Reply content generation method, server device and storage medium

Abstract

The present application provides a reply content generation method, a server device and a storage medium. The method includes: receiving a content generation request sent by a client device; the content generation request includes a question text; performing semantic analysis on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text; obtaining production management information related to the question text according to the keywords and entities of the question text; generating reply content reference data according to the semantic analysis result and the production management information and sending it to the client device, the reply content reference data includes production management information, so that the client device generates and outputs reply content according to the reply content reference data. The method of the present application can apply the artificial intelligence generated content model to actual production management, improving production efficiency and management efficiency.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method for generating reply content, a server device, and a storage medium. Background Art

[0002] The AIGC model can interact with users through dialogue, answer questions raised by users, and improve user work efficiency.

[0003] However, most of the current AIGC models can only answer some general knowledge based on the information input by users and public information on the Internet, and cannot be applied in actual production management. Summary of the invention

[0004] The present application provides a reply content generation method, a server device and a storage medium to solve the problem that the current AIGC model can only answer some general knowledge based on the information input by the user and the public information on the Internet, and cannot be applied in actual production management.

[0005] In a first aspect, the present application provides a method for generating reply content, which is applied to a server device, and the method for generating reply content includes:

[0006] Receiving a content generation request sent by a client device; the content generation request includes a question text;

[0007] Performing semantic analysis on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text;

[0008] Acquire production management information related to the question text based on the keywords and entities in the question text;

[0009] Reply content reference data is generated based on the semantic analysis result and the production management information and sent to the client device. The reply content reference data includes the production management information, so that the client device generates and outputs reply content based on the reply content reference data.

[0010] In a second aspect, the present application provides a reply content generation device, which is applied to a server device, and the reply content generation device includes:

[0011] A receiving module, configured to receive a content generation request sent by a client device; the content generation request includes a question text;

[0012] An analysis module, used to perform semantic analysis on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text;

[0013] An acquisition module, used to acquire production management information related to the question text according to the keywords and entities of the question text;

[0014] A generation module is used to generate reply content reference data based on the semantic analysis results and the production management information and send it to the client device. The reply content reference data includes the production management information, so that the client device generates and outputs the reply content based on the reply content reference data.

[0015] In a third aspect, the present application provides a server device, comprising: a processor, and a memory communicatively connected to the processor;

[0016] The memory stores computer-executable instructions;

[0017] The processor executes the computer-executable instructions stored in the memory to implement the reply content generation method as described in the first aspect.

[0018] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the reply content generation method as described in the first aspect.

[0019] In a fifth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, is used to implement the reply content generation method as described in the first aspect.

[0020] The reply content generation method, server device and storage medium provided in the present application receive a content generation request sent by a client device; the content generation request includes a question text; a semantic analysis is performed on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text; production management information related to the question text is obtained according to the keywords and entities of the question text; reply content reference data is generated according to the semantic analysis result and the production management information and sent to the client device, the reply content reference data includes the production management information, so that the client device generates and outputs the reply content according to the reply content reference data. After the semantic analysis result of the question text is obtained, the production management information related to the semantic analysis result of the question text is queried to obtain the actual production management information related to the question text. Therefore, the reply content reference data generated according to the semantic analysis result and the production management information can be associated with the actual production management, thereby realizing the application of the artificial intelligence generated content model in the actual production management, thereby improving production efficiency and management efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0022] Figure 1 An application scenario diagram of the reply content generation method provided in an embodiment of the present application;

[0023] Figure 2 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 1 ;

[0024] Figure 3 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 2 ;

[0025] Figure 4 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 3 ;

[0026] Figure 5 This is a schematic diagram of the structure of the server device provided in Example 5 of the present application.

[0027] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0028] The terms "first", "second", "third", etc. in the description and claims of the embodiments of the present invention and the drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0030] First, the terms involved in this application are explained:

[0031] Artificial Intelligence Generated Content (AIGC) is a technology that uses machine learning, deep learning, big data and other technical means to automatically generate new content by learning and understanding existing data.

[0032] Semantic Analysis is a technology that deeply understands and analyzes natural language texts to identify semantic information such as entities, relationships, and emotions in the text. Semantic analysis can convert natural language texts into a form that computers can understand and process.

[0033] Keywords, entities, and topics are all concepts in natural language processing (NLP).

[0034] Keywords are words or phrases extracted from text that can represent the theme or core meaning of the text.

[0035] An entity refers to a specific element in a text, such as a person's name, place, organization, date, etc. Entities can usually be clearly identified in a text and carry certain semantic information.

[0036] The topic represents the central idea or main topic expressed by the entire text. Compared with keywords, topics tend to summarize the global information of the text rather than specific words.

[0037] The following is a detailed description and analysis of the prior art involved in this application.

[0038] The AIGC model can interact with users through dialogue, answer questions raised by users, help users extract useful information from large amounts of data, save users information retrieval, and improve user work efficiency.

[0039] However, the current AIGC model usually provides services to users through web applications. Users can enter questions or instructions in the input box of the web page to interact with the AIGC model. The AIGC model can answer some general knowledge based on the information entered by the user and the public information on the Internet. However, the public information on the Internet used by AIGC is usually not the most timely information, but is collected from the Internet in advance and stored in the database. This results in the user being unable to obtain the most timely information during the interaction with the AIGC model because the database has not been updated.

[0040] Regarding the above-mentioned manifestations in the prior art, the inventors discovered through creative research that since the current AIGC model’s method of generating reply content cannot obtain data in real time, the current AIGC model cannot provide the most timely information, and the reply content is divorced from reality. This results in the current AIGC model being often divorced from reality and having no reference value when applied to actual production management with strong timeliness requirements. Therefore, the current AIGC model cannot be applied to actual production management with strong timeliness requirements.

[0041] In the face of problems in the prior art, the present application proposes the following technical concept: In order to enable the AIGC model to be applied in actual production management, it is necessary to enable the AIGC model to respond to time-sensitive information. Therefore, when generating the response content of the AIGC model, it is necessary to combine the actual production management information. Therefore, based on the questions or instructions input by the user, the production management information related to the questions or instructions input by the user can be obtained, and then the response to the question or instruction can be generated in combination with the production management information related to the question or instruction, and the response content will not be separated from the actual production management.

[0042] The following will introduce the application scenarios and network architecture of the reply content generation method provided by the present application. When the following description refers to the drawings, unless otherwise indicated, the same data in different drawings represent the same or similar elements.

[0043] Figure 1 This is an application scenario diagram of the reply content generation method provided in the embodiment of the present application. Figure 1 As shown, in an application scenario of the reply content generation method provided in the embodiment of the present application, it includes: a client device 11, a server device 12, and a production management database 13. The server device 12 is connected to the client device 11 and the production management database 13 in communication.

[0044] The client device 11 may be configured with a client application, and the user may ask questions to the AIGC model in the interactive interface of the client application, asking questions related to production management, including inputting question text or inputting question voice, etc. For example, the user may ask questions about the operating status of device No. 0001 and the current liquid level of device No. 0002. When the user inputs the question voice, the client device 11 may convert the question voice into question text.

[0045] After receiving the user's question text or question voice, the client device 11 sends a content generation request to the server device 12; the content generation request includes the question text; the question text can be the text converted from the question text or the question voice.

[0046] The server device 12 may be configured with a server application for executing the reply content generation method provided in this application.

[0047] The server device 12 receives the content generation request sent by the client device; the content generation request includes the question text.

[0048] The server device 12 performs semantic analysis on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text.

[0049] The server device 12 obtains production management information related to the question text according to the keywords and entities of the question text. The production management database 13 stores production management information generated in the actual production management process, and the server device 12 can obtain production management information related to the question text from the production management database 13.

[0050] The server device 12 generates reply content reference data based on the semantic analysis result and the production management information and sends it to the client device 11. The reply content reference data includes the production management information so that the client device 11 generates and outputs reply content based on the reply content reference data.

[0051] The technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems are described in detail below with specific embodiments. The implementation methods described in the following embodiments do not represent all implementation methods consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims. The following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0052] Embodiment 1

[0053] Figure 2 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 1 .like Figure 2 As shown, the execution subject of this embodiment is a reply content generation device, and the reply content generation device is located in the server device. The reply content generation method provided by this embodiment includes steps 201 to 204.

[0054] Step 201: receiving a content generation request sent by a client device; the content generation request includes a question text.

[0055] In this embodiment, a client application is configured on the client device, and the client application has components such as an input box or an input icon for interacting with the user. The user can input a question text to the client device by interacting with the input box, or input a question voice to the client device by interacting with the input icon. After the client device receives the question text or question voice input by the user, it can generate a content generation request based on the question text or question voice. For example, the client device can determine the question text as the question text; convert the question voice into text form to obtain the question text; correct the question text to obtain the question text.

[0056] After obtaining the question text, the client device sends a content generation request to the server device. Then, the server device receives the content generation request sent by the client device.

[0057] Exemplarily, the question text may be: the commissioning status of equipment No. 0001; the current liquid level of equipment No. 0002; the power consumption trend of equipment No. 0003 this month; the total power consumption of the factory this month.

[0058] Step 202, perform semantic analysis on the question text to obtain the semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text.

[0059] In this embodiment, semantic analysis may include three steps: extracting keywords, identifying entities, and determining topics. For example, the keyword frequency and inverse document frequency TF-IDF algorithm may be used to extract keywords from the question text based on word frequency, word importance, or co-occurrence relationship with other words. The entities of the question text may be identified through a bidirectional long short-term memory LSTM model, a bidirectional encoder representation BERT model, etc., and the topic of the question text may be determined by analyzing the vocabulary distribution of the question text through a Latent Dirichlet Allocation (LDA) algorithm, a BERT model, etc.

[0060] Among them, TF-IDF stands for Term Frequency-Inverse Document Frequency; LSTM stands for Long Short-Term Memory; BERT stands for Bidirectional Encoder Representations from Transformers.

[0061] In this embodiment, by performing semantic analysis on the question text, the keywords, entities and topics of the question text can be obtained. Continuing with the example of step 201, after semantic analysis, the keywords of the question text "the commissioning status of device No. 0001" can include "device No. 0001" and "commissioning status", the entities can include "device No. 0001" and "commissioning status", and the topics can include "device No. 0001" and "commissioning status".

[0062] After semantic analysis, the keywords of the question text "What is the current liquid level of device No. 0002" may include "device No. 0002", "current", "liquid level" and "how much", the entities may include "device No. 0002" and "liquid level", and the topics may include "device No. 0002" and "liquid level".

[0063] After semantic analysis, the keywords of the question text "Power consumption trend of device No. 0003 this month" may include "device No. 0003", "this month", "power consumption" and "trend", the entities may include "device No. 0003", "this month" and "power consumption", and the topics may include "device No. 0003", "this month", "power consumption" and "trend".

[0064] After semantic analysis, the keywords of the question text "Total electricity consumption this month" may include "this month", "total" and "electricity consumption", the entities may include "this month" and "electricity consumption", and the topics may include "this month", "total" and "electricity consumption".

[0065] Step 203, obtaining production management information related to the question text according to the keywords and entities of the question text.

[0066] In this embodiment, production management information refers to data generated by actual production management activities. The production management information can be stored in real time in a production management database. The production management database communicates with the server device, and the server device can obtain production management information related to the question text from the production management database.

[0067] Exemplarily, when production management involves multiple devices, the production management information may include the status data of each device, and the status data of each device may be monitored according to a preset monitoring cycle and stored in the production management database. The status data of different devices may be different. For example, when the device is a feed barrel, the status data may include the liquid level; when the device is a dosing machine, the status data may include the commissioning status, the running time, the last maintenance time, etc. The production management database may store the current status data and historical status data of each device. The current status data is the status data of the current monitoring cycle, and the historical status data is the status data of the monitoring cycle between the current monitoring cycle.

[0068] In this embodiment, the server device can query the production management database for data including keywords of the question text and data including text corresponding to entities of the question text, and then determine the queried data as production management information related to the question text.

[0069] Step 204, generate reply content reference data based on the semantic analysis results and production management information and send it to the client device, the reply content reference data includes production management information, so that the client device generates and outputs reply content based on the reply content reference data.

[0070] In this embodiment, a trained AIGC model may be stored in the server-side device, and the server-side device may input the question text, semantic analysis results, and production management information into the trained AIGC model, so that the trained AIGC model combines the public information, semantic analysis results, and production management information on the Internet to generate and output a response to the question text, and determines the trained AIGC model's response to the question text and the production management information as reference data for the response content.

[0071] In this embodiment, after the server sends the reply content reference data to the client device, the client device can determine the complete reply content reference data or partial reply content reference data as the reply content and output it, so that the user can obtain a reply to the input question text or question voice.

[0072] For example, when the question text input by the user is "What is the current liquid level of device No. 0001", the reply content output by the client device may be "The current liquid level of feed barrel No. 0001 is 51.69ml" or "51.69ml"; when the question text input by the user is "The commissioning status of device No. 0002", the reply content output by the client device may be "The operating status of device No. 0002 is not in operation", "The commissioning status of device No. 0002 is not in use", "Device No. 0002 is not in use" or "Not in operation", etc., without limitation here.

[0073] In this embodiment, since the server-side device performs semantic analysis on the question text and obtains the semantic analysis result of the question text, the semantic analysis result includes the keywords and entities of the question text, and then the production management information related to the question text can be obtained based on the keywords and entities of the question text; and the production management information is information obtained from the data generated by production management. Therefore, reply content reference data is generated based on the semantic analysis result and the production management information and sent to the client device. The reply information generated by the client device based on the reply content reference data may include real-time information in production management. After the user inputs the question text or question voice, the reply information output by the client device is combined with the actual production management, is not separated from the actual production management, and has a certain reference value, so that the AIGC model can be used in actual production management to improve production efficiency and management efficiency.

[0074] The reply content generation method provided in this embodiment receives a content generation request sent by a client device; the content generation request includes a question text; a semantic analysis is performed on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text; production management information related to the question text is obtained based on the keywords and entities of the question text; reply content reference data is generated based on the semantic analysis result and the production management information and sent to the client device, the reply content reference data includes production management information, so that the client device generates and outputs reply content based on the reply content reference data. After the semantic analysis result of the question text is obtained, the production management information related to the semantic analysis result of the question text is queried to obtain information in actual production management related to the question text. Therefore, the reply content reference data generated based on the semantic analysis result and the production management information can be associated with the actual production management, thereby enabling the artificial intelligence generated content model to be applied in actual production management to improve production efficiency and management efficiency.

[0075] Embodiment 2

[0076] Figure 3 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 2 .like Figure 3 As shown, the reply content generation method provided in this embodiment is based on the first embodiment, and step 203 is further refined to include steps 301 to 304.

[0077] Step 301, determining target fields related to the question text according to keywords in the question text.

[0078] In this embodiment, the target field is a field stored in the production management database. Since a large amount of production management information is stored in the production management database, in order to accurately obtain the production management information related to the question text from the production management database, the target field related to the question text can be determined first.

[0079] In this embodiment, the question text may have one or more keywords, and the keywords of the question text may be classified. For example, "equipment No. 0001" and "equipment No. 0002" are equipment identification keywords, and "liquid level", "operating status" and "power consumption" are indicator keywords. In actual production management, the status data of a device may include multiple indicators. If the user only mentions a certain device but not a certain indicator when asking a question, the user may be concerned about all the indicators of the device. If the user mentions both a certain device and a certain indicator when asking a question, the user may be concerned about the indicator of the device. Therefore, the production management information related to the question text may be different due to different keywords in the question text. Furthermore, the target field used to query the production management information related to the question text in the database is also related to the keywords of the question text.

[0080] In this embodiment, when the keywords of the question text only include device identification keywords, the first preset field corresponding to the device identification keywords can be determined as the target field; when the keywords of the question text only include indicator keywords, the second preset field corresponding to the indicator keywords can be determined as the target field; when the keywords of the question text include both device identification keywords and indicator keywords, the field formed by connecting the first preset field corresponding to the device identification keywords and the second preset field corresponding to the indicator keywords using preset characters can be determined as the target field.

[0081] In this embodiment, the target field may be one field or multiple fields.

[0082] Optionally, step 301 is further refined to include steps 401 to 402 .

[0083] Step 401, in response to the fact that the keywords in the question text include device identification keywords and indicator keywords, a plurality of first mapping relationships are acquired, wherein the first mapping relationship is a first mapping relationship between a preset indicator of a preset device identification and a preset field.

[0084] In this embodiment, multiple first mapping relationships can be stored in the server device, and the first mapping relationship is a first mapping relationship between a preset indicator of a preset device identifier and a preset field. By way of example, the multiple first mapping relationships may include: the liquid level of device No. 0001 corresponds to the "0001 liquid level" field, the liquid level of device No. 0002 corresponds to the "0002 liquid level" field, and the power consumption of device No. 0002 corresponds to the "0002 power consumption" field.

[0085] Step 402: Determine, as a target field, a preset field having a first mapping relationship with an indicator keyword of a device identification keyword among a plurality of first preset mapping relationships.

[0086] In this embodiment, when the keywords of the question text include device identification keywords and indicator keywords, a preset field that has a first mapping relationship with the indicator keywords of the device identification keywords can be searched from multiple first preset mapping relationships, and then the preset field that has a first mapping relationship with the indicator keywords of the device identification keywords can be determined as the target field.

[0087] The reply content generation method provided in this embodiment can quickly and accurately determine the target field related to the question text based on the keywords of the question text, and then accurately obtain the production management information related to the question text from the production management database, by determining as the target field a preset field that has a first mapping relationship with the indicator-type keywords of the equipment identification-type keywords among multiple first preset mapping relationships.

[0088] Step 302, determining a target time associated with the question text based on the keywords and entities in the question text.

[0089] In this embodiment, in actual production management, in addition to the need to query real-time production management information, users also have the need to view historical production management information. Therefore, when obtaining production management information related to the question text from the production management database, the time range for obtaining the data also needs to be considered.

[0090] In this embodiment, if there are no time-referencing keywords and no time-related entities in the semantic analysis results, it is possible that the user did not mention any time-related information when asking the question. At this time, it can be considered that the user is concerned about real-time production management information. Therefore, the server-side device can determine that the target time related to the question text is the current time.

[0091] In this embodiment, the target time may be one time or multiple times; the target time may be a time point or a time period.

[0092] Step 303: query the production management database for the target value of the target field at the target time.

[0093] In this embodiment, after the target field and the target time are determined, the target value of the target field at the target time can be queried from the production management database by constructing an SQL query statement or calling the query program interface API of the production management database.

[0094] Step 304: determine the target value as production management information.

[0095] The reply content generation method provided in this embodiment determines the target field related to the question text based on the keywords of the question text; determines the target time related to the question text based on the keywords and entities of the question text; can analyze the objects and time range that the user is concerned about when asking the question, and then query the target value of the target field at the target time from the production management database; determines the target value as production management information, and can accurately obtain the production management information related to the question text from the production management database.

[0096] Optionally, step 302 is further refined to include steps 501 to 504 .

[0097] Step 501: In response to the semantic analysis result including time-referencing keywords, a target time is determined according to the time-referencing keywords.

[0098] In this embodiment, when the semantic analysis result includes time-referencing keywords, the server device can determine the target time according to the time-referencing keywords. Specifically, the server device can obtain the current time and determine the time referred to by the time-referencing keywords according to the current time. For example, if the current time is January 31, 2024, and the time-referencing keywords are today, the target time is January 31, 2024; if the time-referencing keywords are this month, the target time is January 2024; if the time-referencing keywords are last year, the target time is 2023.

[0099] Step 502: In response to the semantic analysis result not including time-referencing keywords but including time-related entities, determine the target time according to the time-related entities.

[0100] In this embodiment, when the semantic analysis result does not include time-referencing keywords and includes time-type entities, the server-side device determines the target time based on the time-type entity. Specifically, the server-side device can obtain the current time and determine the target time corresponding to the time-type entity based on the current time. For example, if the current time is 13:00 on January 31, 2024, if the time-type entity is "now", the target time is 13:00 on January 31, 2024; if the time-type entity is "last Saturday", the target time calculated based on the current time is January 27, 2024.

[0101] Step 503: In response to the semantic analysis result not including time-referencing keywords or time-related entities, a target time is determined according to a default query time configuration.

[0102] In this embodiment, when the semantic analysis result does not include time-referencing keywords or time-related entities, the server-side device can determine the target time according to the default query time configuration. Specifically, the server-side device can obtain the current time and the default query time configuration, and calculate the target time corresponding to the default query time configuration according to the current time. For example, the current time is 13:00 on January 31, 2024. If the default query time configuration is 2 hours ago, the target time is 11:00 on January 31, 2024; if the default query time configuration is 2 hours ago, the target time is from 11:00 to 13:00 on January 31, 2024.

[0103] The reply content generation method provided in this embodiment can accurately analyze the time range that the user is concerned about when asking questions through time-referring keywords and time-related entities, and then determine the exact target time, so as to accurately obtain production management information related to the question text from the production management database.

[0104] Figure 4 Schematic diagram of the process of generating the reply content provided in the embodiment of the present application Figure 3 .like Figure 4 As shown, in step 204, "generating reply content reference data according to semantic analysis results and production management information" is refined to include steps 601 to 602.

[0105] Step 601, in response to the target time being a time period and the target field having multiple target values ​​within the target time, generating at least one preset type of target value graph and / or target value table according to the multiple target values.

[0106] In this embodiment, when the target time is a time period, and the target field has multiple target values ​​within the target time, in order to enable the user to more intuitively obtain the answer content of the question, the server device can generate at least one preset type of target value graph and / or target value table according to the multiple target values. The preset type may include but is not limited to: a line graph, a curve graph, a pie chart, a bar chart, a radar chart, a table including only rows and columns, a table including rows and columns at multiple levels, a table with a title and a summary, a table with a separator line, and a table with color and format.

[0107] Step 602: determine the production management information and various types of target value graphs and / or target value tables as reference data for the reply content.

[0108] In this embodiment, various types of target value graphs and / or target value tables can intuitively display the changing trend and distribution of the target value of the target field within the target time, etc. Therefore, production management information and various types of target value graphs and / or target value tables can be determined as reply content reference data, and then sent to the client device, so that the client device displays various types of target value graphs and / or target value tables to the user.

[0109] The reply content generation method provided in this embodiment generates at least one preset type of target value graph and / or target value table. When the target time is a time period, detailed statistical graphs of historical production management information can be quickly generated in response to user questions, so that the trends and insights of the data are clear at a glance, thereby improving production efficiency and management efficiency in production management.

[0110] Optionally, the reply content reference data also includes display priorities of various types of target value graphs and / or target value tables; step 601 also includes step 701.

[0111] Step 701: Determine the display priority of each type of target value graph and / or target value table according to the display priority corresponding to each preset type.

[0112] In this embodiment, when the number of generated target value graphs and / or target values ​​is too large, each type of target value graph and / or target value table may have a display priority, and the display priority is the priority of the client device in displaying the graph or table. Therefore, the reply content reference data also includes the display priority of each type of target value graph and / or target value table.

[0113] In this embodiment, the server device may store preset display priorities corresponding to each type, and the preset display priorities corresponding to each type may be preset by the user. The server device may determine the preset display priorities corresponding to each type as the display priorities of the target value graphs and / or target value tables of each type.

[0114] The step of “sending the reply content reference data to the client device” in step 602 is refined to include step 702 .

[0115] Step 702, sends each type of target value graph and / or target value table to the client device in order of display priority from high to low, so that the client device determines each type of target value graph and / or target value table as reply content and outputs it in order of display priority from high to low.

[0116] In this embodiment, the server device can send reply content reference data to the client device in the form of a data stream, and the client device can give priority to displaying the target value graph or target value table received first. Therefore, the server device can send various types of target value graphs and / or target value tables to the client device in order of display priority from high to low, so that each type of target value graph and / or target value table can be output on the client device according to the corresponding display priority.

[0117] The reply content generation method provided in the present embodiment sends various types of target value graphs and / or target value tables to the client device in order of display priority from high to low, so that the client device determines and outputs various types of target value graphs and / or target value tables as reply content in order of display priority from high to low, and can intelligently recommend the type of charts to be displayed when too many charts are generated, and display the charts to the customer in sequence, so that the customer can more intuitively obtain the information in the charts and more intuitively understand the actual production management information.

[0118] Optionally, the semantic analysis result also includes the subject of the question text; in step 204, "generating reply content reference data based on the semantic analysis result and production management information" is refined to include steps 801 to 804.

[0119] Step 801 , in response to the subject of the question text including production problems, obtaining reasoning reference data related to the production problems from a production management knowledge base.

[0120] In this embodiment, the production management knowledge base is an internal knowledge base of the enterprise, and the production management knowledge base may include information such as common problems, important problems, and production reports in production management.

[0121] Step 802, input the question text, semantic analysis results, production management information and reasoning reference data into the trained artificial intelligence to generate the content model.

[0122] Step 803, using the trained artificial intelligence generated content model to generate and output a response to the question text.

[0123] Step 804, determining the response of the trained artificial intelligence generated content model to the question text as response content reference information.

[0124] In this embodiment, when the subject of the question text includes production problems, it is possible that the user has discovered the production problems at the production management site. Therefore, in order to better provide information and some solutions about the production problems, the inference reference data related to the production problems can be obtained from the production management knowledge base. The question text, semantic analysis results, production management information and inference reference data are input into the trained artificial intelligence generated content model, so that the trained artificial intelligence generated content model can refer to the production management information and inference reference data to generate a reply to the question text, and then the reply of the trained artificial intelligence generated content model to the question text is determined as the reply content reference information.

[0125] The reply content generation method provided in this embodiment, through intercommunication with the production management knowledge base, increases the data that can be referenced when the trained artificial intelligence content generation model generates replies, thereby ensuring that the replies generated by the trained artificial intelligence content generation model do not deviate from actual production management, thereby improving the accuracy of the replies.

[0126] Optionally, the content generation request also includes a user identifier; in step 204, “generating reply content reference data according to semantic analysis results and production management information” is refined to include steps 901 to 903.

[0127] Step 901, determining the inquiry intention of the question text according to the user identifier; the inquiry intention includes data inspection intention and cause analysis intention.

[0128] In this embodiment, the server device can generate different reply content according to different roles or identities of the questioning user when generating the reply content reference data. Specifically, the user identifier can correspond to different role or identity pools, and different roles or identity pools can correspond to different inquiry intentions.

[0129] The server device can determine the inquiry intention of the question text according to the role or identity pool corresponding to the user identifier. For example, when asking about the current status of a certain device, the inspection personnel may ask when checking the current status, while the product personnel may ask when the device fails.

[0130] Step 902, in response to the inquiry intention being a data inspection intention, determining the production management information as a reference data for the reply content.

[0131] In this embodiment, when the inquiry intention is for data inspection, as exemplified in step 901, it may be that the inspector is checking the current condition of the equipment. Therefore, only the production management information can be determined as the reference data for the reply content, and then, the client device can only reply to the user about the current condition of the equipment.

[0132] Step 903, in response to the inquiry intention being the cause analysis intention, the production management information is input into the trained cause analysis model, the trained cause analysis model is used to analyze the cause of the production management information, and the production management information and its cause are determined as reference data for the reply content.

[0133] In this embodiment, when the inquiry intention is the intention of cause analysis, as in the example in step 901, it may be that the product personnel find that the device has a fault and inquire about the current situation of the device. Therefore, the production management information can be input into the trained cause analysis model, and the trained cause analysis model is used to analyze the cause of the production management information. Then, the production management information and its cause are determined as the reference data of the reply content, so that the client device can reply to the user about the current situation of the device and the possible cause of the fault.

[0134] The reply content generation method provided in this embodiment determines the inquiry intention of the question text through the user identifier, and when generating reply content reference data, generates different reply content reference data according to different inquiry intentions, so that the generated reply content reference data can be applicable to different scenarios and different personnel in production management, thereby improving the production efficiency and management efficiency in each production management link.

[0135] Optionally, in the reply content generation method provided by the present application, the server device can also infer the problems that may arise in current or future production management based on the user's historical questions and the reference information of the reply content to the historical questions, and send alarm information to the client device when the possible problems are inferred, so that the client device outputs the alarm information. In turn, the user can obtain the alarm information, prevent or deal with the problems that may arise in current or future production management, and ensure production safety.

[0136] Embodiment 4

[0137] An embodiment of the present application also provides a reply content generation device, which is located in a server-side device and includes: a receiving module, an analyzing module, an acquiring module and a generating module.

[0138] The receiving module is used to receive a content generation request sent by a client device; the content generation request includes a question text.

[0139] The analysis module is used to perform semantic analysis on the question text to obtain the semantic analysis results of the question text; the semantic analysis results include keywords and entities of the question text.

[0140] The acquisition module is used to acquire production management information related to the question text based on the keywords and entities in the question text.

[0141] A generation module is used to generate reply content reference data based on semantic analysis results and production management information and send it to the client device. The reply content reference data includes production management information so that the client device generates and outputs reply content based on the reply content reference data.

[0142] Optionally, the acquisition module is specifically used to: determine a target field related to the question text according to keywords of the question text;

[0143] Determine the target time related to the question text based on the keywords and entities in the question text;

[0144] Query the target value of the target field at the target time from the production management database;

[0145] The target value is determined as production management information.

[0146] Optionally, the acquisition module is further used to:

[0147] In response to the keyword of the question text including the device identification keyword and the indicator keyword, a plurality of first mapping relationships are acquired, wherein the first mapping relationship is a first mapping relationship between a preset indicator of a preset device identification and a preset field;

[0148] A preset field that has a first mapping relationship with an indicator-type keyword of a device identification-type keyword among the plurality of first preset mapping relationships is determined as a target field.

[0149] Optionally, the acquisition module is further used to:

[0150] In response to the semantic analysis result including time-referencing keywords, determining a target time according to the time-referencing keywords;

[0151] In response to the semantic analysis result not including the time reference keyword but including the time entity, determining the target time according to the time entity;

[0152] In response to the semantic analysis result not including a time-referencing keyword or a time-related entity, a target time is determined according to a default query time configuration.

[0153] Optionally, the generation module is specifically used for:

[0154] In response to the target time being a time period, and the target field having a plurality of target values ​​within the target time, generating at least one preset type of target value graph and / or target value table according to the plurality of target values;

[0155] Production management information and various types of target value graphs and / or target value tables are determined as reference data for the reply content.

[0156] Optionally, the reply content reference data also includes display priorities of various types of target value graphs and / or target value tables; the generating module is further specifically used for:

[0157] Determine the display priority of each type of target value graph and / or target value table according to the display priority corresponding to each preset type;

[0158] Each type of target value graph and / or target value table is sent to the client device in order of display priority from high to low, so that the client device determines and outputs each type of target value graph and / or target value table as reply content in order of display priority from high to low.

[0159] Optionally, the semantic analysis result also includes the subject of the question text; the generation module is specifically used for:

[0160] In response to the topic of the question text including production problems, obtaining reasoning reference data related to the production problems from a production management knowledge base;

[0161] Input question text, semantic analysis results, production management information, and reasoning reference data into the trained AI-generated content model;

[0162] Using a trained artificial intelligence content generation model to generate and output a response to the question text;

[0163] The response to the question text generated by the trained artificial intelligence content generation model is determined as response content reference information.

[0164] Optionally, the content generation request further includes a user identifier; the generation module is specifically configured to:

[0165] Determine the inquiry intention of the question text according to the user identifier; the inquiry intention includes data inspection intention and cause analysis intention;

[0166] In response to the inquiry intention being a data inspection intention, determining the production management information as a reference data for the reply content;

[0167] In response to the inquiry intention being the cause analysis intention, the production management information is input into the trained cause analysis model, the trained cause analysis model is used to analyze the cause of the production management information, and the production management information and its cause are determined as the reference data for the reply content.

[0168] The reply content generation device provided in this embodiment can execute the reply content generation method provided in any of the above embodiments. The specific implementation method and principle are similar and will not be repeated here.

[0169] Embodiment 5

[0170] Figure 5This is a schematic diagram of the structure of the server device provided in Example 5 of this application. Figure 5 As shown, the server device 50 provided in this embodiment includes: a processor 52 and a memory 51 communicatively connected to the processor 52 .

[0171] The memory 51 stores computer-executable instructions.

[0172] The processor 52 executes the computer execution instructions stored in the memory 51 to implement the reply content generation method provided by any of the above embodiments. The specific implementation method and principle are similar and will not be repeated here.

[0173] The memory 51 and the processor 52 can be connected to each other through a bus. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0174] In this embodiment, the memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk, etc.

[0175] In this embodiment, the server device can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to execute the reply content generation method provided in any of the above embodiments.

[0176] The embodiment of the present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the reply content generation method provided in any of the above embodiments. The specific implementation method and principle are similar and will not be repeated here. Exemplarily, the computer-readable storage medium can be a read-only memory, a random access memory, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0177] An embodiment of the present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the reply content generation method provided by any of the above embodiments. The specific implementation method and principle are similar and will not be repeated here.

[0178] It should be understood that the above method embodiments are only illustrative, and the method of the present application can also be implemented in other ways. For example, the division of modules in the above embodiments is only a logical function division, and there may be other division methods in actual implementation. For example, multiple modules can be combined, or can be integrated into another system, or some features can be ignored or not executed.

[0179] In addition, unless otherwise specified, the functional modules in the various embodiments of the present application may be integrated into one module, or each module may exist physically separately, or two or more modules may be integrated together. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software program modules.

[0180] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present application.

[0181] It should be further noted that, although the various steps in the flow chart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flow chart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0182] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0183] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for generating reply content, characterized in that: Applied to a server device, the method comprises: Receiving a content generation request sent by a client device; the content generation request includes a question text; Performing semantic analysis on the question text to obtain a semantic analysis result of the question text; the semantic analysis result includes keywords and entities of the question text; Acquire production management information related to the question text according to keywords and entities of the question text; the production management information refers to data generated by actual production management activities; the production management information is stored in a production management database in real time; Generate reply content reference data according to the semantic analysis result and the production management information and send it to the client device, wherein the reply content reference data includes the production management information, so that the client device generates reply content according to the reply content reference data and outputs it; The step of obtaining production management information related to the question text according to the keywords and entities of the question text includes: Determine the target fields related to the question text based on the keywords of the question text; Determine the target time related to the question text based on the keywords and entities in the question text; Querying the target value of the target field at the target time from the production management database; determining the target value as the production management information; When the target time is a time period, and the target field has multiple target values ​​within the target time, determining the target field related to the question text according to the keywords of the question text includes: In response to the keyword of the question text including a device identification keyword and an indicator keyword, a plurality of first mapping relationships are acquired, wherein the first mapping relationship is a first mapping relationship between a preset indicator of a preset device identification and a preset field; Determine, as a target field, a preset field having a first mapping relationship with the indicator keyword of the device identification keyword among the plurality of first preset mapping relationships; When the content generation request further includes a user identifier, generating the reply content reference data according to the semantic analysis result and the production management information includes: Determine the inquiry intention of the question text according to the user identifier; the inquiry intention includes data inspection intention and cause analysis intention; In response to the inquiry intention being a data inspection intention, determining the production management information as the reply content reference data; In response to the inquiry intention being a cause analysis intention, inputting the production management information into a trained cause analysis model, using the trained cause analysis model to analyze the generation cause of the production management information, and determining the production management information and the generation cause thereof as the reference data for the reply content; When the semantic analysis result also includes the subject of the question text, generating the answer content reference data according to the semantic analysis result and the production management information includes: In response to the subject of the question text including production problems, acquiring reasoning reference data related to the production problems from a production management knowledge base; Inputting the question text, the semantic analysis result, the production management information and the reasoning reference data into a trained artificial intelligence generated content model; Using a trained artificial intelligence content generation model to generate and output a response to the question text; The response of the trained artificial intelligence generated content model to the question text is determined as the response content reference information.

2. The method for generating reply content according to claim 1, characterized in that: Determining the target time related to the question text according to the keywords and entities of the question text includes: In response to the semantic analysis result including time-referencing keywords, determining the target time according to the time-referencing keywords; In response to the semantic analysis result not including time-referencing keywords but including time-related entities, determining the target time according to the time-related entities; In response to the semantic analysis result not including time-referencing keywords or time-related entities, the target time is determined according to a default query time configuration.

3. The method for generating reply content according to claim 1, characterized in that: The generating of the reply content reference data according to the semantic analysis result and the production management information comprises: In response to the target time being a time period, and the target field having a plurality of target values ​​within the target time, generating at least one preset type of target value graph and / or target value table according to the plurality of target values; The production management information and various types of target value graphs and / or target value tables are determined as the reply content reference data.

4. The method for generating reply content according to claim 3, characterized in that: The reply content reference data also includes display priorities of various types of target value graphs and / or target value tables; After generating at least one preset type of target value graph and / or target value table according to the plurality of target values, the method further includes: Determine the display priority of each type of target value graph and / or target value table according to the display priority corresponding to each preset type; Sending the reply content reference data to the client device includes: Each type of target value graph and / or target value table is sent to the client device in order of display priority from high to low, so that the client device determines and outputs each type of target value graph and / or target value table as reply content in order of display priority from high to low.

5. A server device, characterized in that: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the reply content generating method according to any one of claims 1 to 4.

6. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the reply content generation method as described in any one of claims 1 to 4.

Images