Cloud computing scheme solution information generation method and device, computer equipment and medium

Abstract

The invention relates to a cloud computing scheme solution information generation method and device, computer equipment and a medium. The method comprises the following steps: in response to a target instruction obtained by an operation of inputting to-be-solved information on a solution interface of a cloud computing solution system, extracting target entity data of the to-be-solved information; from a pre-constructed cloud computing knowledge graph, obtaining attributes of each category in multiple layers of target categories to which the target entity data belongs and target entity associated data of the target entity data; cloud computing scheme solution information matched with the to-be-solved information is generated according to attributes of all categories in the multiple layers of target categories and target entity associated data; and outputting cloud computing scheme solution information. By adopting the method, the efficiency of obtaining the cloud computing scheme solution information can be improved.

Description

Technical Field

[0001] The present application relates to the field of cloud computing technology, and in particular to a method, device, computer equipment and medium for generating information for a cloud computing solution. Background Art

[0002] Cloud computing technology can improve the level of enterprise informatization and plays a very important role in the development of enterprise informatization. In the field of cloud computing technology, massive cloud computing knowledge data has multiple data sources, multiple data structures, multiple data formats and multiple storage paths.

[0003] In the related art, when a user wants to know the cloud computing solution information for a cloud computing problem, the user needs to obtain cloud computing data multiple times and through multiple channels, and browse the obtained cloud computing data to find out the related information of the cloud computing related problems, and determine the cloud computing solution information based on the related information.

[0004] In this way, when a computer device presents a massive amount of cloud computing knowledge data, it is necessary for the computer device to obtain cloud computing solution information based on the large amount of cloud computing knowledge data browsed by the user before it can receive the cloud computing solution information input by the user. In other words, the efficiency of computer devices in obtaining cloud computing solution information in related technologies is low. Summary of the invention

[0005] Based on this, the present application provides a method, apparatus, computer equipment and medium for generating cloud computing solution information, which can improve the efficiency of obtaining cloud computing solution information.

[0006] In a first aspect, the present application provides a method for generating cloud computing solution information, the method comprising:

[0007] In response to a target instruction obtained by inputting information to be solved in a solution interface of the cloud computing solution system, extracting target entity data of the information to be solved;

[0008] Obtaining, from a pre-built cloud computing knowledge graph, the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity associated data of the target entity data;

[0009] Generate cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity association data;

[0010] Output cloud computing solution information.

[0011] In some embodiments, the method further comprises:

[0012] In response to a specified instruction obtained by inputting supplementary knowledge data in a knowledge supplement interface of the cloud computing solution system, sending the supplementary knowledge data to a target device;

[0013] In response to confirmation information of the supplementary knowledge data sent by the target device, the cloud computing knowledge graph is updated according to the supplementary knowledge data; the confirmation information is used to indicate that the supplementary knowledge data is accurate;

[0014] Output accurate indication information of supplementary knowledge data on the knowledge supplement interface.

[0015] In some embodiments, the method further comprises:

[0016] Preprocessing the acquired cloud computing knowledge data to determine cloud computing text data;

[0017] According to the attributes of each category in the multi-layer category obtained by category attribute recognition of cloud computing text data, the pattern layer of the cloud computing knowledge graph is constructed; the pattern layer represents the hierarchical association relationship of the attributes of each category in the multi-layer category; the multi-layer target category is a part of the categories in the multi-layer category;

[0018] Construct a cloud computing knowledge graph based on the model layer and cloud computing text data.

[0019] In some embodiments, preprocessing the acquired cloud computing knowledge data to determine cloud computing text data includes:

[0020] Identify text data and multimedia data in cloud computing knowledge data;

[0021] determining text data in the cloud computing knowledge data as first text data, and sending multimedia data to a target device;

[0022] Receiving second text data of the multimedia data sent by the target device; the second text data is text description information of the multimedia data;

[0023] The first text data and the second text data are included in the cloud computing text data.

[0024] In some embodiments, according to the attributes of each category in the multi-layer category obtained by performing category attribute recognition on the cloud computing text data, a pattern layer of the cloud computing knowledge graph is constructed, including:

[0025] Obtaining architectural information of a pre-built pattern layer; the architectural information represents a hierarchical association relationship;

[0026] Perform category attribute recognition on cloud computing text data to obtain the attributes of each category in the multi-layer category;

[0027] The attributes of each category in the multi-layer category are mapped to the architectural information to obtain the pattern layer.

[0028] In some embodiments, the attributes of each category in the multi-layer category are mapped to the architecture information to obtain the pattern layer, including:

[0029] Perform category concept recognition on cloud computing text data to obtain the concepts of each category in the multi-layer category;

[0030] Analyze the concepts of each category in the multi-layer category to obtain the hierarchical relationship of the categories in the multi-layer category;

[0031] According to the hierarchical association relationship of categories in the multi-layer category, the attributes of each category in the multi-layer category are mapped to the architecture information to obtain the pattern layer.

[0032] In some embodiments, category attribute recognition is performed on cloud computing text data to obtain attributes of each category in the multi-layer category, including:

[0033] A large language model is used to identify the category attributes of part of the cloud computing text data, and the attributes of each category in the multi-layer category of the part of the data are obtained;

[0034] In response to modification prompt information for the attributes of each category in the multi-layer category of the partial data, modify the attributes of each category in the multi-layer category of the partial data;

[0035] The large language model is trained using the attributes of each category in the multi-layer categories of partial data and modified partial data;

[0036] The trained large language model is used to perform category attribute recognition on another part of the cloud computing text data to obtain the attributes of each category in the multi-layer categories of the other part of the data.

[0037] In some embodiments, a cloud computing knowledge graph is constructed based on the model layer and the cloud computing text data, including:

[0038] Extract entity data and entity-related data corresponding to the attributes of each category in the multi-layer categories in the pattern layer from the cloud computing text data;

[0039] Construct the data layer of cloud computing knowledge graph based on entity data and entity-related data;

[0040] Construct a cloud computing knowledge graph based on the model layer and data layer.

[0041] In some embodiments, a data layer of a cloud computing knowledge graph is constructed based on entity data and entity-related data, including:

[0042] Disambiguate the entity data to obtain the target data;

[0043] Determine the target association relationship based on the target data and entity association data;

[0044] Perform knowledge reasoning on the target data and target association relationships to obtain the data layer of the cloud computing knowledge graph.

[0045] In some embodiments, the method further comprises:

[0046] Identify structured and semi-structured data in cloud computing knowledge data;

[0047] Perform structural transformation on structured data and semi-structured data to obtain a first knowledge graph;

[0048] Based on the model layer and cloud computing text data, a cloud computing knowledge graph is constructed, including:

[0049] Construct the second knowledge graph based on the model layer and cloud computing text data;

[0050] The first knowledge graph and the second knowledge graph are integrated to obtain a cloud computing knowledge graph.

[0051] In some embodiments, the solution interface includes a cloud computing knowledge query interface, the information to be solved includes cloud computing file description information, and the cloud computing solution information includes link information of the target cloud computing file; or,

[0052] The solution interface includes a cloud computing dialogue interface, the information to be solved includes cloud computing question information, and the cloud computing solution information includes cloud computing answer information; or,

[0053] The solution interface includes a cloud configuration solution generation interface, and the information to be solved includes the demand information of the cloud configuration solution; the cloud computing solution solution information includes the cloud configuration solution result.

[0054] In a second aspect, the present application provides a cloud computing solution information generation device, the cloud computing solution information generation device comprising:

[0055] An extraction module, configured to extract target entity data of the information to be solved in response to a target instruction obtained by inputting the information to be solved in a solution interface of the cloud computing solution system;

[0056] An acquisition module, used to acquire the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity associated data of the target entity data from a pre-built cloud computing knowledge graph;

[0057] A generation module, for generating cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity association data;

[0058] The output module is used to output cloud computing solution information.

[0059] In a third aspect, the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of any of the above methods when executing the computer program.

[0060] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, characterized in that when the computer program is executed by a processor, the steps of any of the above methods are implemented.

[0061] In the technical solution of the embodiment of the present application, the computing device generates cloud computing solution information matching the information to be solved based on the acquired information to be solved and the pre-built cloud computing knowledge graph, and outputs the cloud computing solution information. In this way, the cloud computing solution information matching the information to be solved can be output by inputting the information to be solved once. Compared with the method in the related art that the computer device needs to obtain and browse a large amount of cloud computing knowledge data multiple times and through multiple channels and analyze the cloud computing solution information before it can obtain the cloud computing solution information input by the user, this can greatly shorten the time for obtaining the cloud computing solution information, thereby improving the efficiency of obtaining the cloud computing solution information. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A flowchart of a method for generating cloud computing solution information provided in some embodiments;

[0064] Figure 2 A flowchart of a method for supplementing knowledge data provided in some embodiments;

[0065] Figure 3 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some embodiments;

[0066] Figure 4 A flowchart of a method for determining cloud computing text data provided by some embodiments;

[0067] Figure 5 A schematic diagram of a process for constructing a model layer of a knowledge graph provided in some embodiments;

[0068] Figure 6 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some other embodiments;

[0069] Figure 7 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some further embodiments;

[0070] Figure 8 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some further embodiments;

[0071] Fig. 9 A schematic diagram of a process for obtaining multi-source multi-modal data and preprocessing the data provided in some embodiments;

[0072] Fig.10 A schematic diagram of a process for constructing a domain scenario ontology provided in some embodiments;

[0073] Fig.11 A schematic diagram of system functional division provided for some embodiments;

[0074] Fig.12 A schematic diagram of the structure of a cloud computing solution information generation device provided in some embodiments of the present application;

[0075] Fig.13 A schematic diagram of the structure of a computer device is provided for some embodiments. DETAILED DESCRIPTION

[0076] The following embodiments of the technical solution of the present application are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0077] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0078] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0079] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0080] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0081] Before introducing the cloud computing solution information generation method according to the embodiment of the present application, the terms involved in the embodiment of the present application are explained first:

[0082] Ontology: Ontology is a set of concepts or terms used to describe or express knowledge in a certain field. It can be used to organize higher-level knowledge abstractions in a knowledge base, or it can be used to describe knowledge in a specific field.

[0083] Region: An independent geographical area is called a Region. Regions in a public cloud are designed to be completely physically isolated from each other. Most resources belong to a certain Region, and cross-region resource isolation is provided. Communication between regions is done through the public network (Internet) or a dedicated cross-region data channel.

[0084] Availability Zone (AZ): Multiple isolated locations in the same region are called multiple availability zones in the region. Availability zones are introduced to avoid availability failures (natural disasters, power failures in the data center, or operator network failures, etc.) of all instances hosted in the same location in the data center. Users can deploy their instances and resources to multiple AZs in the same region to help avoid the impact of such failures. For example, a public cloud can map one AZ to one data center.

[0085] Cluster: A logical business concept. A cluster is a group of independent computers that are connected via a high-speed network. They form a group and are managed as a single system. Cluster configuration is used to improve availability and scalability.

[0086] Virtual Private Cloud (VPC): A private network that can build a secure and isolated network environment.

[0087] Load Balance (LB) platform: refers to distributing client requests to different service units. For example, there may be 4-layer and 7-layer load balancing.

[0088] Natural Language Processing (NLP): It is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can enable effective communication between humans and computers using natural language.

[0089] Hybrid Cloud Solution Design System (HCSDS): This article uses a custom abbreviation to describe the system designed in this article. It can also be called a cloud computing solution system, or the hybrid cloud solution design system is included in the cloud computing solution system.

[0090] The computer device may include one or a combination of at least two of the following: a server, a mobile phone (MobilePhone), a tablet computer (Pad), a computer with wireless transceiver function, a PDA, a desktop computer, a personal digital assistant, a portable media player, an intelligent speaker, a navigation device, a smart watch, a smart glasses, a smart necklace and other wearable devices, a pedometer, a digital TV, a virtual reality (VirtualReality, VR) terminal device, an augmented reality (AugmentedReality, AR) terminal device, a wireless terminal in industrial control (Industrial Control), a wireless terminal in self-driving (Self Driving), a wireless terminal in remote medical surgery (Remote Medical Surgery), a wireless terminal in smart grid (Smart Grid), a wireless terminal in transportation safety (Transportation Safety), a wireless terminal in a smart city (Smart City), a wireless terminal in a smart home (Smart Home), a car in a vehicle networking system, a vehicle-mounted device, a vehicle-mounted module, and the like.

[0091] Figure 1 A flow chart of a method for generating cloud computing solution information provided in some embodiments, the method being applied to a computer device, wherein the method for generating cloud computing solution information includes:

[0092] S101 , in response to a target instruction obtained by inputting information to be resolved in a solution interface of a cloud computing solution system, extracting target entity data of the information to be resolved.

[0093] In some embodiments, the cloud computing solution system may include a hybrid cloud solution design system. Exemplarily, the cloud computing solution system may be a hybrid cloud solution design system. In some embodiments, the cloud computing solution may be a hybrid cloud solution, etc. In other embodiments, the cloud computing solution may include at least one of the following: a hybrid cloud solution, a private cloud solution, a public cloud solution, etc.

[0094] In some embodiments, the information to be resolved may include content related to the cloud computing field that the user is concerned about. Exemplarily, the information to be resolved may be one of the following: "What certificate is required for hybrid cloud solution project A", "What qualification materials are required for cloud computing solution project B", "Are the numbers of the following cloud computing related content standardized?", "What is the server architecture of hybrid cloud solution project C?", "Provide an example of the certificate used in project A", etc. It should be noted that the information to be resolved may be any information that users need to resolve in the field of cloud computing, and this article will not list them one by one.

[0095] Exemplarily, entity data refers to the description of specific objects, which may be computer rooms, servers, assets, switches, etc. Taking the information to be solved as "What certificates are needed for hybrid cloud solution project A", entity data may include entities such as project A and certificates. Taking the information to be solved as "What is the server architecture of hybrid cloud solution project C", entity data may include entities such as project C and servers.

[0096] In some embodiments, the computer device may be installed with a cloud computing solution system and be able to display a solution interface of the cloud computing solution system. The user may input information to be resolved in the solution interface, and the computer device extracts target entity data of the information to be resolved in response to the user inputting the information to be resolved.

[0097] In other embodiments, the computer device may receive a target instruction sent by the electronic device, and the target instruction carries the information to be solved. For example, the electronic device may be installed with a cloud computing solution system and be able to display a solution interface of the cloud computing solution system. The user may input the information to be solved in the solution interface. The electronic device generates a target instruction in response to the user inputting the information to be solved, and sends the target instruction to the computer device. Exemplarily, the electronic device is connected to the computer device for communication, and the electronic device may be any device with a display function, for example, the electronic device may be a mobile phone or a computer.

[0098] S102. Obtain, from a pre-built cloud computing knowledge graph, the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data.

[0099] In some embodiments, the pre-built cloud computing knowledge graph may be continuously updated, for example, the cloud computing knowledge graph may be updated according to newly obtained cloud computing knowledge data. The pre-built cloud computing knowledge graph used in S102 may be the cloud computing knowledge graph currently stored in the computer device.

[0100] In some embodiments, the cloud computing knowledge graph may include attributes of each category in multiple layers of categories, entity data under each category, and association data between entity data under each category.

[0101] For example, each category of a multi-layer category indicates that there is a hierarchical relationship between each category in the multi-layer category. The hierarchical relationship can also be called a structural hierarchy. The hierarchical relationship of each category indicates the subordinate relationship of each category. For example, each category of a multi-layer category includes Block, POD, AZ, Cluster, and Region. POD represents the smallest unit that can be created and deployed. The subordinate relationship of these categories can be Block∈(POD)∈AZ∈Cluster∈Region.

[0102] The attributes of each category in the multi-layer target category are included in the attributes of each category in the multi-layer category. The target entity associated data of the target entity data are included in the entity data under each category and the associated data between the entity data under each category.

[0103] Exemplarily, the attributes of a category may include at least one of various features, parameters, and descriptive information, and the attributes of a category may be used to define and distinguish different categories. For example, in the field of cloud computing, there may be computing resource categories, storage resource categories, etc. The computing resource category may have attributes such as computing speed and supported parallel computing scale, and the storage resource category may have attributes such as storage capacity and storage type (such as solid state drive storage or mechanical hard drive storage).

[0104] The entity association data includes: other entity data associated with the target entity data, and / or entity association data between the target entity data and other entity data. Exemplarily, the target entity data may include a cloud computing task, a user account, or an application, etc. For example, the target entity data is a cloud computing task for image processing, and the target entity association data may include: at least one cloud computing image and / or an association relationship between the cloud computing task and at least one cloud computing image. For another example, the target entity data includes server A, and the target entity association data may include: other servers associated with server A and / or an association relationship between server A and other associated servers.

[0105] In some embodiments, the computer device may store a pre-built cloud computing knowledge graph, and the computer device may obtain the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data from the stored cloud computing knowledge graph.

[0106] In other embodiments, a pre-built cloud computing knowledge graph may be stored in a cloud server, and a computer device may send target entity data of information to be resolved to the cloud server, so that the cloud server determines the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data based on the cloud computing knowledge graph, and sends the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data to the computer device, so that the computer device can also obtain the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data from the pre-built cloud computing knowledge graph.

[0107] S103: Generate cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity association data.

[0108] In some embodiments, the attributes and entity association data that match the information to be resolved can be determined from the attributes of each category in the multi-layer target category and the target entity association data, and cloud computing solution information that matches the information to be resolved can be generated based on the matching attributes and entity association data.

[0109] In other embodiments, semantic analysis can be performed on the information to be resolved to determine a reply template for cloud computing solution information that matches the information to be resolved, the reply template including at least one area to be filled in, and based on the attributes of the at least one area to be filled in, at least one field that can be filled in at least one area to be filled in is determined from the attributes of each category in the multi-layer target category and the target entity-related data, and the at least one field is filled in at least one area to be filled in to obtain cloud computing solution information that matches the information to be resolved.

[0110] In some other embodiments, a reply template of cloud computing solution information matching the information to be resolved can be determined based on the attributes of each category in the multi-layer target category, the reply template includes at least one area to be filled in, and based on the attributes of the at least one area to be filled in, at least one field that can be filled in the at least one area to be filled in is determined from the attributes of each category in the multi-layer target category and the target entity associated data, and the at least one field is filled in the at least one area to be filled in, to obtain the cloud computing solution information matching the information to be resolved. Exemplarily, the attributes of each category in the multi-layer target category may include a link to the reply template. Also exemplarily, the attributes of each category in the multi-layer target category may include a reply result, and the entity data, the attributes of the entity, and the associated data of the entity in the reply result are removed to obtain the reply template.

[0111] For example, if the information to be solved is "What certificates are needed for hybrid cloud solution project A", the cloud computing solution solution information matching the information to be solved may include "at least one certificate name and / or at least one certificate image corresponding to the name of at least one certificate". For another example, if the information to be solved is "What is the server architecture of hybrid cloud solution project C", the cloud computing solution solution information matching the information to be solved may include "the topology diagram of the server architecture of hybrid cloud solution project C", etc.

[0112] S104: Output cloud computing solution information.

[0113] In some embodiments, the computer device may display the cloud computing solution information in a solution interface displayed in the computer device. In other embodiments, the computer device may output the cloud computing solution information to the electronic device so that the electronic device displays the cloud computing solution information in the solution interface.

[0114] In the technical solution of the embodiment of the present application, the computing device generates cloud computing solution information matching the information to be solved based on the acquired information to be solved and the pre-built cloud computing knowledge graph, and outputs the cloud computing solution information. In this way, the cloud computing solution information matching the information to be solved can be output by inputting the information to be solved once. Compared with the method in the related art that the computer device needs to obtain and browse a large amount of cloud computing knowledge data multiple times and through multiple channels and analyze the cloud computing solution information before it can obtain the cloud computing solution information input by the user, this can greatly shorten the time for obtaining the cloud computing solution information, thereby improving the efficiency of obtaining the cloud computing solution information.

[0115] Some implementation scenarios are described below: In some implementation scenarios, the solution interface includes a cloud computing knowledge query interface, the information to be solved includes cloud computing file description information, and the cloud computing solution solution information includes link information of the target cloud computing file. In other implementation scenarios, the solution interface includes a cloud computing dialogue interface, the information to be solved includes cloud computing question information, and the cloud computing solution solution information includes cloud computing answer information. In still other implementation scenarios, the solution interface includes a cloud configuration solution generation interface, the information to be solved includes cloud configuration solution requirement information; and the cloud computing solution solution information includes cloud configuration solution results.

[0116] Figure 2 A flow chart of a method for supplementing knowledge data provided in some embodiments, the method is applied to a computer device, and the method for supplementing knowledge data can be Figure 1 Executed before or may be performed in Figure 1 After the implementation of the embodiment, the knowledge data supplement method can be executed once every preset time, for example, the preset time can be 1 day, 2 days, a week or a month, etc., so that the impact of the update of the cloud computing knowledge graph on the output of the cloud computing solution information can be reduced. The knowledge data supplement method includes:

[0117] S201 . In response to a designated instruction obtained by inputting supplementary knowledge data in a knowledge supplement interface of a cloud computing solution system, sending supplementary knowledge data to a target device.

[0118] Exemplarily, the supplementary knowledge data may be structured data, semi-structured data or unstructured data. Unstructured data may include unstructured text data or unstructured multimedia data.

[0119] In some embodiments, the computer device may be installed with a cloud computing solution system and be able to display a knowledge supplement interface of the cloud computing solution system. The user may input supplementary knowledge data in the knowledge supplement interface, and the computer device sends the supplementary knowledge data to the target device in response to the user inputting the supplementary knowledge data.

[0120] In other embodiments, the computer device may receive a designated instruction sent by the electronic device, and the designated instruction carries the supplementary knowledge data. For example, the electronic device may be installed with a cloud computing solution system and be able to display a knowledge supplement interface of the cloud computing solution system. The user may input the supplementary knowledge data in the knowledge supplement interface. The electronic device generates a designated instruction in response to the user inputting the supplementary knowledge data and sends the designated instruction to the computer device.

[0121] Exemplarily, the target device may be a device at the cloud computing knowledge review end or a device at the cloud computing expert end. Exemplarily, the target device may be any device with a display function, for example, the target device may be a mobile phone or a computer.

[0122] In some embodiments, the target device can display supplementary knowledge data. When the personnel on the target device side determines that the supplementary knowledge data is accurate, they send confirmation information to the computer device; when the personnel on the target device side determines that the supplementary knowledge data is inaccurate, they send negative information and negative reasons to the computer device.

[0123] In some implementation scenarios, the computer device may display the negation information and the negation reason in the knowledge supplement interface displayed by the computer device. In other implementation scenarios, upon receiving the negation information and the negation reason, the computer device sends the negation information and the negation reason to the electronic device, so that the electronic device displays the negation information and the negation reason in the knowledge supplement interface.

[0124] S202: In response to confirmation information on the supplementary knowledge data sent by the target device, update the cloud computing knowledge graph according to the supplementary knowledge data.

[0125] Among them, the confirmation information is used to indicate that the supplementary knowledge data is accurate.

[0126] In some embodiments, when the supplementary knowledge data is structured data or semi-structured data, the supplementary knowledge data is structurally transformed to obtain a transformed knowledge graph, and the cloud computing knowledge graph is updated based on the transformed knowledge graph.

[0127] In some embodiments, when the supplementary knowledge data is unstructured text data, the following scheme is used to update the cloud computing knowledge graph: category attribute recognition is performed on the text data corresponding to the supplementary knowledge data to obtain attributes of each category in the multi-layer specified category, and the designated entity data and designated entity associated data corresponding to the attributes of each category in the multi-layer specified category are extracted from the text data corresponding to the supplementary knowledge data; the cloud computing knowledge graph is updated according to the attributes of each category in the multi-layer specified category, the designated entity data and the designated entity associated data.

[0128] In the case where the supplementary knowledge data is unstructured multimedia data, the supplementary knowledge data is sent to a target device, and text data corresponding to the supplementary knowledge data sent by the target device (i.e., text description data of the supplementary knowledge data) is received, and the cloud computing knowledge graph is updated using the following scheme: attributes of each category in the multi-layer specified category are obtained by performing category attribute recognition on the text data corresponding to the supplementary knowledge data, and designated entity data and designated entity association data corresponding to the attributes of each category in the multi-layer specified category are extracted from the text data corresponding to the supplementary knowledge data; the cloud computing knowledge graph is updated according to the attributes of each category in the multi-layer specified category, the designated entity data, and the designated entity association data.

[0129] S203: Output information indicating that the supplementary knowledge data is accurate.

[0130] In some embodiments, the computer device may display the indication information that the supplementary knowledge data is accurate on the knowledge supplement interface. In other embodiments, the computer device may output the indication information that the supplementary knowledge data is accurate to the electronic device, so that the electronic device displays the indication information that the supplementary knowledge data is accurate on the knowledge supplement interface.

[0131] Figure 3 A flowchart of a method for constructing a cloud computing knowledge graph provided in some embodiments, the method is applied to a computer device, the method is executed before a target instruction is obtained in response to an operation of inputting information to be solved in a solution interface of a cloud computing solution system, and the method includes:

[0132] S301: Preprocess the acquired cloud computing knowledge data to determine cloud computing text data.

[0133] In some embodiments, unstructured data in cloud computing knowledge data is identified, the identified unstructured data is preprocessed, and cloud computing text data is determined.

[0134] Unstructured data includes unstructured multimedia data and unstructured text data, and unstructured multimedia data includes image data, voice data, video data, etc. The following describes how to preprocess unstructured text data, image data, voice data, and video data to obtain multiple sub-cloud computing text data, and a combination of at least part of the multiple sub-cloud computing text data is the cloud computing text data.

[0135] Exemplarily, the unstructured text data in the cloud computing knowledge data is determined as the first sub-cloud computing text data corresponding to the unstructured text data.

[0136] As another example, optical character recognition (OCR) is performed on the image data to obtain second sub-cloud computing text data corresponding to the image data.

[0137] Also illustratively, voice recognition is performed on the voice data to obtain third sub-cloud computing text data corresponding to the voice data.

[0138] As another example, OCR recognition and speech recognition are performed on each frame of data in the video data to obtain fourth sub-cloud computing text data corresponding to the video data.

[0139] As another example, unstructured multimedia data may be sent to a target device, and fifth sub-cloud computing text data corresponding to the unstructured multimedia data sent by the target device may be received.

[0140] In this way, the computer device can determine the corresponding cloud computing text data according to the type of the unstructured text data. For example, if the unstructured text data includes unstructured text data, image data, voice data, and video data, then the corresponding cloud computing text data includes the first to fifth sub-cloud computing text data, or includes the first to fourth sub-cloud computing text data, or includes the first, second, and fifth sub-cloud computing text data, or includes the first and fifth sub-cloud computing text data. For another example, if the unstructured text data includes unstructured text data and image data, then the corresponding cloud computing text data includes the first, second, and fifth sub-cloud computing text data, or the corresponding cloud computing text data includes the first and second sub-cloud computing text data, or the corresponding cloud computing text data includes the first and fifth sub-cloud computing text data.

[0141] S302. Construct a model layer of a cloud computing knowledge graph based on the attributes of each category in the multi-layer categories obtained by performing category attribute recognition on the cloud computing text data.

[0142] The pattern layer represents the hierarchical association relationship of the attributes of each category in the multi-layer category. In some embodiments, the pattern layer can also be called the concept layer.

[0143] Among them, the multi-layer target category is a part of the categories in the multi-layer category.

[0144] The attributes of a category refer to the common characteristics of the category. The pattern layer of the cloud computing knowledge graph can include the hierarchical associations of categories in the multi-layer categories and the attributes of each category.

[0145] The model layer is built on the data layer and is the core of the knowledge graph. The ontology library is usually used to manage the model layer of the knowledge graph. Ontology is a conceptual template for a structured knowledge base. The knowledge base formed by the ontology library not only has a strong hierarchical structure, but also has a low degree of redundancy.

[0146] S303. Construct a cloud computing knowledge graph based on the model layer and cloud computing text data.

[0147] In some embodiments, entity recognition and entity relationship recognition can be performed on cloud computing text data based on the pattern layer to obtain a data layer that matches the pattern layer. The data layer that matches the pattern layer includes entity data and entity-related data that match the pattern layer. Based on the pattern layer and the data layer, a cloud computing knowledge graph is constructed.

[0148] Figure 4 A flowchart of a method for determining cloud computing text data provided in some embodiments is provided. The method is applied to a computer device. The method is a description of S301. The method includes:

[0149] S401: Identify text data and multimedia data in cloud computing knowledge data.

[0150] In some embodiments, unstructured data in cloud computing knowledge data is identified, and text data and multimedia data in the unstructured data are acquired.

[0151] S402: Determine text data in the cloud computing knowledge data as first text data, and send multimedia data to a target device.

[0152] In some embodiments, multimedia data is sent to a target device so that the target device outputs the multimedia data. The target device receives text description information of the multimedia data input by a user, obtains second text data of the multimedia data, and sends the second text data to a computer device.

[0153] S403, receiving second text data of the multimedia data sent by the target device; the second text data is text description information of the multimedia data.

[0154] The first text data and the second text data are included in the cloud computing text data.

[0155] Figure 5 A schematic flow chart of a method for constructing a pattern layer of a knowledge graph provided in some embodiments, the method being applied to a computer device, the method comprising:

[0156] S501. Acquire the architecture information of the pre-built pattern layer; the architecture information represents the hierarchical association relationship.

[0157] Exemplarily, the architectural information of the pattern layer is similar to a mind map in the field of cloud computing, and the architectural information of the pattern layer may include subclasses under each knowledge classification in at least one knowledge classification. Exemplarily, the subclasses under a knowledge classification may also have a hierarchical association relationship. Another exemplary embodiment, the subclasses under a knowledge classification are all categories of the next layer of the knowledge classification. For example, at least one knowledge classification may include at least one of development, application, infrastructure, security, platform, operation and maintenance, etc. For example, the subclasses under development may include subclasses of processes and / or languages. Exemplarily, the next layer of the language subclass may have subclasses of at least one of Java, C, Python, etc.

[0158] S502: Perform category attribute recognition on cloud computing text data to obtain attributes of each category in the multi-layer category.

[0159] In some embodiments, each category in the multi-layer category can be extracted from cloud computing text data without the guidance of architecture information. In other embodiments, the cloud computing text data can be guided by architecture information to perform category attribute recognition to obtain the attributes of each category in the multi-layer category.

[0160] Exemplarily, each category in the multi-layer category may include each category in the architecture information, and new categories under the guidance of each category in the architecture information, and the new categories are supplementary categories to each category in the architecture information. Also exemplarily, each category in the multi-layer category may be each category in the architecture information.

[0161] S503: Map the attributes of each category in the multi-layer category to the architecture information to obtain a pattern layer.

[0162] In this way, the pattern layer includes not only the structure between categories in the multi-layer categories, but also the attributes of each category in the multi-layer categories.

[0163] In some embodiments, the attributes of each category in the multi-layer category are mapped to the architecture information to obtain the pattern layer, including: performing category concept recognition on cloud computing text data to obtain the concepts of each category in the multi-layer category; analyzing the concepts of each category in the multi-layer category to obtain the hierarchical association relationship of the categories in the multi-layer category; and mapping the attributes of each category in the multi-layer category to the architecture information according to the hierarchical association relationship of the categories in the multi-layer category to obtain the pattern layer.

[0164] In some implementations, a concept may include a term and explanatory information of the term.

[0165] In some implementations, the construction of the model layer may also be referred to as the construction of the ontology.

[0166] In some embodiments, category attribute recognition is performed on cloud computing text data to obtain attributes of each category in the multi-layer category, including: using large language models (LLMs) to perform category attribute recognition on part of the cloud computing text data to obtain attributes of each category in the multi-layer category of the part of the data; in response to modification prompt information for the attributes of each category in the multi-layer category of the part of the data, modifying the attributes of each category in the multi-layer category of the part of the data; using the attributes of each category in the multi-layer category of the part of the data and the modified part of the data to train the large language model; using the trained large language model to perform category attribute recognition on another part of the cloud computing text data to obtain attributes of each category in the multi-layer category of the other part of the data.

[0167] Exemplarily, the large language model may include models such as ChatGPT, Wenxinyiyan, or Copilot.

[0168] Using a large language model to perform category attribute recognition on part of the data in cloud computing text data can include: providing part of the data in the cloud computing text data to the large language model, and inputting instruction information for performing category attribute recognition to the large language model, so that the large language model outputs attributes of each category in the multi-layer category of the part of the data.

[0169] In some embodiments, a computer device may display the attributes of each category in a multi-layer category of partial data, or send the attributes of each category in a multi-layer category of partial data to an electronic device, so that the electronic device displays the attributes of each category in the multi-layer category of partial data. A user may input modification prompt information of the attributes of each category in the multi-layer category of partial data to the computer device or the electronic device through the displayed attributes of each category in the multi-layer category of partial data, and the electronic device sends the modification prompt information to the computer device. In this way, when the computer device obtains the modification prompt information, it may input the modification prompt information to the large language model, so that the large language model re-outputs the attributes of each category in the multi-layer category of partial data.

[0170] By adopting the attributes of each category in the multi-layer categories of partial data and modified partial data, the large language model is trained, so that the performance of the large language model in category attribute recognition is getting better and better.

[0171] In some embodiments, the attributes of each category in the multi-layer category of the modified portion of data and the attributes of each category in the multi-layer category of another portion of data are determined as the attributes of each category in the multi-layer category.

[0172] Figure 6A schematic flow chart of a method for constructing a cloud computing knowledge graph provided in some other embodiments, the method is applied to a computer device, the method is a description of S303, and the method includes:

[0173] S601. Extracting entity data and entity-related data corresponding to attributes of each category in the multi-layer categories in the pattern layer from the cloud computing text data.

[0174] In some embodiments, entity data under each category in the multi-layer category can be extracted from the cloud computing text data. Exemplarily, one entity data can be included in at least one category.

[0175] Through S601, all entity data in the cloud computing text data that matches the attributes of each category in the multi-layer category and the associated data between all entity data can be obtained.

[0176] S602: Construct a data layer of a cloud computing knowledge graph based on entity data and entity-related data.

[0177] The data layer of the knowledge graph is the basic layer of the knowledge graph, which is responsible for storing and managing all the raw data in the knowledge graph. The data layer is mainly composed of a series of facts, which exist in a structured form and are usually stored in the graph database in the form of "entity-relationship-entity" triples or "entity-attribute-attribute value" attribute graphs. Exemplarily, entity association data may include association relationships between entities and / or association relationships between entities and entity attributes.

[0178] S603. Construct a cloud computing knowledge graph based on the model layer and the data layer.

[0179] In this way, the constructed model layer includes the hierarchical association relationship of the attributes of each category in the multi-layer categories, and the constructed data layer includes the entity data and entity association data corresponding to the model layer. Therefore, based on the model layer and the data layer, a cloud computing knowledge graph is constructed, which can structurally integrate knowledge from all aspects of the cloud computing field, and can more effectively organize and display the logical relationship between knowledge, which is conducive to knowledge query and retrieval, thereby helping to improve the efficiency of knowledge query and retrieval.

[0180] In some embodiments, a data layer of a cloud computing knowledge graph is constructed based on entity data and entity-related data, including: performing disambiguation processing on entity data to obtain target data; determining target association relationships based on target data and entity-related data; and performing knowledge reasoning on target data and target association relationships to obtain a data layer of a cloud computing knowledge graph.

[0181] Exemplarily, performing disambiguation processing on entity data may include: performing entity disambiguation processing on the entity data and / or performing coreference disambiguation processing on the entity data.

[0182] The essence of entity disambiguation is that a word has many possible meanings. Entity disambiguation is to map a word with different meanings to different identifiers and / or descriptions. For example, the meaning of "apple" in the two sentences "My phone is an apple" and "I like to eat apples" is different. The former represents a mobile phone, and the latter represents fruit.

[0183] Co-reference disambiguation is also called reference resolution. Because natural language is full of ambiguity, multiple signals and knowledge must be used to disambiguate. These references need to be understood based on the understanding of the surrounding world. For example, "it is a server node". It and server node both mean server node, so the entity data about it is discarded.

[0184] In some embodiments, the entity association data is adjusted using the target data to determine the target association relationship.

[0185] In some embodiments, an original knowledge graph can be constructed based on the target data and the target association relationship, and knowledge reasoning can be performed on the original knowledge graph to obtain the data layer of the cloud computing knowledge graph. In other embodiments, an original knowledge graph can be constructed based on the pattern layer, target data and target association relationship of the knowledge graph, and knowledge reasoning can be performed on the original knowledge graph to obtain the data layer of the cloud computing knowledge graph.

[0186] Exemplarily, the reasoning of the knowledge graph mainly revolves around the reasoning of relationships, that is, inferring unknown facts or relationships based on existing facts or relationships in the knowledge graph, and focusing on the characteristic information of entities, relationships and graph structures. Knowledge graph reasoning can mainly assist in reasoning out new facts, new relationships, new axioms and new rules, etc. Exemplarily, the reasoning tasks mainly include at least one of the following: completing the knowledge graph through rule mining (Knowledge Base Completion, KBC) and quality verification, link prediction, association relationship reasoning and conflict detection, etc. Exemplarily, the main methods of knowledge reasoning include one of the following: reasoning based on logical rules, reasoning based on graph structure, reasoning based on distributed representation learning, reasoning based on neural networks, and hybrid reasoning.

[0187] Exemplarily, when the data layer of the cloud computing knowledge graph is obtained, the data layer of the cloud computing knowledge graph can also be updated according to the new cloud computing knowledge data, so that the knowledge of the data layer of the cloud computing knowledge graph will become more and more abundant and accurate.

[0188] Figure 7 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some embodiments, the method is applied to a computer device, and the method includes:

[0189] S701: Identify structured data and semi-structured data in cloud computing knowledge data, and pre-process the acquired cloud computing knowledge data to determine cloud computing text data.

[0190] In some embodiments, structured data and semi-structured data in the cloud computing knowledge data may be identified based on the format information of the cloud computing knowledge data.

[0191] S702: Perform structural transformation on structured data and semi-structured data to obtain a first knowledge graph.

[0192] Exemplarily, the structured data and semi-structured data can be structurally transformed by a transformation rule mapping tool to obtain a first knowledge graph. Exemplarily, the transformation rule mapping tool can include at least one of the following: database to resource description framework D2R (Database to RDF), relational database to resource description framework mapping language (RDB to RDFmapping language, R2RML), graph mapping, etc. RDB stands for relational database (Relational Database), and RDF stands for resource description framework (Resource Description Framework).

[0193] S703. Construct a model layer of the cloud computing knowledge graph based on the attributes of each category in the multi-layer categories obtained by performing category attribute recognition on the cloud computing text data.

[0194] The description of S703 may refer to the description of S302 above.

[0195] S704: Construct a second knowledge graph based on the model layer and cloud computing text data.

[0196] In some embodiments, entity recognition and entity relationship recognition can be performed on cloud computing text data based on the pattern layer to obtain a data layer that matches the pattern layer. The data layer that matches the pattern layer includes entity data and entity-related data that match the pattern layer. Based on the pattern layer and the data layer, a second knowledge graph is constructed.

[0197] S705: Merge the first knowledge graph and the second knowledge graph to obtain a cloud computing knowledge graph.

[0198] Exemplarily, knowledge fusion is achieved by fusing similar or identical subjects through one or several attributes of the subject and the relationship between the subjects through a fusion algorithm to achieve the purpose of reference disambiguation.

[0199] In some embodiments, the knowledge graph is logically divided into two layers, namely the pattern layer and the data layer. The former is a pattern that describes the data and is a refinement and constraint on the data layer, while the latter is a layer that stores specific data information, also known as the instance layer. In some embodiments, there are only two major ways to construct the graph, bottom-up and top-down. The top-down construction is to first determine the pattern layer and then add entity data to the knowledge base, while the bottom-up construction is to first determine the data layer of the knowledge graph and then refine the data pattern. Different from the bottom-up construction method mostly used in general domain knowledge graphs, the general industry domain knowledge graph is small in scale and relatively closed in domain. It is easy to construct the pattern layer according to the characteristics of the domain, and it is mostly constructed in a top-down manner.

[0200] The model layer of the knowledge graph expresses a "meta-concept", that is, the schema corresponding to the transaction entity type, which is a structured definition of knowledge with the same data structure, defining the attributes, relationships and related constraints of the knowledge concept, and mainly realizing the structuring of entities and the definition of relationships between entities. The industry generally uses the philosophical concept of ontology to standardize the design pattern of the graph and reduce the subjective bias of knowledge cognition. Ontology is the basic terms and relationships that constitute the vocabulary of related fields, as well as the rules that define the extension of these terms and relationships.

[0201] Figure 8 A schematic diagram of a process for constructing a cloud computing knowledge graph provided in some further embodiments, the method is applied to a computer device, and the method includes:

[0202] Analyze the main data sources and data characteristics of the field scenario, adopt the data acquisition mode combining the two sub-modules of "active acquisition" and "passive input", and combine interactive annotation metadata and other methods to analyze the multi-source and multi-modal data (including Figure 8 At least one of the data in the database, white papers, hybrid cloud construction solutions, industry standards and specifications, data released by media platforms, etc. in the database is preprocessed to provide a data basis for building a knowledge graph. Exemplarily, by analyzing multi-source multimodal data, structured data, semi-structured data and unstructured data are obtained. The data acquisition module (including active acquisition module and / or passive acquisition module) can obtain these structured data, semi-structured data and unstructured data (also multivariate multimodal original data), preprocess the multivariate multimodal data to obtain structured data, semi-structured data and unstructured text data, and transmit the unstructured text data (i.e., cloud computing text data in the unstructured data) to the concept layer construction module so that the concept layer construction module can construct the concept layer.

[0203] Secondly, based on the cloud computing text data, the concept layer is constructed to construct the concept layer of the cloud computing knowledge graph (also called the ontology or pattern layer of the cloud computing knowledge graph). The concept layer of the cloud computing knowledge graph includes the following steps: determining the scope, considering ontology reuse, refining terms, establishing relationships, defining attributes, performing attribute restrictions (also called attribute facets), and constructing the ontology. In determining the scope, the business departments and business goals mainly used in the hybrid cloud solution design domain scenario, as well as the functional and performance requirements, can be designed in advance based on the relevant experience of the business department, construction standards and specifications, etc., prompt templates that meet specific fields are pre-designed. The prompt template is the architecture information of the pattern layer. In considering ontology reuse, the ontology of cloud computing related fields can be obtained, and the ontology of cloud computing related fields can be integrated with the ontology determined by the following steps to obtain the concept layer of the cloud computing knowledge graph: refining terms, establishing relationships, defining attributes, and performing attribute restrictions. By refining terms, establishing relationships, defining attributes, and performing attribute restrictions, the concept of category terms can be extracted from cloud computing text data, the hierarchical association between categories in multi-layer categories can be determined based on the concept of category terms, the attributes of each category in multi-layer categories can be extracted from cloud computing text data, and attribute restrictions can be performed to obtain at least one facet of the attributes of each category in each multi-layer category, and the attributes of each category in the multi-layer category can be included in at least one facet, and the ontology can be determined based on the attributes of each category in the multi-layer category. In the process of extracting cloud computing text data, the powerful semantic understanding and context perception capabilities of ChatGPT can be used to refine entity recognition and classification of pre-processed data, and through reasonable questions and feedback, ChatGPT can be guided to provide more refined ontology extraction results, thereby constructing the pattern layer of the knowledge graph. The pattern layer of the knowledge graph can be mapped to the knowledge graph.

[0204] Next, when the concept layer is constructed, knowledge extraction, entity alignment, and knowledge representation are performed according to the concept layer of the knowledge graph. For example, relationship extraction and attribute extraction are performed on cloud computing text data, entity data and entity-related data are determined, and entity data is disambiguated (including entity disambiguation and coreference disambiguation) to obtain target data (i.e., triples and attribute graphs). The target data can be the data layer of the knowledge graph. The original knowledge graph is constructed according to the concept layer and data layer of the knowledge graph. The computer device can perform the following operations on the original knowledge graph: knowledge reasoning, knowledge updating, and knowledge storage. Through knowledge storage, the cloud computing knowledge graph can be stored. The cloud computing knowledge graph can be applied to the knowledge-driven hybrid cloud solution design system to realize the knowledge graph application. In this way, through the links of knowledge extraction, knowledge fusion, knowledge reasoning, and knowledge updating, the knowledge graph is constructed under the mapping of the pattern layer. Figure 8 The formal expression of the knowledge graph is shown.

[0205] In some implementations, a knowledge-driven hybrid cloud solution design system can be designed around the core demands of the daily work of hybrid cloud solution designers to realize the application of knowledge graphs. The functional points include at least one of the following: 1) real-time status query and display of the knowledge base; 2) semantic understanding of questions and answers and traceability of answer support materials; 3) hybrid cloud configuration solution recommendations; 4) knowledge and experience sharing and expert maintenance.

[0206] The following is an explanation of obtaining multi-source multi-modal data and preprocessing it: The analysis and processing of key business or scenario data directly determines the quality of knowledge graph construction. This step aims to uniformly preprocess multi-source multi-modal data to facilitate better model-to-model matching later. Specifically, the data involved in the design of hybrid cloud solutions comes from a wide range of sources and has different modalities, such as: construction guidance plans for hybrid cloud construction, server configuration manuals, various product capability lists, various software copyright certificates or mutual authentication certificate images, various work environment instant messaging software platforms or communication information between work mailboxes, various back-end databases, etc. Therefore, this step involves a unified adaptive receiving module for multi-source multi-modal data, including two sub-modules: active acquisition and passive input. Through the cooperation of two different sub-modules, the cleaning and preprocessing of multi-source multi-modal data can be completed. For unstructured data (multimedia data such as pictures, videos, and audio), an interactive additional metadata description module will be designed in this link and integrated into the overall system. The unstructured data stream will be imported into this module through diversion triggering, and the textual processing of unstructured multimedia data will be realized through the method of additional annotation metadata (metadata is the text description information of multimedia data) through manual interaction.

[0207] Fig. 9 A schematic diagram of a process for obtaining multi-source multi-modal data and preprocessing the data provided in some embodiments, such as Fig. 9 As shown, the data acquisition module (e.g., a unified adaptive receiving module for multivariate multimodal data) can acquire multi-source multimodal data, and the multi-source multimodal data can include at least one of the following: a word (Word) file, a table (e.g., Excel) file, a text (TEXT) file, an email file, a portable document format (PDF) file, a chat file, etc. The multi-source multimodal data can be actively acquired or passively input. The computer device can acquire unstructured multimedia data in the multi-source multimodal data, and obtain cloud computing text data of the multimedia data by interactively attaching metadata descriptions, so that the multi-source multimodal data can be classified as structured data, semi-structured data, or unstructured text data.

[0208] After the above process, multi-source multi-modal data can be uniformly processed into structured / semi-structured or unstructured free text data. Generally speaking, for structured / semi-structured data, because they already have structured semantic features, they are essentially highly condensed knowledge expressions. With the help of conversion rule mapping tools such as D2R and R2RML, the relational data structure can be converted into a graph data structure in order to build a knowledge graph. Based on the unified processing of unstructured free text data completed in the above steps, the next step is to carry out subsequent knowledge extraction, knowledge fusion, knowledge processing, and knowledge updating for unstructured free text data to complete the construction of the knowledge graph.

[0209] The following is an explanation of how to implement the knowledge graph model layer (i.e., the model layer of the cloud computing knowledge graph, also called ontology) based on domain characteristics: Due to differences in domains and industries, the model layer construction of the knowledge graph is also different. This step requires determining the main business departments and business goals, as well as functional and performance requirements, and combining the relevant experience of the business departments, construction standards and specifications, to carry out the design of the domain ontology.

[0210] Fig.10 A flow chart of domain scenario ontology construction provided for some embodiments, wherein the domain scenario may include a domain scenario for hybrid cloud solution design, and the ontology construction may include the following steps: determining the scope, ontology reuse, refining terms, establishing relationships, defining attributes, attribute restrictions, and constructing the ontology.

[0211] Determine the scope of the domain involved in the ontology. This step extensively refers to the hybrid cloud construction plan, product white papers of various production lines, product capability lists, relevant industry standards and literature, and the Cloud Computing Open Source Industry Alliance Hybrid Cloud White Paper. At the same time, combined with the knowledge of domain experts, in-depth analysis of various elements and features involved in the hybrid cloud solution design scenario, and design of the ontology architecture.

[0212] Examine and reuse existing ontologies. For example, there is currently no definition or construction method for knowledge graphs related to hybrid cloud solution design in the cloud computing industry. Therefore, in this step, the ontologies in weakly related fields can be used for reference only. Exemplarily, the ontologies in weakly related fields in the hybrid cloud solution field may include: ontologies in the solution design field and / or ontologies in the cloud solution field, etc.

[0213] Identify the important concepts or terms in this field. In this step, based on the field defined in the first step, list the important concepts and terms in the field of hybrid cloud solution design, such as Region, AZ, Cluster, POD, VPC, LB, etc.

[0214] Combine concepts or terms to analyze the hierarchical relationship and attribute relationship between classes. This step mainly defines each class according to the importance of the concept or class, and then establishes the association relationship and hierarchical hierarchy between concepts or classes to achieve the hierarchical division of ontology concepts. For example: for Region, AZ, Cluster, POD, Block, there is a mutual inclusion relationship: Block∈(POD)∈AZ∈Cluster∈Region.

[0215] Describe the attributes of the class based on domain knowledge. This step aims to assign corresponding attributes to the defined class, which generally include internal attributes, external attributes, and object attributes. For example, the attributes of "server" can include: "model", "manufacturer", "price", "type", etc.

[0216] Attribute restrictions can include defining facets of attribute values. Facets of attributes refer to limiting the validity by defining other characteristics of attribute values, including the type, range, quantity, etc. of attribute values. Among them, dividing attributes into multiple facets usually means in software development, dividing the attributes or characteristics of an object into multiple parts, each part is called a "facet", and the attributes in each facet are encoded separately according to the rules.

[0217] In building the ontology. Based on the above steps, build classes, subclasses, attributes, etc. according to the requirements of ontology construction, and add individuals (i.e. node labels) to the classes to complete the ontology construction.

[0218] Pre-design prompt templates that are in line with the specific fields of hybrid cloud solution design, and then use the powerful semantic understanding and context awareness capabilities of large language models such as ChatGPT to refine entity recognition and classification of pre-processed data, and fill them into the pre-designed templates accordingly. By asking reasonable questions and providing feedback to large language models such as ChatGPT, the model parameters can be gradually corrected to guide the model to provide more refined ontology extraction results.

[0219] The following is an explanation of the implementation method of building a domain knowledge graph (i.e., cloud computing knowledge graph): This step is the main part of building a knowledge graph, including knowledge extraction, knowledge fusion, knowledge reasoning, and knowledge updating. Knowledge extraction includes the extraction of entities and relationships. Among them, entity extraction for unstructured free text is to detect named entities from the text and classify them into predefined categories, such as: people, organizations, groups, equipment, time, place, etc. In general, the existing entity extraction methods can be divided into rule-based methods (such as rules, dictionaries, etc.), statistical model-based methods, and deep learning-based methods. The statistical model-based method can use conditional random field (CRF), hidden Markov model (HMM) or maximum entropy Markov model (MEMM). The deep learning-based method can use bidirectional long short-term memory network (Bidirectional Long Short-Term Memory, Bi-LSTM) + CRF or LSTM + Convolutional Neural Networks (CNN) + CRF, etc. Relation extraction refers to extracting the relationship between entities from text. Generally, after identifying the entities, possible relationships between entities are extracted. At present, relationship extraction methods can be divided into template-based methods, supervised learning-based methods, and weakly supervised learning-based methods. Among them, the supervised learning-based methods can use contrastively-reinforced convolutional neural networks (CR-CNN) and extract transform load-span (ETL-Span), and the weakly supervised learning-based methods can use pulse coupled neural networks (PCNNs). Constructing a domain knowledge graph is actually to perform modulus matching on the corresponding knowledge graph pattern layer. Under the guidance of the pattern layer, multi-source and multi-modal knowledge is based on entity relationship extraction and other methods to obtain the basic triple units that make up the knowledge graph, and construct the data layer of the domain knowledge graph of the hybrid cloud solution.

[0220] In the process of knowledge extraction, due to the heterogeneity of data from multiple sources, there may be a variety of expressions when explaining the same concept. The extracted entities, attributes, relationships and other reference items may have information redundancy or ambiguity. Therefore, it is very necessary to perform semantic disambiguation and knowledge fusion in the process of knowledge graph construction. In the process of knowledge fusion, it is mainly through calculating the similarity between entities or relationships to determine whether there is a co-reference between the reference items to be matched. The similarity of entities can be understood as describing the differences in their own metadata information in multiple dimensions. The n-item features of the metadata can be formed into an n-dimensional feature vector. For example, entity a is expressed as an n-dimensional feature vector V. a =(x1,x2,...,x n ). Similarly, entity b is expressed as an n-dimensional feature vector V b =(y1,y2,...,y n ), then the semantic similarity between the two can be expressed by the cosine value of the angle between their vectors, that is, For entities or relationships expressed in the entire sentence text, a supervised sentence embedding model can be used, such as generating a fixed-length vector representation of the sentence (Sentence Bidirectional Encoder Representations from Transformers, Sentence-BERT), and representing the acquired text with sentence vector features. The vectors i and j of the two sentences still use the vector angle cosine similarity formula to calculate the similarity, that is, A threshold α is pre-set for the cosine of the above vector angles. When the calculated result is greater than α, there is reason to believe that the two vectors are highly semantically similar and entity fusion alignment can be performed.

[0221] Reasoning and updating of knowledge are important means to supplement or improve existing knowledge graphs. The former is based on the existing knowledge base and, with the support of reasonable association rules, further enriches and improves the content that is not explicitly expressed in the original knowledge system. It mainly includes two parts: reasoning about entity relationships and reasoning about entity attributes. The reasoning methods are mainly divided into graph-based reasoning and logic-based reasoning. The latter takes into account the timeliness of knowledge, and based on strategies such as regular updates, event-driven updates, or user feedback updates, the existing knowledge graphs are improved and updated to ensure that the knowledge or solutions output by the system comply with the latest current policies or relevant standards and are authentic and reliable.

[0222] After the above steps, we can obtain hybrid cloud domain solution design domain data with clear structure, diverse entity attributes and association relationships. The storage method of graph database is obviously more in line with the network storage mode of diversified knowledge in this field. Therefore, this article will use the Neo4j database based on the graph model to store the established knowledge graph, and then we can also mine related knowledge based on the graph structure. For example, we can mine whether the manufacturers of several servers are the same manufacturer.

[0223] The following is an explanation of the knowledge-driven hybrid cloud solution design system: Based on the cloud computing knowledge graph obtained above, a knowledge-driven hybrid cloud solution design system is designed to help hybrid cloud solution designers to design solutions efficiently and conveniently.

[0224] The system can focus on the core demands of the daily work of hybrid cloud solution designers and summarize the main functional points in this field: 1) real-time status query and display of the knowledge base; 2) semantic understanding of questions and answers and traceability of answer support materials; 3) hybrid cloud configuration solution recommendations; 4) knowledge and experience sharing and expert maintenance.

[0225] The real-time status query and display function of the knowledge base aims to provide solution personnel with fast and efficient data retrieval. The data sources involved in the hybrid cloud solution design process are extensive and the modalities are very different. For example: construction guidance plans for hybrid cloud construction, server configuration manuals, various product capability lists, various software copyright certificates or mutual certification certificate images, etc. In the daily solution design stage, there may be business scenarios where some qualification materials or capability lists and other materials are directly provided to customers. Currently, different materials are scattered on many platforms, and it is difficult to retrieve the desired materials in an integrated and convenient manner. Therefore, solution designers need a query system that can obtain the required information as comprehensively and quickly as possible. The system needs to integrate multiple platform interfaces, collect various modal data, and establish a knowledge graph for storage, providing solution designers with a simple query interface.

[0226] In semantic understanding of questions and answers and tracing the source of answer support materials, the main role is to use relevant models such as NLP to fully understand the deep knowledge reflected in multimodal data such as documents, tables, and pictures, and extract necessary knowledge points. In the daily solution design process, it greatly saves solution designers the tedious work of searching for relevant knowledge points in massive documents, and for the answers given, the system will simultaneously provide the corresponding traceability materials to clearly indicate the location, so that there is a basis to rely on.

[0227] In the hybrid cloud configuration plan recommendations, on the basis of a comprehensive understanding of hybrid cloud construction plans and other related technical documents, point-to-point analysis of user needs has been added. For a long text description entered into the system, the corresponding field information can be automatically extracted according to the demand template. For missing fields, they can be supplemented through user interaction and other methods, and then the cloud platform configuration plan can be generated by calling the relevant application programming interface (Application Programming Interface, API) or the built-in solution generation platform, gradually reducing the workload of solution designers.

[0228] In the knowledge and experience sharing and expert maintenance, the aim is to further enrich the experience knowledge of the knowledge platform and maintain the authority of the knowledge platform. For solution designers, they not only have the right to obtain knowledge, but also have the right to share knowledge. By resubmitting relevant knowledge points and supporting materials in the system, they can be added to the knowledge platform node after being reviewed and approved by the system knowledge administrator. The expert administrator is responsible for identifying knowledge points that ordinary system administrators cannot determine, and ensuring the validity and accuracy of the knowledge.

[0229] In some embodiments, the system may adopt a browser / server (B / S) architecture, in which the user requests the corresponding service on the browser side and the backend server returns the result. The system may be divided into a data layer, a logic layer, and an interactive display layer according to its functions. Figure 4 This is a brief design diagram of the functional division of the system.

[0230] The above is an explanation of the method in the embodiment of the present application using a computer as an example. In other embodiments, the method in the embodiment of the present application can be applied to a processor. The method of the embodiment of the present application executed by the processor is similar to the method of the embodiment of the present application executed by a computer device, and will not be repeated herein.

[0231] Fig.11 A schematic diagram of system function division provided for some embodiments, such as Fig.11 As shown, the system includes a data layer, a logic layer and an interactive display layer.

[0232] The data layer is mainly used to store various types of system data, including system user data, system multimodal data, knowledge graph data, etc. According to the characteristics of various types of data, the system selects three types of storage databases. Relational databases (such as MySQL) are used to store some structured data and semi-structured data, non-relational databases (such as MongoDB or NoSQL) are used to store some unstructured data, and Neo4j databases are used to store relevant data of knowledge graphs.

[0233] The logic layer connects the data layer and the interaction layer respectively, and mainly involves some logical interfaces of the business or some model algorithms required in the process of building the knowledge graph. Its main function is to process the original data obtained through the bottom layer and build the knowledge graph according to the rules, as well as respond to the operation requests of the platform user business layer (obtained through the business API) (i.e. request response).

[0234] The interactive display layer mainly presents the knowledge graph constructed by the logic layer to the user intuitively and provides a simple query entry to facilitate user operation. The interactive display layer can be used to display at least one of the following interfaces: knowledge base query, semantic understanding question and answer, configuration solution recommendation, and knowledge sharing.

[0235] In the process of acquiring multi-source multi-modal data and preprocessing it, the active acquisition module of the embodiment of the present application mainly obtains chat context information in a timely manner through various work environment instant messaging software platforms or work mailbox-related dialogue interface APIs, and identifies whether the current communication content contains relevant entities through the NLP natural language processing module, and then performs subsequent entity or relationship extraction; or uses scripts to obtain structured data from the background database and uses crawlers and other technical means to obtain semi-structured data on the front-end related web pages for entity or relationship extraction. The passive input module is mainly for data input scenarios that require human intervention, such as various hybrid cloud construction documents or technical specification manuals, software copyright certificates or compatible mutual authentication certificates. For original text documents, document content recognition is directly realized through NLP technology. For image data containing text content, optical character recognition is realized through the built-in OCR module to convert it into text information representing the image content for further processing. For non-text content images and other data, administrators with knowledge system management authority need to modify and process them uniformly through additional metadata. For example, for some typical hybrid cloud project cases, although the on-site photos delivered do not have any text information, they are important for solution engineers to write project summary plans. Therefore, such data materials require administrators to first describe them according to the category and specific content of the data with additional metadata (metadata is data that describes information resources or data objects), and then construct knowledge graphs based on the information contained in the metadata. The disk location information stored in the data itself exists as the basic attribute of the ontology, which is convenient for users to further query the original data. Non-text data materials that appear in chat or email contexts can also be processed in this way of interactive additional metadata description.

[0236] In the process of constructing a domain ontology (i.e., the model layer of the cloud computing knowledge graph) in the embodiment of the present application, according to the characteristics of the hybrid cloud solution design domain scenario, by summarizing the real things and terminology in the field, it can be roughly summarized into three categories of abstract ontologies: personnel, equipment, and methods. By defining these three categories of ontology concept templates, cognitive modeling can be further performed, because most of the scenarios in the field of hybrid cloud solution design can be roughly summarized as "who uses certain equipment to build an environment and provide a complete hybrid cloud solution". Of course, the concepts are all general. For example, personnel can refer to specific solution designers, department leaders, a planning team, or even a department. Equipment refers to servers, switches, security equipment, including software and hardware equipment, which are the basis for building a cloud platform; methods refer to various technical documents, training manuals, some principles for configuring equipment, and other methodological aspects, which are the key links in the organic combination of personnel and equipment. In this stage, after defining a reasonable template, the preprocessed text data and model input are combined, and the encapsulated API provided by open Artificial Intelligence (openAI) is used to specifically perform the task of extracting ontology concepts and relationships from the preprocessed data. By leveraging the powerful semantic perception and understanding capabilities of the large language model and performing real-time interactive model parameter feedback corrections, the refined extraction and construction of the model layer ontology concepts can be achieved.

[0237] In the process of building a domain knowledge graph (i.e., cloud computing knowledge graph), after the data preprocessing and cleaning in the aforementioned steps, the knowledge graph construction in this step is mainly faced with relatively standardized formatted data (such as relational databases and / or manually annotated JSON documents, etc.). For such relatively standardized data, the entities and relationships can be obtained by directly reading the corresponding fields. Mainly in the face of large amounts of text, it is still necessary to use natural language models to deconstruct the various vocabulary and short sentences expressed in the text, and match them one by one according to the entity relationship templates generated in the previous ontology design phase to complete the entity relationship extraction.

[0238] For example, when faced with a text in a hybrid cloud construction solution document, "The cloud management platform divides each functional module into four clusters for deployment. It can be a physical machine or a virtual machine. At present, the main solution is to deploy it on the virtual machine in the management area of ​​the full-stack hybrid cloud. In the case of no full-stack hybrid cloud platform, the cloud management platform can be deployed on the private cloud virtual machine or public cloud host specified by the customer", the corresponding NLP model can interpret it into entity concepts (also called entity data) such as "cloud management platform", "cluster", "physical machine", "virtual machine", "solution", "customer", "private cloud", and "public cloud". Accordingly, for example, "customer" corresponds to the sub-concept under the concept of "personnel" ontology (also called category or type), and "physical machine" corresponds to the sub-concept under the "equipment" type; "solution" is a sub-concept of "method", and correspondingly, the relationship between entities such as "divided into" and "deployment" can also be extracted.

[0239] The extraction of entities and relationships above is the basis for building a knowledge graph. The subsequent knowledge fusion stage involves situations such as "virtual machine" and "virtual machine" which may have differences in description, but actually refer to the same concept. The fusion can be performed by calculating the cosine value of the angle between word vectors and setting a threshold to reduce the redundancy of the knowledge graph. For the constructed knowledge graph, the built-in graph reasoning algorithm and periodic update mechanism can be used to ensure the self-update and improvement of the graph. Finally, the constructed knowledge graph is stored in the Neo4j database designed by the system.

[0240] In the construction of the knowledge-driven hybrid cloud solution design system, an HCSDS system based on the B / S architecture is built. From the bottom to the top, the data layer stores all the basic data required for the system operation, including personnel information, technical documents, knowledge graphs, etc. Through the various algorithm models of the logic layer and the processing and calculation of the API interface, a relatively complete response service can be generated, providing the necessary basis for the functional display of the interactive display layer. In the interactive display layer, users can not only intuitively see the overall picture of the knowledge graph of the current system, but also drag and click any node to understand the entity represented by the node and its relationship with other surrounding entities. In the concise semantic understanding question and answer box, enter the question you want to ask to trigger the knowledge-driven question and answer system. For example, you can enter "Has the ** cloud platform passed the ** trusted cloud certification?", and the system will trigger semantic understanding accordingly, quickly retrieve the key entities "cloud platform" and "trusted cloud certification", and finally give an affirmative answer based on the triple relationship of <cloud platform, trusted cloud certificate, obtained>, and focus the page on the entity node of "** trusted cloud certificate". Clicking on the node will display detailed attribute information, including acquisition time, expiration time, issuing authority, and the corresponding address link for downloading the original file.

[0241] The embodiment of the present application introduces knowledge graph technology to comprehensively integrate various types of multi-source heterogeneous and multi-modal target data, realize imperceptible access to multi-scenario data, and generate a knowledge graph based on this. The design of the hybrid cloud solution breaks the inherent pattern of strong manual reliance on solution design scenarios under normal circumstances, making the entire solution design process accurately controllable, based on evidence, efficient and convenient.

[0242] The embodiments of the present application can make the constructed knowledge graph compatible with various existing platform systems through API docking, and combine with technologies such as NLP to achieve unified query and answer, unified management and maintenance of domain knowledge on each platform, thereby reducing the duplication of knowledge engineering.

[0243] Based on the same inventive concept, the embodiment of the present application also provides a cloud computing solution information generation device for implementing the cloud computing solution information generation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the one or more cloud computing solution information generation device embodiments provided below can refer to the limitations of the cloud computing solution information generation method above, and will not be repeated here.

[0244] In an exemplary embodiment, Fig.12 A schematic diagram of a cloud computing solution information generation device provided in some embodiments of the present application, such as Fig.12 As shown, the cloud computing solution information generation device 1200 includes:

[0245] An extraction module 1201 is used to extract target entity data of the information to be solved in response to a target instruction obtained by inputting the information to be solved in the solution interface of the cloud computing solution system;

[0246] The acquisition module 1202 is used to acquire the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity association data of the target entity data from the pre-built cloud computing knowledge graph;

[0247] A generating module 1203, for generating cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity associated data;

[0248] The output module 1204 is used to output cloud computing solution information.

[0249] In some embodiments, the cloud computing solution information generating device 1200 also includes a communication module and an update module; wherein the communication module is used to send supplementary knowledge data to the target device in response to a specified instruction obtained by inputting supplementary knowledge data in the knowledge supplement interface of the cloud computing solution system; the update module is used to respond to confirmation information on the supplementary knowledge data sent by the target device, and update the cloud computing knowledge graph according to the supplementary knowledge data; the confirmation information is used to indicate that the supplementary knowledge data is accurate; the output module 1204 is also used to output indication information that the supplementary knowledge data is accurate.

[0250] In some embodiments, the cloud computing solution information generating device 1200 also includes an identification module, a pattern layer construction module and a knowledge graph construction module; wherein the identification module is used to pre-process the acquired cloud computing knowledge data and determine the cloud computing text data; the pattern layer construction module is used to construct a pattern layer of the cloud computing knowledge graph based on the attributes of each category in the multi-layer category obtained by category attribute identification of the cloud computing text data; the pattern layer represents the hierarchical association relationship of the attributes of each category in the multi-layer category; the multi-layer target category is a part of the categories in the multi-layer category; the knowledge graph construction module is used to construct a cloud computing knowledge graph based on the pattern layer and the cloud computing text data.

[0251] In some embodiments, the recognition module includes a recognition unit, a text data determination unit, and a communication unit; wherein the recognition unit is used to identify text data and multimedia data in cloud computing knowledge data; the text data determination unit is used to determine the text data in the cloud computing knowledge data as first text data; the communication unit is used to: send multimedia data to a target device; receive second text data of the multimedia data sent by the target device; the second text data is text description information of the multimedia data; wherein the first text data and the second text data are included in the cloud computing text data.

[0252] In some embodiments, the model layer construction module includes an architecture acquisition unit, an attribute identification unit and a model layer determination unit; wherein the architecture acquisition unit is used to obtain the architecture information of a pre-built model layer; the architecture information represents a hierarchical association relationship; the attribute identification unit is used to perform category attribute identification on cloud computing text data to obtain the attributes of each category in the multi-layer category; the model layer determination unit is used to map the attributes of each category in the multi-layer category to the architecture information to obtain the model layer.

[0253] In some embodiments, the knowledge graph construction module includes an extraction unit, a data layer construction unit and a knowledge graph construction unit; wherein the extraction unit is used to extract entity data and entity-related data corresponding to the attributes of each category in the multi-layer categories in the pattern layer from the cloud computing text data; the data layer construction unit is used to construct the data layer of the cloud computing knowledge graph based on the entity data and the entity-related data; the knowledge graph construction unit is used to construct the cloud computing knowledge graph based on the pattern layer and the data layer.

[0254] In some embodiments, the cloud computing solution information generating device 1200 also includes a structure conversion module; the identification module is also used to identify structured data and semi-structured data in cloud computing knowledge data; the structure conversion module is used to perform structural conversion on structured data and semi-structured data to obtain a first knowledge graph; the knowledge graph construction module includes a second knowledge graph construction unit and a cloud computing knowledge graph construction unit; the second knowledge graph construction unit is used to construct a second knowledge graph based on the model layer and cloud computing text data; the cloud computing knowledge graph construction unit is used to fuse the first knowledge graph and the second knowledge graph to obtain a cloud computing knowledge graph.

[0255] The description of the above device embodiment is similar to the description of the above method embodiment, and has similar beneficial effects as the method embodiment. For technical details not disclosed in the device embodiment of the present application, please refer to the description of the method embodiment of the present application for understanding.

[0256] Each module in the above-mentioned cloud computing solution information generation device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0257] In an exemplary embodiment, Fig.13A schematic diagram of the structure of a computer device provided for some embodiments, the computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through wireless fidelity (Wireless Fidelity, WIFI), a mobile cellular network, near field communication (Near Field Communication, NFC) or other technologies. When the computer program is executed by the processor, a cloud computing solution is implemented to generate information. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0258] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0259] For example, a computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method of any of the above embodiments when executing the computer program.

[0260] For example, in an exemplary embodiment, the processor is used to execute a computer program to implement: in response to a target instruction obtained by inputting information to be solved in a solution interface of a cloud computing solution system, extracting target entity data of the information to be solved; obtaining, from a pre-built cloud computing knowledge graph, attributes of each category in a multi-layer target category to which the target entity data belongs and target entity association data of the target entity data; generating cloud computing solution information that matches the information to be solved based on the attributes of each category in the multi-layer target category and the target entity association data; and outputting the cloud computing solution information.

[0261] In one embodiment, a computer-readable storage medium is provided, and when a computer program is executed by a processor, the steps of the method provided in any of the above embodiments are implemented.

[0262] For example, in an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented: in response to a target instruction obtained by inputting information to be solved in a solution interface of a cloud computing solution system, target entity data of the information to be solved is extracted; from a pre-constructed cloud computing knowledge graph, attributes of each category in a multi-layer target category to which the target entity data belongs and target entity association data of the target entity data are obtained; based on the attributes of each category in the multi-layer target category and the target entity association data, cloud computing solution information matching the information to be solved is generated; and the cloud computing solution information is output.

[0263] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0264] The processor, each functional module or each functional unit in any embodiment of the present application may include any one or more of the following integrations: general-purpose processor, application-specific integrated circuit (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), central processing unit (CPU), graphics processing unit (GPU), embedded neural network processor (neural-network processing units, NPU), controller, microcontroller, microprocessor, programmable logic device, discrete gate or transistor logic device, discrete hardware component, data processing logic based on quantum computing, artificial intelligence (AI) processor, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0265] The memory or computer-readable storage medium in any embodiment of the present application may include at least one of a non-volatile memory and a volatile memory. Non-volatile memory includes the integration of one or more of the following: Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Ferromagnetic Random Access Memory (FRAM), Flash Memory, Magnetic Surface Storage, Optical Disc, Compact Disc Read-Only Memory (CD-ROM), Magnetic Tape, Floppy Disk, Flash Memory, Optical Storage, High-density Embedded Non-volatile Memory, Resistive Random Access Memory (ReRAM), Magnetoresistive Random Access Memory (MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (PCM), Graphene Memory, Volatile Memory, etc. Volatile memory includes an integration of one or more of the following: Random Access Memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), etc.

[0266] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0267] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for generating cloud computing solution information, characterized in that: The method comprises: In response to a target instruction obtained by inputting information to be solved in a solution interface of the cloud computing solution system, extracting target entity data of the information to be solved; Obtaining, from a pre-built cloud computing knowledge graph, attributes of each category in the multi-layer target category to which the target entity data belongs and target entity associated data of the target entity data; Generate cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity associated data; Output the cloud computing solution information.

2. The method according to claim 1, characterized in that: The method further comprises: In response to a designated instruction obtained by inputting supplementary knowledge data in a knowledge supplement interface of the cloud computing solution system, sending the supplementary knowledge data to a target device; In response to confirmation information of the supplementary knowledge data sent by the target device, updating the cloud computing knowledge graph according to the supplementary knowledge data; the confirmation information is used to indicate that the supplementary knowledge data is accurate; Output accurate indication information of the supplementary knowledge data.

3. The method according to claim 1 or 2, characterized in that: The method further comprises: Preprocessing the acquired cloud computing knowledge data to determine cloud computing text data; According to the attributes of each category in the multi-layer category obtained by performing category attribute recognition on the cloud computing text data, a pattern layer of the cloud computing knowledge graph is constructed; the pattern layer represents the hierarchical association relationship of the attributes of each category in the multi-layer category; the multi-layer target category is a part of the categories in the multi-layer category; The cloud computing knowledge graph is constructed according to the pattern layer and the cloud computing text data.

4. The method according to claim 3, characterized in that The step of preprocessing the acquired cloud computing knowledge data to determine cloud computing text data includes: Identifying text data and multimedia data in the cloud computing knowledge data; Determining the text data in the cloud computing knowledge data as first text data, and sending the multimedia data to a target device; receiving second text data of the multimedia data sent by the target device; the second text data is text description information of the multimedia data; The first text data and the second text data are included in the cloud computing text data.

5. The method according to claim 3, characterized in that: The method of constructing a pattern layer of a cloud computing knowledge graph based on the attributes of each category in the multi-layer category obtained by performing category attribute recognition on the cloud computing text data includes: Acquire pre-built architecture information of the model layer; the architecture information represents a hierarchical association relationship; Performing category attribute recognition on the cloud computing text data to obtain attributes of each category in the multi-layer category; The attributes of each category in the multi-layer category are mapped to the architecture information to obtain the mode layer.

6. The method according to claim 3, characterized in that The step of constructing the cloud computing knowledge graph according to the model layer and the cloud computing text data includes: Extracting entity data and entity-related data corresponding to the attributes of each category in the multi-layer categories in the pattern layer from the cloud computing text data; Constructing a data layer of the cloud computing knowledge graph according to the entity data and the entity-related data; The cloud computing knowledge graph is constructed according to the model layer and the data layer.

7. The method according to claim 3, characterized in that The method further comprises: identifying structured data and semi-structured data in the cloud computing knowledge data; Performing structural transformation on the structured data and the semi-structured data to obtain a first knowledge graph; The step of constructing the cloud computing knowledge graph according to the model layer and the cloud computing text data includes: Constructing a second knowledge graph according to the model layer and the cloud computing text data; The first knowledge graph and the second knowledge graph are merged to obtain the cloud computing knowledge graph.

8. A cloud computing solution information generation device, characterized in that: The cloud computing solution information generating device comprises: An extraction module, configured to extract target entity data of the information to be solved in response to a target instruction obtained by inputting the information to be solved in a solution interface of the cloud computing solution system; An acquisition module, used to acquire, from a pre-built cloud computing knowledge graph, the attributes of each category in the multi-layer target category to which the target entity data belongs and the target entity associated data of the target entity data; A generating module, configured to generate cloud computing solution information matching the information to be solved according to the attributes of each category in the multi-layer target category and the target entity associated data; The output module is used to output the cloud computing solution information.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images