Question and answer method and device for training content and electronic equipment

Abstract

The invention relates to the technical field of data processing, and provides a question answering method and device for training content and electronic equipment, and the method comprises the steps: obtaining a question description text of a natural language, which is input by a user side and is related to the training content; performing semantic analysis on the question description text to obtain a key entity and an entity category which represent the intention of a user; target training content matched with the key entity and the entity category is retrieved based on a preset training knowledge base, the target training content serves as an answer of the question description text, and the answer is returned to the user side to be displayed. According to the method, the key entity and the entity category are obtained by semantically analyzing the natural language question input by the user side, and the corresponding target training content is retrieved to serve as the question answer, so that the professional requirement of the user is reduced, and the user experience is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a question and answer method, device and electronic equipment for training content. BACKGROUND

[0002] In the prior art, a training management system is used to manage training content such as training outlines, related units, related training equipment and associated standards of training targets (for example, flight training and driving school car training). The management mode is realized by a traditional database mode, that is, the training content-related keywords are stored in the database, and the corresponding training content is queried by inputting the corresponding keywords in the input box of the training management system interface. Moreover, the keywords need to match the keywords stored in the training management system. Users need to have a certain understanding of a training professional and be familiar with the keywords related to the training content. The professional requirements for users are high, and users cannot use natural language question and answer methods to retrieve related training content, resulting in poor user experience. SUMMARY

[0003] The present application provides a question and answer method, device and electronic equipment for training content to solve the problem that users cannot use natural language question and answer methods to retrieve related training content in the prior art.

[0004] The present application provides a question and answer method for training content, comprising: obtaining a natural language question description text related to training content input by a user terminal; performing semantic analysis on the question description text to obtain a key entity and an entity category representing the user's intention; retrieving target training content matching the key entity and the entity category based on a preset training knowledge base, using the target training content as the answer to the question description text, and returning the answer to the user terminal for display; wherein the training knowledge base stores a knowledge graph of training content, the knowledge graph of training content includes information of key entities and entity categories of training content, and the knowledge graph of training content is constructed based on key entities and related attributes and relationships extracted from natural language texts in training outlines through semantic analysis.

[0005] According to the question and answer method for training content provided by the present application, the question description text is subjected to semantic analysis to obtain a key entity and an entity category representing the user's intention, comprising: dividing the question description text into at least one question sentence according to punctuation marks; analyzing the grammatical structure of each question sentence to extract the subject, predicate, object and conjunction representing the logical relationship of the question sentence; In the case that the logical relation representing conjunction is not extracted from the question sentence, the question sentence is determined as a sub-problem description text; In the case that the logical relation representing conjunction is extracted from the question sentence, the question sentence is split into multiple sub-problem description texts based on the subject or object connected by the conjunction; Each of the sub-problem description texts is subjected to semantic parsing to obtain a sub-key entity and a sub-entity category representing the user's intention in each of the sub-problem description texts.

[0006] According to the question and answer method for training content provided by the application, the target training content matched with the key entity and the entity category is retrieved, and the target training content is taken as the answer of the question description text, which comprises: The sub-target training content matched with the sub-key entity and the sub-entity category is retrieved, and the sub-target training content is taken as a sub-answer, and the sub-answer is returned to the user terminal for display; The sub-answers are merged to obtain the answer of the question description text.

[0007] According to the question and answer method for training content provided by the application, the target training content matched with the key entity and the entity category is retrieved, and the target training content is taken as the answer of the question description text, which comprises: In the case that the target training content retrieved has a time attribute, the target training content and the corresponding time attribute are taken as the answer of the question description text.

[0008] According to the question and answer method for training content provided by the application, it further comprises: The target training content is pushed to the user terminal for display in the order of the time attribute; Two comparison time points selected by the user are received, and the difference information of the target training content corresponding to the two time points is displayed.

[0009] According to the question and answer method for training content provided by the application, it further comprises: According to the target key entity in the target training content, an associated key entity associated with the target key entity is found; The associated training content is generated based on the associated key entity, and the associated training content is pushed to the user terminal for display; The target key entity and the associated key entity belong to the same entity category.

[0010] According to the question and answer method for training content provided by the application, when the target training content includes a training subject, the training equipment and training standards related to the training subject are determined based on a mapping table of the training subject and the training equipment and training standards, and the training equipment and training standards related to the training subject are pushed to the user end for display.

[0011] According to the question and answer method for training content provided by the application, the natural language question description text input by the user end and related to the training content is acquired, including: The original description text of natural language input by the user end and related to the training content is acquired. The original description text is subjected to fuzzy completion processing based on the last N question description texts before the original description text.

[0012] The application further provides a question and answer device for training content, including: A text acquisition module is configured to acquire the natural language question description text input by the user end and related to the training content. A semantic analysis module is configured to perform semantic analysis on the question description text to obtain key entities and entity categories representing user intentions. A retrieval module is configured to retrieve target training content matched with the key entities and entity categories based on a preset training knowledge base, use the target training content as an answer to the question description text, and return the answer to the user end for display. The training knowledge base stores a knowledge graph of the training content, the knowledge graph of the training content includes information of key entities and entity categories of the training content, and the knowledge graph of the training content is constructed based on key entities and related attributes and relationships extracted from natural language texts in a training outline through semantic analysis.

[0013] The application further provides an electronic device including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the question and answer method for training content as described above when executing the program.

[0014] The application provides a question and answer method, device and electronic equipment for training content. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor.

[0016] Figure 1 FIG. 1 is a flowchart of the question and answer method for training content provided by the application.

[0017] Figure 2 FIG. 2 is a structural diagram of the question and answer device for training content provided by the application.

[0018] Figure 3 FIG. 3 is a structural diagram of the electronic equipment provided by the application. DETAILED DESCRIPTION

[0019] In order to make the objects, technical solutions and advantages of the application clearer, the technical solutions in the application will be described clearly and completely below with reference to the drawings in the application. Obviously, the described embodiments are some embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0020] The question and answer method for training content provided by the embodiment of the application can be the training management system, and the specific process is as shown in FIG. 1, including steps S110 to S130. Figure 1 The question and answer method for training content provided by the embodiment of the application can be the training management system, and the specific process is as shown in FIG. 1, including steps S110 to S130.

[0021] Step S110: Obtain the natural language question description text related to the training content input by the user end. Specifically, the user inputs the natural language question description text related to the training content to be queried in the question dialogue box of the training management system user end, and the training management system obtains the natural language question description text related to the training content through the question dialogue box.

[0022] Step S120: Perform semantic analysis on the question description text to obtain key entities and entity categories representing user intent. Specifically, in this step, the Named Entity Recognition (NER) technology can be used to identify key entities and determine entity categories in the question description text. For example, if the user inputs "What is the training outline of a certain unit in 2023?", the key entities after analysis include "a certain unit", "2023", and "training outline", and the categories of the key entities are unit name, time, and outline category, respectively.

[0023] Step S130: Based on the pre-set training knowledge base, retrieve the target training content matching the key entities and entity categories, use the target training content as the answer to the question description text, and return the answer to the user end for display.

[0024] Specifically, based on the pre-set training knowledge base retrieval, the training knowledge base stores a knowledge graph of training content, which includes information of key entities and entity categories (one of the attributes) of the training content. The knowledge graph of the training content is constructed based on the natural language text in the training outline after semantic analysis (named entity recognition) to extract key entities and related attributes and relationships. Therefore, the key entities and entity categories obtained by semantic analysis of the natural language input by the user end can be used to retrieve the corresponding target training content.

[0025] The question and answer method for training content of the embodiment, by acquiring the natural language question description text input by the user terminal and related to the training content; performing semantic analysis on the question description text to obtain the key entity and entity category representing the user's intention; retrieving the target training content matching the key entity and entity category based on the pre-set training knowledge base, taking the target training content as the answer to the question description text, and returning the answer to the user terminal for display; wherein the training knowledge base stores the knowledge graph of the training content, the knowledge graph of the training content includes information of the key entity and entity category of the training content, and the knowledge graph of the training content is constructed based on the key entity and related attributes and relationships extracted by semantic analysis on the natural language text in the training outline. Therefore, the key entity and entity category can be obtained by semantic analysis on the natural language question input by the user terminal, and the corresponding target training content is retrieved as the answer to the question, thereby reducing the professional requirement of the user and improving the user experience.

[0026] Since the natural language question description text input by the user terminal can include multiple questions, in order to more accurately analyze the key entity and entity category in the multiple questions, in some embodiments, step S120 specifically includes: Step S121: dividing the question description text into at least one question sentence according to the punctuation marks, because each question sentence usually contains one question, for example: "What is the training outline of a certain unit in 2023, and what equipment and standards are required for a certain training subject in the training outline". The question description text has two question sentences "What is the training outline of a certain unit in 2023" and "What equipment and standards are required for a certain training subject in the training outline", and the second question sentence involves two questions of equipment and standards.

[0027] Step S122: analyzing the grammatical structure of each question sentence to extract the subject, predicate, object and conjunction representing the logical relationship of the question sentence. For example: for the second question sentence above, after extracting the subject, predicate, object and conjunction representing the logical relationship, we get: Subject: a certain training subject in the training outline; predicate: required; object: equipment and standards; conjunction: and.

[0028] Step S123: in the case where the question sentence does not extract the conjunction representing the logical relationship, the question sentence is determined as a sub-question description text. For example: the first question sentence above has no conjunction, which is a sub-question description text, i.e. "What is the training outline of a certain unit in 2023".

[0029] Step S124: In the case that the conjunction representing the logical relationship is extracted from the question sentence, the question sentence is split into multiple sub-question description texts based on the subject or object connected by the conjunction. Specifically, the question sentence is split into multiple sub-question description texts according to the number of the subject or object connected by the conjunction.

[0030] For example, the conjunction "and" is extracted from the second question sentence, and the object connected by the conjunction has two: equipment and standard, and thus the second question sentence is split into two sub-question description texts, which are "the equipment required for a training subject in the training outline" and "the standard required for a training subject in the training outline", respectively. Thus, the question description text "what is the training outline of a certain unit in 2023, and what is the equipment and standard required for a training subject in the training outline" is split into three sub-question description texts.

[0031] It should be noted that the standard of the training subject (i.e., the training standard) is usually defined in the training outline, and can be obtained by searching the knowledge graph corresponding to the training outline in the training knowledge base.

[0032] Step S125: Semantic analysis is performed on each of the sub-question description texts to obtain the sub-key entity and the sub-entity category representing the user's intention in each of the sub-question description texts. The sub-key entity and the sub-entity category obtained by performing semantic analysis on the above three sub-question description texts are as follows.

[0033] Sub-question description text one: "what is the training outline of a certain unit in 2023", the semantic analysis obtains the sub-key entity as: a certain unit, 2023, and training outline, and the corresponding sub-entity categories are: unit name, time, and outline category.

[0034] Sub-question description text two: "the equipment required for a training subject in the training outline", the semantic analysis obtains the sub-key entity as: the training outline, a training subject, and equipment, and the corresponding sub-entity categories are: outline, subject, and equipment category.

[0035] Sub-question description text three: "the standard required for a training subject in the training outline", the semantic analysis obtains the sub-key entity as: the training outline, a training subject, and standard, and the corresponding sub-entity categories are: outline, subject, and standard category.

[0036] In this embodiment, the complex natural language question description text is decomposed into simple sub-question description texts, so that the key entities and their entity categories in multiple questions can be more accurately parsed, some key entities and their entity categories can be avoided to be missed, and thus the final answer is more consistent with the question description text.

[0037] In some embodiments, based on the processing of steps S121 to S125, step S130 specifically comprises: Retrieving sub-target training content matching the sub-key entity and the sub-entity category, taking the sub-target training content as a sub-answer, and returning the sub-answer to the user terminal for display to facilitate user viewing.

[0038] Merging each sub-answer to obtain an answer to the problem description text. Specifically, each sub-answer is merged into a piece of text as an answer through natural language processing (NLP) technology. For example, the sub-answers can be merged according to user instructions after the user views the sub-answers and confirms that each sub-answer is correct.

[0039] In this embodiment, after retrieving the sub-answers according to the sub-key entity and the sub-entity category, the sub-answers are returned to the user terminal for display, which facilitates user viewing and confirmation of whether the returned answers meet their own requirements, and then each sub-answer is merged into an answer.

[0040] Further, after obtaining each sub-problem description text, the sub-problem description text and the corresponding problem description text are displayed in a hierarchical structure on the user terminal, and each sub-problem description text is displayed in association with the corresponding sub-answer. This allows the user to click to view the independent analysis sub-answers of each sub-problem, thereby improving user experience.

[0041] In some embodiments, retrieving target training content matching the key entity and the entity category, and taking the target training content as an answer to the problem description text, comprises: In the case where the retrieved target training content has a time attribute, the target training content and the corresponding time attribute are taken as an answer to the problem description text. The time attribute includes the formulation time and / or the applicable time range of the training content. For example, a training outline of a certain training target is retrieved. The training outline usually has a formulation time and an applicable time range (e.g., formulated and published in January 2023, and applicable for the period 2023-2025). The time attribute of the training outline is taken as an answer, i.e., the returned answer is: a certain training outline, formulation time: January 2023, applicable time range: 2023-2025.

[0042] In this embodiment, in the case where the retrieved target training content has a time attribute, the target training content and the corresponding time attribute are taken as an answer to the problem description text, which can help the user quickly locate the required training content.

[0043] Further, after taking the target training content and the corresponding time attribute as an answer to the problem description text, the method further comprises: The target training content is pushed to the user end in the order of the time attribute, for example, a training outline of a certain training target has multiple different versions, which can be displayed on the user end in the order of the time attribute of different versions.

[0044] The two selected comparison time points are received, and the difference information of the target training content corresponding to the two time points is displayed. Since the training content is displayed in the order of the time attribute, the user can easily select the training content of at least two time points to compare the differences. For example, the training outline in 2020 contains "basic training", and the training outline in 2023 adds "special training".

[0045] In this embodiment, the historical change trend of the training content (such as training subjects, training outlines, and equipment standards) can be more intuitively presented in the form of a time axis, helping the user to understand the adjustment record and application scope of a certain content at different time points.

[0046] In some embodiments, the question and answer method for training content further comprises: According to the target key entity in the target training content, an associated key entity associated with the target key entity is found, wherein the target key entity and the associated key entity belong to the same entity category. For example, the associated relationship can be determined based on the attribute similarity of the key entities of the same entity category, for example, the similarity of the attributes of the target key entity and any key entity of the same entity category is calculated, and when the similarity is greater than a preset threshold, it is considered that any key entity is an associated key entity associated with the target key entity.

[0047] Based on the associated key entity, an associated training content is generated, and the associated training content is pushed to the user end for display.

[0048] For example, the target key entity in the target training content (such as the basic training outline of A unit in 2023) is "A unit, 2023, basic training outline", and the key entity "A unit, 2024, advanced training outline" is found to have a high similarity with the target key entity. Therefore, the associated training content generated according to "A unit, 2024, advanced training outline" is "A unit 2024 advanced training outline", which is pushed to the user end as an associated answer.

[0049] In this embodiment, by recommending the training content of the associated key entity related to the target key entity in the target training content, the user is quickly provided with high-relevance alternative training content, improving the user experience.

[0050] In some embodiments, the method for training content further comprises, in the case that the target training content includes a training subject, determining the training equipment and training standards related to the training subject based on a mapping table of training subjects and training equipment and training standards respectively, and pushing the training equipment and training standards related to the training subject to the user end for display.

[0051] For example, the retrieved training subject is physical training. In addition to returning the physical training to the user end, the corresponding training equipment is obtained based on the mapping table of physical training and training equipment and training standards respectively, which is: Equipment 1: treadmill, quantity: 5.

[0052] Equipment 2: dumbbell, specification: 10 kg, quantity: 10.

[0053] The corresponding standards are: Standard 1: Physical Training Operation Specification 2023 Edition.

[0054] Standard 2: Physical Training Assessment Index 2023 Edition.

[0055] The above training equipment and standard information are returned to the user end together with the physical training for display.

[0056] In this embodiment, the training equipment and standards related to the training subject are found through the mapping table of training subjects and training equipment and training standards, which helps users to fully understand the equipment configuration and standard requirements required for a certain training subject.

[0057] In some embodiments, the natural language question description text related to the training content input by the user end is obtained, including: The original description text of natural language related to the training content input by the user end is obtained; the original description text is processed for fuzzy completion based on the last N (for example, N=5) question description texts before the original description text. The N-Gram model can be used to process the original description text for fuzzy completion.

[0058] For example, when the current original description text input by the user is "What training subjects are included in the training outline", the last N question description texts are directly completed as "a certain unit" accessed recently, accordingly, the completed current question description text is "What training subjects are included in the training outline of a certain unit". Or prompt "whether it refers to the training outline of 'A unit' or 'B unit'", according to the unit confirmed by the user, such as 'A unit', the completion obtains "What training subjects are included in the training outline of A unit".

[0059] In this embodiment, the original description text of the natural language input by the user is completed to obtain a more complete problem description text, which is beneficial to obtaining a more accurate answer.

[0060] The question and answer device for training content provided by the present application is described below, and the question and answer device for training content described below can be correspondingly referred to the question and answer method for training content described above.

[0061] The question and answer device for training content of the embodiment of the present application, as shown in the figure, comprises: Figure 2 The text acquisition module 210 is configured to acquire the natural language problem description text related to the training content input by the user terminal.

[0062] The semantic analysis module 220 is configured to perform semantic analysis on the problem description text to obtain key entities and entity categories representing the user's intention.

[0063] The retrieval module 230 is configured to retrieve target training content matching the key entities and entity categories based on the preset training knowledge base, use the target training content as the answer to the problem description text, and return the answer to the user terminal for display.

[0064] The training knowledge base stores a knowledge graph of the training content, and the knowledge graph of the training content includes information of key entities and entity categories of the training content. The knowledge graph of the training content is constructed based on key entities and related attributes and relationships extracted from the natural language text in the training outline through semantic analysis.

[0065] In some embodiments, the semantic analysis module 220 is specifically configured to: divide the problem description text into at least one question sentence according to punctuation marks.

[0066] analyze the grammatical structure of each question sentence, extract the subject, predicate, object, and conjunction representing the logical relationship of the question sentence.

[0067] In the case where the conjunction representing the logical relationship is not extracted from the question sentence, the question sentence is determined as a sub-problem description text.

[0068] In the case where the conjunction representing the logical relationship is extracted from the question sentence, the question sentence is split into multiple sub-problem description texts based on the subject or object connected by the conjunction.

[0069] Each sub-problem description text is subjected to semantic analysis to obtain sub-key entities and sub-entity categories representing the user's intention in each sub-problem description text. ​

[0070] In some embodiments, the retrieving module 230 is specifically configured to: retrieve sub-target training content matching the sub-key entity and the sub-entity category, take the sub-target training content as a sub-answer, and return the sub-answer to the user terminal for display.

[0071] merge the sub-answers to obtain an answer to the problem description text.

[0072] In some embodiments, the retrieving module 230 is specifically configured to, in a case where the retrieved target training content has a time attribute, take the target training content and the corresponding time attribute as an answer to the problem description text.

[0073] In some embodiments, the question and answer device for training content further includes a difference display module configured to push the target training content to the user terminal for display in the order of the time attribute, receive two user-selected comparison time points, and display difference information of the target training content corresponding to the two time points.

[0074] In some embodiments, the question and answer device for training content further includes an associated content generation module configured to find, according to a target key entity in the target training content, an associated key entity associated with the target key entity, generate associated training content based on the associated key entity, and push the associated training content to the user terminal for display, wherein the target key entity and the associated key entity belong to the same entity category.

[0075] In some embodiments, the question and answer device for training content further includes a training subject analysis module configured to, in a case where the target training content includes a training subject, determine training equipment and training standards related to the training subject based on a mapping table of the training subject with respect to the training equipment and the training standards, and push the training equipment and the training standards related to the training subject to the user terminal for display.

[0076] In some embodiments, the text acquisition module 210 is specifically configured to: acquire a natural language original description text input by a user terminal and related to training content.

[0077] perform fuzzy completion processing on the original description text based on the last N pieces of problem description text before the original description text.

[0078] Figure 3 An example of an entity structure diagram of an electronic device is shown in FIG. 1. Figure 3As shown, the electronic device can include a processor 310, a communications interface 320, a memory 330, and a communications bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other through the communications bus 340. The processor 310 can invoke the logical instructions in the memory 330 to execute a question and answer method for training content, which includes: Obtaining a natural language question description text input by a user terminal and related to the training content.

[0079] Performing semantic analysis on the question description text to obtain a key entity and an entity category representing the user's intention.

[0080] Retrieving target training content matching the key entity and the entity category based on a preset training knowledge base, taking the target training content as an answer to the question description text, and returning the answer to the user terminal for display.

[0081] Wherein, the training knowledge base stores a knowledge graph of the training content, the knowledge graph of the training content includes information of key entities and entity categories of the training content, and the knowledge graph of the training content is constructed based on key entities and related attributes and relationships extracted from natural language texts in a training outline through semantic analysis.

[0082] In addition, the logical instructions in the memory 330 described above can be implemented in the form of a software function unit and sold or used as an independent product, which can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or parts of the prior art that essentially contribute or the part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for making a computer device (which can be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various storage medium that can store program codes.

[0083] On the other hand, the present application also provides a computer program product, which includes a computer program, the computer program can be stored on a non-transitory computer readable storage medium, and the computer program is executed by a processor, the computer can execute the question and answer method for training content provided by the above-mentioned method, which includes: Acquire natural language question description text input by a user terminal and related to training content.

[0084] Perform semantic analysis on the question description text to obtain key entities and entity categories representing user intent.

[0085] Retrieval of target training content matching the key entities and entity categories based on a preset training knowledge base, use of the target training content as an answer to the question description text, and return of the answer to the user terminal for display.

[0086] The training knowledge base stores a knowledge graph of training content, the knowledge graph of training content includes information of key entities and entity categories of training content, and the knowledge graph of training content is constructed based on key entities and related attributes and relationships extracted from natural language text in a training outline through semantic analysis.

[0087] In another aspect, the present application also provides a non-transitory computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor to implement the question and answer method for training content provided by the above-mentioned methods, the method comprising: Acquire natural language question description text input by a user terminal and related to training content.

[0088] Perform semantic analysis on the question description text to obtain key entities and entity categories representing user intent.

[0089] Retrieval of target training content matching the key entities and entity categories based on a preset training knowledge base, use of the target training content as an answer to the question description text, and return of the answer to the user terminal for display.

[0090] The training knowledge base stores a knowledge graph of training content, the knowledge graph of training content includes information of key entities and entity categories of training content, and the knowledge graph of training content is constructed based on key entities and related attributes and relationships extracted from natural language text in a training outline through semantic analysis.

[0091] The device embodiments described above are only schematic, wherein the units illustrated as separate components can or can not be physically separate, and the components illustrated as units can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the present embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0092] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0093] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A question and answer method for training content, characterized by, The method comprises the following steps: acquiring natural language question description text related to training content input by a user terminal; performing semantic analysis on the question description text to obtain key entities and entity categories representing user intent; retrieving target training content matching the key entities and entity categories based on a preset training knowledge base, using the target training content as the answer to the question description text, and returning the answer to the user terminal for display; wherein the training knowledge base stores a knowledge graph of training content, the knowledge graph of training content includes information of key entities and entity categories of training content, and the knowledge graph of training content is constructed based on key entities and related attributes and relationships extracted from natural language text in a training outline through semantic analysis. 2.The question and answer method for training content of claim 1, wherein, The semantic analysis on the question description text to obtain key entities and entity categories representing user intent comprises the following steps: dividing the question description text into at least one question sentence according to punctuation marks; analyzing the grammatical structure of each question sentence to extract the subject, predicate, object, and conjunction representing logical relationships of the question sentence; in the case where no conjunction representing logical relationships is extracted from the question sentence, determining the question sentence as a sub-question description text; in the case where a conjunction representing logical relationships is extracted from the question sentence, splitting the question sentence into multiple sub-question description texts based on the subject or object connected by the conjunction; performing semantic analysis on each sub-question description text to obtain sub-key entities and sub-entity categories representing user intent in each sub-question description text. 3.The question and answer method for training content of claim 2, wherein, The retrieval of target training content matching the key entities and entity categories, using the target training content as the answer to the question description text, comprises the following steps: retrieving sub-target training content matching the sub-key entities and sub-entity categories, using the sub-target training content as a sub-answer, and returning the sub-answer to the user terminal for display; merging each sub-answer to obtain the answer to the question description text. 4.The question and answer method for training content of claim 1, wherein, The retrieval of target training content matching the key entities and entity categories, using the target training content as the answer to the question description text, comprises the following steps: in the case where the retrieved target training content has a time attribute, using the target training content and the corresponding time attribute as the answer to the question description text.

5. The question and answer method for training content of claim 4, wherein, Further comprising: pushing the target training content to the user terminal for display in the order of the time attribute; receiving two comparison time points selected by the user and displaying the difference information of the target training content corresponding to the two time points.

6. The question and answer method for training content of claim 1, wherein, Further comprising: finding associated key entities associated with the target key entities in the target training content based on the target key entities; generating associated training content based on the associated key entities and pushing the associated training content to the user terminal for display; wherein the target key entities and the associated key entities belong to the same entity category. 7.The question and answer method for training content of claim 1, wherein, Further comprising: In the case that the target training content includes a training subject, the training subject related training equipment and training standard are determined based on a mapping table of training subjects and training equipment and training standard respectively, and the training subject related training equipment and training standard are pushed to the user end for display.

8. The question and answer method for training content according to any one of claims 1 to 7, characterized in that, The natural language question description text related to the training content input by the user end is acquired, including: The natural language original description text related to the training content input by the user end is acquired; The original description text is fuzzily completed based on the last N question description texts before the original description text.

9. A question and answer apparatus for training content, characterized by, Including: A text acquisition module is configured to acquire the natural language question description text related to the training content input by the user end; A semantic analysis module is configured to perform semantic analysis on the question description text to obtain key entities and entity categories representing user intentions; A retrieval module is configured to retrieve target training content matching the key entities and entity categories based on a preset training knowledge base, use the target training content as the answer to the question description text, and return the answer to the user end for display; The training knowledge base stores a knowledge graph of training content, the knowledge graph of training content includes information of key entities and entity categories of the training content, and the knowledge graph of training content is constructed based on key entities and related attributes and relationships extracted from natural language texts in a training outline through semantic analysis.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and running on the processor, characterized in that, The processor executes the computer program to implement the question and answer method for training content according to any one of claims 1 to 8.

Images