Question and answer knowledge multiplexing method and device, equipment and medium

Abstract

The invention discloses a question and answer knowledge multiplexing method and device, equipment and a medium, and relates to the technical field of artificial intelligence. The method comprises the following steps: answering a target user question through a predefined decision tree instance to obtain target question and answer knowledge; the decision tree instance has a node for calling a private large model; storing the target question and answer knowledge into a knowledge object library, and determining whether the target question and answer knowledge has matched in-transit questions and answers; and if the target question and answer knowledge has the matched in-transit question and answer, multiplexing the target question and answer knowledge, and updating the popularity of the target question and answer knowledge in the knowledge object library. According to the embodiment of the invention, the answer effect of the locally deployed large language model can be improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method, device, equipment and medium for reusing question and answer knowledge. Background Art

[0002] As various large language models continue to be open sourced, more and more companies are trying to deploy private large models locally instead of directly calling third-party large models. For enterprises, local deployment can ensure that data does not leave the enterprise's firewall, thereby better controlling and protecting data security and privacy, which is especially important for industries such as finance and healthcare that have strict requirements on data security. On the other hand, enterprises can modify and expand the functions of the model more freely to meet specific business needs without relying on external suppliers.

[0003] However, private big models are not a main business for most companies, and the resource cost invested is quite limited. This leads to obvious deficiencies in reasoning efficiency and processing capabilities of locally deployed big models compared to third-party big models. A more obvious manifestation is that third-party big models can simultaneously handle a large number of individual user usage requests and interface call requests, while private big models can only handle a limited number of requests at the same time. Excessive processing requests can easily lead to unstable operation or even crash of private big models. Summary of the invention

[0004] The present invention provides a method, device, equipment and medium for reusing question-answering knowledge to improve the answer effect of a locally deployed large language model.

[0005] According to one aspect of the present invention, a method for reusing question-answer knowledge is provided, comprising:

[0006] Answer the target user's questions through a predefined decision tree instance to obtain target question-answering knowledge; the decision tree instance has a node that calls a private large model;

[0007] Storing the target question-answering knowledge in a knowledge object library, and determining whether the target question-answering knowledge has a matching in-transit question-answer;

[0008] If the target question and answer knowledge has a matching in-transit question and answer, the target question and answer knowledge is reused, and the popularity of the target question and answer knowledge is updated in the knowledge object library.

[0009] According to another aspect of the present invention, a device for reusing question-answer knowledge is provided, comprising:

[0010] A question-answer generation module is used to answer the target user's questions through a predefined decision tree instance to obtain target question-answer knowledge; the decision tree instance has a node that calls a private large model;

[0011] An in-transit matching module is used to store the target question and answer knowledge into a knowledge object library and determine whether the target question and answer knowledge has a matching in-transit question and answer;

[0012] The knowledge reuse module is used to reuse the target question and answer knowledge if there is a matching in-transit question and answer for the target question and answer knowledge, and to update the popularity of the target question and answer knowledge in the knowledge object library.

[0013] According to another aspect of the present invention, a computer program product is provided, including a computer program, wherein when the computer program is executed by a processor, the computer program implements the method for reusing question and answer knowledge according to any embodiment of the present invention.

[0014] According to another aspect of the present invention, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the question-and-answer knowledge reuse method described in any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the question-and-answer knowledge reuse method described in any embodiment of the present invention when executed.

[0016] The embodiment of the present invention reuses the generated question-answering knowledge. For similar user questions, there is no need to repeatedly run decision tree instances and call private large models to generate corresponding answers, which reduces the consumption of software and hardware resources in the process of answering similar questions and improves the answering effect of the large language model.

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

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 is a flow chart of a method for reusing question-answer knowledge according to an embodiment of the present invention;

[0020] Figure 2 is a flow chart of a method for reusing question-answer knowledge according to another embodiment of the present invention;

[0021] Figure 3 is a structural schematic diagram of a question-and-answer knowledge reuse device provided according to another embodiment of the present invention;

[0022] Figure 4 It is a schematic diagram of the structure of an electronic device implementing an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first", "second", etc. in the present invention are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices.

[0025] Figure 1 This is a flowchart of a method for reusing question-answer knowledge provided by an embodiment of the present invention. This embodiment is applicable to the situation where the generated knowledge is reused after the localized deployment of a large language model. The method can be executed by a question-answer knowledge reuse device, which can be implemented in the form of hardware and / or software, and can be configured in an electronic device with corresponding data processing capabilities. Figure 1 As shown, the method includes:

[0026] S110. Answer the target user's question through a predefined decision tree instance to obtain target question-answering knowledge.

[0027] S120: storing the target question and answer knowledge into a knowledge object library, and determining whether the target question and answer knowledge has matching in-transit questions and answers.

[0028] S130: If the target question and answer knowledge has a matching in-transit question and answer, the target question and answer knowledge is reused, and the popularity of the target question and answer knowledge is updated in the knowledge object library.

[0029] Among them, the question-and-answer in progress refers to the question-and-answer that has been processed but has not yet been answered. The decision tree instance has a node that calls a private large model. The knowledge object library is used to store newly generated and existing question-and-answer knowledge.

[0030] Specifically, the user questions that third-party big models have to deal with are relatively complex and rarely repeated, often involving various aspects such as circuits, history, art, engineering, and economics. The reuse of question-and-answer knowledge is not very meaningful. Compared with third-party big models, the types of user questions that private big models have to deal with are very limited, and most of them are related to a specific business of the enterprise. The value of reusing question-and-answer knowledge is more obvious.

[0031] Create and configure the @AIDP tag in the system in advance, and build a postprocessor based on the spring framework. When the spring framework starts, it determines whether the processing flow contains the @AIDP tag based on the loaded bean definition. If it does, it builds an AIDP container to store the metadata of the class, method, and tag involved in the @AIDP tag, and builds a decision tree instance based on the metadata information.

[0032] When receiving a user question, a question-answering session is created and a decision tree example is run to process the target user question and obtain the target answer to the target user question. The target user question and the corresponding target answer are associated and saved as target question-answering knowledge.

[0033] For the newly obtained target question and answer knowledge, on the one hand, the target question and answer knowledge is organized into a standard expression and output to the user. On the other hand, it is stored in the knowledge object library and determines whether there are similar questions of other users waiting for the system to answer, that is, whether there are matching in-transit questions and answers.

[0034] If so, it means that the target question-answering knowledge can be reused. By reusing the target question-answering knowledge, the corresponding answers to user questions in the in-transit question-answering can be obtained without re-running the decision tree instance, saving computing resources and reducing the processing pressure of the locally deployed large language model.

[0035] At the same time, as a cache space with limited storage capacity, the knowledge object library needs to remove the internal question and answer knowledge regularly. In the present invention, each time the question and answer knowledge in the knowledge object library is reused, its heat is increased by 1, so that when necessary, it can be determined which question and answer knowledge needs to be removed and which does not need to be removed based on the heat information.

[0036] The embodiment of the present invention reuses the generated question-answering knowledge. For similar user questions, there is no need to repeatedly run decision tree instances and call private large models to generate corresponding answers, which reduces the consumption of software and hardware resources in the process of answering similar questions and improves the answering effect of the large language model.

[0037] In the above-mentioned knowledge object library including a hot data area and a cold data area, storing the target question-answering knowledge into the knowledge object library includes:

[0038] Inserting the target question-answering knowledge into the cold data area of ​​the knowledge object library;

[0039] Accordingly, the method further includes:

[0040] For each question-and-answer knowledge in the cold data area, if the popularity data of the question-and-answer knowledge within a preset time meets the removal condition, the question-and-answer knowledge is removed from the cold data area;

[0041] If the popularity data of the question and answer knowledge within a preset time meets the migration condition, the question and answer knowledge is migrated from the cold data area to the hot data area.

[0042] Specifically, the knowledge object library is divided into hot data areas and cold data areas. When data is inserted into the two data areas, it is sorted according to the data popularity. The frontmost question and answer knowledge is the hottest question knowledge in the current area.

[0043] Question and answer knowledge will be inserted into the cold data area first, with an initial heat of 0. When there is knowledge reuse, its heat is increased by 1. The knowledge object library is monitored, and the window is sliding at a preset time (for example, 5 minutes). If the heat of a certain question and answer knowledge in the window is lower than the removal threshold (for example, 1), the question and answer knowledge is removed from the knowledge object library. If the heat of a certain question and answer knowledge in the window is greater than the migration threshold (for example, 50), the question and answer knowledge is migrated from the cold data area to the hot data area.

[0044] Based on the above embodiment, optionally, the method further includes:

[0045] If the area proportion of the hot data area in the knowledge object library exceeds a preset proportion, the question-answering knowledge in the hot data area is eliminated and optimized according to the popularity until the area proportion is lower than the preset proportion.

[0046] The management mechanism of hot data areas is different from that of cold data areas. For hot data areas, when the sliding window is executed, the area ratio of the hot data area in the knowledge object library is monitored. If the area ratio exceeds the preset ratio (for example, 30%), the question-answering knowledge in the hot data area is sorted in descending order according to the popularity, and then the question-answering knowledge is eliminated from the back to the front until the area ratio of the data area in the knowledge object library is lower than the preset ratio (for example, reaching 20%).

[0047] Figure 2 This is a flowchart of a method for reusing question-answer knowledge provided by another embodiment of the present invention. This embodiment is optimized and improved on the basis of the above embodiment. Figure 2 As shown, the method includes:

[0048] S210, creating a temporary variable library of question-answering knowledge, storing the target user's question in the temporary variable library; answering the target user's question in the temporary variable library through a predefined decision tree instance to obtain target question-answering knowledge.

[0049] Among them, the temporary variable library is a temporary storage space used to save all information related to user questions.

[0050] Specifically, for the target user question to be answered, the system first creates a temporary variable library of the question and answer object, puts the target user question in for identification, and starts scheduling in the decision tree instance to run the first node in the decision tree instance. If the knowledge type of the data returned by the first node is static knowledge, the static knowledge is determined as the target answer to the target user question, the target question and answer are obtained, and the question and answer ends.

[0051] If the knowledge type of the data returned by the first node is not static knowledge, it is necessary to acquire knowledge again based on the business system or private large model. A transition object needs to be created. The object contains the following fields: session id, question, answer (the method and parameters for the next step of execution). The transition object is placed in the transition object library and used as the context object of the decision tree instance to trigger the next decision tree node for execution. When the decision tree node is executed downward, the method for acquiring the next step of knowledge is obtained based on the answer of the context transition object. Based on the method for acquiring knowledge, the nodes of the decision tree are triggered downward. When each node is executed, the answer field in the transition object needs to be updated, and the next node is triggered until the target answer is obtained.

[0052] It should be noted that during the entire decision tree node triggering process, the decision tree is a triggering process of the entire process, but knowledge may be cyclical. The return of one piece of knowledge will depend on the result of another knowledge question. When each node is triggered, the node of the current context needs to be checked to determine whether to execute a second time or the knowledge result. The knowledge branch execution needs to be clearly specified. The current transition object needs to be upgraded to a semi-connected knowledge object, and a sub-session needs to be created to schedule the process. The semi-connected knowledge object is associated with the sub-session. The monitoring of the question and answer object library is added. Once the knowledge of the session id appears in the question and answer object library and the knowledge is obtained, the main session is triggered to execute downward. After all decision trees are executed, the knowledge obtained is identified as static knowledge, the results are updated, and the current knowledge object is moved to the question and answer object library.

[0053] S220: storing the target question-answering knowledge in a knowledge object library, and calculating similarities between the target question-answering knowledge and other user questions in a temporary variable library one by one.

[0054] S230. If there are other user questions whose similarity with the target question and answer knowledge is greater than a threshold, the other user questions are determined as in-transit questions and answers, and it is determined that the target question and answer knowledge has matching in-transit questions and answers; if there are no other user questions whose similarity with the target question and answer knowledge is greater than a threshold, it is determined that the target question and answer knowledge has no matching in-transit questions and answers.

[0055] Specifically, the temporary variable library stores all recent user questions to be answered and related information of these questions. After the target question and answer knowledge is stored in the knowledge object library, the feature vectors of other user questions in the temporary variable library and the feature vector of the target user question are obtained; the feature vectors of other user questions in the temporary variable library are calculated for similarity with the feature vector of the target user question one by one. If there are other user questions with a similarity greater than a threshold (for example, 0.8), the other user questions are determined as in-transit questions and answers, and it is determined that the target question and answer knowledge has a matching in-transit question and answer; if there are no other user questions with a similarity greater than the threshold with the target question and answer knowledge, it is determined that the target question and answer knowledge has no matching in-transit question and answer.

[0056] S240. If the target question and answer knowledge has a matching in-transit question and answer, a semi-connected object is constructed using the target question and answer knowledge and the session ID of the in-transit question and answer, and stored in a semi-connected object library; the target question and answer knowledge is reused as an answer to the in-transit question and answer through the semi-connected object.

[0057] Specifically, if a similar question is found, that is, an on-the-way question and answer, a semi-connected object containing the target question and answer knowledge and the session ID of the on-the-way question and answer is created and stored in the semi-connected object library to indicate that the target question and answer knowledge can be reused when processing the on-the-way question and answer. When the subsequent system processes the on-the-way question and answer, if it monitors the session ID of the on-the-way question and answer in the semi-connected object library, it obtains the target question and answer knowledge stored together with the session ID from the semi-connected object, and determines the target answer in the target question and answer knowledge as the answer to the on-the-way question and answer, thereby realizing the reuse of the target question and answer knowledge.

[0058] S250: Update the popularity of the target question and answer knowledge in the knowledge object library.

[0059] The embodiment of the present invention determines the on-the-way questions and answers by calculating the similarity, thereby achieving accurate screening of the on-the-way questions and answers.

[0060] Figure 3 A schematic diagram of the structure of a question-answer knowledge reuse device provided by another embodiment of the present invention. Figure 3 As shown, the device comprises:

[0061] A question-answer generation module 310 is used to answer the target user's question through a predefined decision tree instance to obtain target question-answer knowledge; the decision tree instance has a node that calls a private large model;

[0062] An in-transit matching module 320 is used to store the target question and answer knowledge into a knowledge object library and determine whether the target question and answer knowledge has a matching in-transit question and answer;

[0063] The knowledge reuse module 330 is used to reuse the target question and answer knowledge if there is a matching in-transit question and answer for the target question and answer knowledge, and to update the popularity of the target question and answer knowledge in the knowledge object library.

[0064] The question-and-answer knowledge reuse device provided in the embodiment of the present invention can execute the question-and-answer knowledge reuse method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0065] Optionally, the question and answer generating module 310 includes:

[0066] A variable library creation unit, used to create a temporary variable library of question-answering knowledge and store the target user's questions in the temporary variable library;

[0067] The question and answer generating unit is used to answer the target user questions in the temporary variable library through a predefined decision tree instance to obtain target question and answer knowledge.

[0068] Optionally, the in-transit matching module 320 includes:

[0069] A matching unit, used to calculate similarity between the target question-answering knowledge and other user questions in the temporary variable library one by one;

[0070] A first analyzing unit is configured to determine, if there are other user questions whose similarity with the target question-and-answer knowledge is greater than a threshold, the other user questions as in-transit questions and answers, and determine whether the target question-and-answer knowledge has a matching in-transit question and answer;

[0071] The second analyzing unit is used to determine that the target question and answer knowledge has no matching in-transit question and answer if there are no other user questions whose similarity with the target question and answer knowledge is greater than a threshold.

[0072] Optionally, the knowledge reuse module 330 includes:

[0073] A semi-connection construction unit, used to construct a semi-connection object using the target question-answering knowledge and the session ID of the in-transit question-answering, and store the object in a semi-connection object library;

[0074] A multiplexing unit is used to reuse the target question and answer knowledge as an answer to the in-transit question and answer through the semi-connected object.

[0075] Optionally, the knowledge object library includes a hot data area and a cold data area, and the knowledge reuse module 330 includes:

[0076] A storage unit, used for inserting the target question-answering knowledge into the cold data area of ​​the knowledge object library;

[0077] Accordingly, the device further comprises:

[0078] A removal module is used to remove each question-and-answer knowledge in the cold data area from the cold data area if the popularity data of the question-and-answer knowledge within a preset time meets the removal condition;

[0079] The migration module is used to migrate the question and answer knowledge from the cold data area to the hot data area if the popularity data of the question and answer knowledge within a preset time meets the migration condition.

[0080] Optionally, the device further comprises:

[0081] The elimination optimization module is used to eliminate and optimize the question-answering knowledge in the hot data area according to the heat level if the area proportion of the hot data area in the knowledge object library exceeds the preset proportion, until the area proportion is lower than the preset proportion.

[0082] The further described question-and-answer knowledge reuse device can also execute the question-and-answer knowledge reuse method provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0083] Figure 4 A schematic diagram of an electronic device 40 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0084] like Figure 4 As shown, the electronic device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In the RAM 43, various programs and data required for the operation of the electronic device 40 can also be stored. The processor 41, the ROM 42, and the RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0085] A number of components in the electronic device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the electronic device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0086] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 41 executes the various methods and processes described above, such as the multiplexing method of question-answering knowledge.

[0087] In some embodiments, the method for reusing question-and-answer knowledge may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the method for reusing question-and-answer knowledge described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to execute the method for reusing question-and-answer knowledge in any other appropriate manner (e.g., by means of firmware).

[0088] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0089] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0090] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, device, or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0091] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0092] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0093] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0094] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0095] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for reusing question-answer knowledge, characterized in that: The method comprises: Answer the target user's question through a predefined decision tree instance to obtain target question-answering knowledge; the decision tree instance includes a node for calling a private large model; Storing the target question-answering knowledge in a knowledge object library, and determining whether the target question-answering knowledge has a matching in-transit question-answer; If the target question and answer knowledge has a matching in-transit question and answer, the target question and answer knowledge is reused, and the popularity of the target question and answer knowledge is updated in the knowledge object library.

2. The method for reusing question-answer knowledge according to claim 1, characterized in that: The target user's question is answered by a predefined decision tree instance to obtain target question-answering knowledge including: Creating a temporary variable library of question-answering knowledge, and storing the target user's questions in the temporary variable library; The target user questions in the temporary variable library are answered through a predefined decision tree instance to obtain target question-answering knowledge.

3. The method for reusing question-answer knowledge according to claim 2, characterized in that: Determining whether the target question-and-answer knowledge has a matching in-transit question-and-answer comprises: Calculate the similarity between the target question-answering knowledge and other user questions in the temporary variable library one by one; If there are other user questions whose similarity with the target question-and-answer knowledge is greater than a threshold, the other user questions are determined as in-transit questions and answers, and it is determined that the target question-and-answer knowledge has matching in-transit questions and answers; If there are no other user questions whose similarity with the target question and answer knowledge is greater than a threshold, it is determined that the target question and answer knowledge has no matching in-transit question and answer.

4. The method for reusing question-answer knowledge according to claim 3, characterized in that: The reusing of the target question-answering knowledge includes: Constructing a semi-connected object with the question-and-answer knowledge and the session ID of the in-transit question-and-answer session, and storing the object in a semi-connected object library; The target question and answer knowledge is reused as an answer to the in-transit question and answer through the semi-connected object.

5. The method for reusing question-answer knowledge according to claim 1, characterized in that: The knowledge object library includes a hot data area and a cold data area, and storing the target question-answering knowledge into the knowledge object library includes: Inserting the target question-answering knowledge into the cold data area of ​​the knowledge object library; Accordingly, the method further includes: For each question-and-answer knowledge in the cold data area, if the popularity data of the question-and-answer knowledge within a preset time meets the removal condition, the question-and-answer knowledge is removed from the cold data area; If the popularity data of the question and answer knowledge within a preset time meets the migration condition, the question and answer knowledge is migrated from the cold data area to the hot data area.

6. The method for reusing question-answer knowledge according to claim 5, characterized in that: The method further comprises: If the area proportion of the hot data area in the knowledge object library exceeds a preset proportion, the question-answering knowledge in the hot data area is eliminated and optimized according to the popularity until the area proportion is lower than the preset proportion.

7. A question-answering knowledge reuse device, characterized in that: The device comprises: A question-answer generation module is used to answer the target user's questions through a predefined decision tree instance to obtain target question-answer knowledge; the decision tree instance has a node that calls a private large model; An in-transit matching module is used to store the target question and answer knowledge into a knowledge object library and determine whether the target question and answer knowledge has a matching in-transit question and answer; The knowledge reuse module is used to reuse the target question and answer knowledge if there is a matching in-transit question and answer for the target question and answer knowledge, and to update the popularity of the target question and answer knowledge in the knowledge object library.

8. The question-answer knowledge reuse device according to claim 7, characterized in that: The question and answer generation module comprises: A variable library creation unit is used to create a temporary variable library of question-answering knowledge and store the target user's questions in the temporary variable library; The question and answer generating unit is used to answer the target user questions in the temporary variable library through a predefined decision tree instance to obtain target question and answer knowledge.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the question-and-answer knowledge reuse method described in any one of claims 1-6.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the question-and-answer knowledge reuse method described in any one of claims 1 to 6 when executed.

Images