowledge content and the input content. The specific execution operations of the server 301 can be found in the previous embodiments, such as the embodiment shown in FIG2 , and will not be repeated here.

[0143] The feedback information generated by the server 301 can be sent to the user terminal 303 so that the user terminal 303 can display the feedback information to the user and complete the intelligent question-answering operation.

[0144] In some embodiments, in order to further improve data security, etc., the above-mentioned combination of user characteristic information and knowledge base characteristic information to determine the target knowledge base that the user can access may include:

[0145] Importing input into a trusted execution environment;

[0146] In a trusted execution environment, the execution combines user feature information and knowledge base feature information to determine the target knowledge base that the user can access.

[0147] The above-mentioned searching for target knowledge content that matches the input content from the target knowledge base is also searching for target knowledge content that matches the input content from the target knowledge base in the trusted execution environment.

[0148] The above-mentioned use of the large language model to generate feedback information based on the target knowledge content and the input content may also be the use of the large language model to generate feedback information based on the target knowledge content and the input content in a trusted execution environment.

[0149] Of course, the knowledge content vectorization and user input content vectorization mentioned above can also be executed in a trusted execution environment.

[0150] In addition, the method may further include: exporting the feedback information from the trusted execution environment, so that the feedback information can be sent to the user terminal for the user to view.

[0151] The trusted execution environment may be, for example, a sandbox created in the device system; or the trusted execution environment may be created based on a security unit built into the CPU, such as Intel SGS ( It may be created by Intel Software Guard Extensions, ARM Trustzone (an embedded platform security technology) or AMD PSP (AMD Platform Security Processor); or it may be created by an external secure element (SE) configured in the device, etc., which is not specifically limited in this disclosure.

[0152] Among them, the trusted execution environment can be an operating environment that coexists with the operating system. It can run independently in the device, ensure the security of data and code in the trusted execution environment, and achieve the purpose of isolation from the external execution environment. The external execution environment referred to in this article can refer to the operating system of the device.

[0153] By combining the technical solution of the disclosed embodiment with the knowledge base knowledge content, the accuracy of the large language model is improved, avoiding the illusion that it is inconsistent with the input content or facts. In addition, the permission mapping relationship between different knowledge bases and different users can be defined, so that the scope of knowledge that users can obtain when asking questions only includes the knowledge bases to which they have permission to access. The large language model will summarize the questions and answers based on the knowledge base content accessible to the user, ensuring that users cannot access knowledge for which they do not have permission, thereby avoiding data leakage and ensuring data security.

[0154] FIG4 is a schematic structural diagram of an embodiment of an information processing device provided by an embodiment of the present disclosure. The device may include:

[0155] Content acquisition module 401, used to acquire user input content;

[0156] The knowledge base determination module 402 is used to determine the target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information;

[0157] The knowledge search module 403 is used to search the target knowledge content that matches the input content from the target knowledge base;

[0158] The information generation module 404 is used to generate feedback information based on the target knowledge content and the input content using a large language model.

[0159] In some embodiments, user characteristic information includes user identification; knowledge base characteristic information includes knowledge base identification; the above-mentioned knowledge base determination module is specifically used to search for the first permission mapping relationship based on the user identification and the knowledge base identification to determine the target knowledge base to which the user has access rights; the first permission mapping relationship is used to indicate whether the user has access rights to the knowledge base.

[0160] In some embodiments, the apparatus may further include:

[0161] The first relationship determination module is used to determine the first permission mapping relationship between the user and different knowledge bases based on the first judgment result of whether the user meets the first access requirements configured for different knowledge bases, and save the first permission mapping relationship; or, based on the user registration information, define the first permission mapping relationship between the user and different knowledge bases, and save the first permission mapping relationship.

[0162] In some embodiments, user characteristic information includes a user identifier; knowledge base characteristic information includes a knowledge base provider identifier; the above-mentioned knowledge base determination module is specifically used to search for a second permission mapping relationship based on the user identifier and the knowledge base provider identifier to determine the target knowledge base provider to which the user has access rights; the second permission mapping relationship is used to indicate whether the user has access rights to the knowledge base provider; and at least one knowledge base provided by the target knowledge base provider is used as the target knowledge base.

[0163] In some embodiments, the apparatus may further include:

[0164] The second relationship determination module is used to determine a second permission mapping relationship between the user and the knowledge base provider according to a second determination result of whether the user meets the second access requirement corresponding to any knowledge base provider, and save the second permission mapping relationship.

[0165] Alternatively, a second permission mapping relationship between the user and different knowledge base providers is defined according to the user registration information, and the second permission mapping relationship is saved.

[0166] In some embodiments, the user characteristic information includes a user identifier and / or at least one attribute characteristic; the knowledge base characteristic information includes a knowledge base identifier, a knowledge base provider identifier, a first access requirement corresponding to the knowledge base, and a second access requirement corresponding to the knowledge base provider to which the knowledge base belongs;

[0167] The above-mentioned knowledge base determination module is specifically used to combine user characteristic information to determine whether the user meets at least one judgment condition of the first access requirement corresponding to any knowledge base and the second access requirement corresponding to any knowledge base provider; and use at least one knowledge base whose user characteristic information meets the first access requirement and / or at least one knowledge base provided by the knowledge base provider whose user characteristic information meets the second access requirement as the target knowledge base.

[0168] In some embodiments, the apparatus may further include:

[0169] A first registration module, configured to send a first registration prompt message to a knowledge base provider;

[0170] According to the first registration request sent by the knowledge base provider, knowledge base registration information is generated; the knowledge base registration information includes the first access requirement and / or the second access requirement.

[0171] In some embodiments, the apparatus may further include:

[0172] The second registration module is used to send a second registration prompt message to the user terminal; create user registration information according to the registration request sent by the user terminal; the user feature information includes at least part of the information in the user registration information.

[0173] In some embodiments, the first access requirement or the second access requirement includes one or more of the following access conditions:

[0174] having a configured predetermined user identification;

[0175] having one or more configured predetermined attribute characteristics;

[0176] as well as,

[0177] During the validity period of the permissions set by one or more of the above access conditions.

[0178] In some embodiments, the knowledge base determination module may be specifically configured to select a knowledge base whose matching degree between knowledge base feature information and user feature information meets a matching requirement as a target knowledge base accessible to the user.

[0179] In some embodiments, the information generation module is specifically configured to generate first prompt information based on target knowledge content and input content; and input the first prompt information into the large language model to obtain feedback information.

[0180] In some embodiments, the apparatus may further include:

[0181] The first processing module is used to call at least one search engine to perform a network search based on the input content when the determination result of the target knowledge base is empty; extract the target search content from the search results; and use the large language model to obtain feedback information of the input content based on the target search content and the input content.

[0182] In some embodiments, the apparatus may further include:

[0183] When the determination result of the target knowledge base is empty, the intention recognition model is used to identify the target intention corresponding to the input content; based on the target intention and the input content, the large language model is used to obtain feedback information of the input content.

[0184] In some embodiments, the knowledge search module is specifically configured to convert input content into a feature vector;

[0185] The feature vector is searched for similarity in the vector data corresponding to the knowledge base content of the target knowledge base to determine the target knowledge content corresponding to the target vector data that meets the similarity requirement with the feature vector.

[0186] In some embodiments, the apparatus may further include:

[0187] A secure execution module, configured to import input content into a trusted execution environment, wherein the knowledge base is deployed in the trusted execution environment;

[0188] The above-mentioned knowledge base determination module is specifically executed in a trusted execution environment, combining user feature information and knowledge base feature information to determine the target knowledge base that the user can access;

[0189] The apparatus may further include: an information export module, configured to export the feedback information to the trusted execution environment.

[0190] The information processing device shown in FIG4 can execute the information processing method described in the embodiment shown in FIG1. ​​The implementation principle and technical effects thereof are not described in detail here. The specific manner in which the various modules and units of the information processing device in the above embodiment perform operations has been described in detail in the embodiment of the method and will not be elaborated on here.

[0191] The present disclosure also provides a computing device, as shown in FIG5 , which may include a storage component 501 and a processing component 502 ;

[0192] The storage component 501 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 502 to implement the information processing method described in the embodiment shown in FIG1 .

[0193] Of course, a computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0194] The input / output interface provides an interface between the processing component and peripheral interface modules, which may be output devices, input devices, etc. The communication component is configured to facilitate wired or wireless communication between the computing device and other devices.

[0195] The processing component may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented as one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above method.

[0196] The storage component is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0197] The display component may be an electroluminescent (EL) element, a liquid crystal display or a micro display having a similar structure, or a direct retinal display or a similar laser scanning display.

[0198] It should be noted that when the computing device is deployed as a server, it can be a physical device or an elastic computing host provided by a cloud computing platform. It can be implemented as a distributed cluster consisting of multiple servers or terminal devices, or as a single server or a single terminal device.

[0199] When the above-mentioned computing device is deployed on the user side, it can be specifically implemented as an electronic device. The electronic device can refer to a device used by the user and has the computing, Internet access, communication and other functions required by the user, such as a mobile phone, tablet computer, personal computer, wearable device, etc.

[0200] The present disclosure also provides a computer-readable storage medium storing a computer program. When executed by a computer, the computer program implements the information processing method described in the embodiment shown in FIG1 . The computer-readable medium may be included in the electronic device described in the above embodiment, or may exist independently and not be incorporated into the electronic device.

[0201] The present disclosure also provides a computer program product, comprising a computer program carried on a computer-readable storage medium. When executed by a computer, the computer program can implement the information processing method described in the embodiment shown in FIG1 . In such an embodiment, the computer program can be downloaded and installed from a network and / or installed from a removable medium. When executed by a processor, the computer program performs the various functions defined in the system of the present disclosure.

[0202] It should be noted that the embodiments of the present disclosure may involve the use of user data. In actual applications, user-specific personal data may be used in the scenarios described herein within the scope permitted by applicable laws and regulations, subject to compliance with applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0203] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0204] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0205] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, or of course, by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the existing technology can be embodied in the form of a software product. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or certain parts of the embodiments.

[0206] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than to limit them. Although the present disclosure has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present disclosure.

Claims

1. An information processing method, wherein: include: Get user input; Determine a target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information; Searching the target knowledge base for target knowledge content that matches the input content; Feedback information is generated using a large language model according to the target knowledge content and the input content.

2. The method according to claim 1, wherein: The user characteristic information includes a user identifier; the knowledge base characteristic information includes a knowledge base identifier; and the step of combining the user characteristic information and the knowledge base characteristic information to determine a target knowledge base that the user can access includes: According to the user identifier and the knowledge base identifier, a first permission mapping relationship is searched to determine the target knowledge base to which the user has access rights; the first permission mapping relationship is used to indicate whether the user has access rights to the knowledge base.

3. The method according to claim 2, wherein: Also includes: Determine first permission mapping relationships between the user and different knowledge bases according to a first determination result of whether the user meets first access requirements respectively configured for different knowledge bases, and save the first permission mapping relationships; Alternatively, a first permission mapping relationship between the user and different knowledge bases is defined according to the user registration information, and the first permission mapping relationship is saved.

4. The method according to claim 1, wherein: The user characteristic information includes a user ID; the knowledge base characteristic information includes a knowledge base provider ID; The determining of the target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information comprises: According to the user identifier and the knowledge base provider identifier, searching for a second permission mapping relationship to determine the target knowledge base provider to which the user has access rights; the second permission mapping relationship is used to indicate whether the user has access rights to the knowledge base provider; At least one knowledge base provided by the target knowledge base provider is used as the target knowledge base.

5. The method according to claim 4, wherein: Also includes: Determine a second permission mapping relationship between the user and the knowledge base provider according to a second determination result of whether the user meets a second access requirement corresponding to any knowledge base provider, and save the second permission mapping relationship; Alternatively, a second permission mapping relationship between the user and different knowledge base providers is defined according to the user registration information, and the second permission mapping relationship is saved.

6. The method according to claim 1, wherein: The user characteristic information includes a user identifier and / or at least one attribute characteristic; the knowledge base characteristic information includes a knowledge base identifier, a knowledge base provider identifier, a first access requirement corresponding to the knowledge base, and a second access requirement corresponding to the knowledge base provider to which the knowledge base belongs; the determination of the target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information includes: Determine, in combination with the user characteristic information, whether the user meets at least one of the first access requirements corresponding to any knowledge base and the second access requirements corresponding to any knowledge base provider; At least one knowledge base corresponding to the first access requirement for the user characteristic information and / or at least one knowledge base provided by a knowledge base provider for the second access requirement for the user characteristic information are respectively used as target knowledge bases.

7. The method according to claim 3, 5 or 6, wherein: Also includes: Sending a first registration prompt message to the knowledge base provider; Generate knowledge base registration information according to the first registration request sent by the knowledge base provider; The knowledge base registration information includes a first access requirement and / or a second access requirement.

8. The method according to claim 3, 5 or 6, wherein: Also includes: Sending a second registration prompt message to the user terminal; User registration information is created according to the registration request sent by the user terminal; the user characteristic information includes at least part of the information in the user registration information.

9. The method according to claim 6, wherein: The first access requirement or the second access requirement includes one or more of the following access conditions: having a configured predetermined user identification; having one or more configured predetermined attribute characteristics; as well as, During the validity period of the permissions set by one or more of the above access conditions.

10. The method according to any one of claims 1 to 9, wherein: Also includes: In the case where the determination result of the target knowledge base is empty, based on the input content, calling at least one search engine to perform a network search; Extract target search content from search results; According to the target search content and the input content, feedback information of the input content is obtained using the large language model.

11. The method according to any one of claims 1 to 9, wherein: Also includes: In the case where the determination result of the target knowledge base is empty, using an intention recognition model to identify the target intention corresponding to the input content; According to the target intention and the input content, feedback information of the input content is obtained using the large language model.

12. The method according to any one of claims 1 to 11, wherein: The step of searching the target knowledge base for target knowledge content that matches the input content includes: Converting the input content into a feature vector; The feature vector is searched for similarity in vector data corresponding to the knowledge base content of the target knowledge base to determine the target knowledge content corresponding to the target vector data that meets the similarity requirement with the feature vector.

13. The method according to any one of claims 1 to 12, wherein: Also includes: Importing the input content into a trusted execution environment; The determining of the target knowledge base that the user can access by combining the user characteristic information and the knowledge base characteristic information comprises: In the trusted execution environment, combining the user characteristic information and the knowledge base characteristic information to determine a target knowledge base that the user can access; After the feedback information is generated by using the large language model, the method further includes: The feedback information is exported to the trusted execution environment.

14. A computing device, wherein: including a processing component and a storage component; The storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the information processing method according to any one of claims 1 to 13.

15. A computer-readable storage medium, wherein: A computer program is stored, and when the computer program is executed by a computer, the information processing method according to any one of claims 1 to 13 is implemented.

16. A computer program product comprising a computer program, wherein: When the computer program is executed by a processor, the information processing method according to any one of claims 1 to 13 is implemented.

Images