Interactive processing method, device, computer equipment, storage medium and computer program product in the field of intelligent manufacturing

Abstract

The present application relates to an interactive processing method, device, computer equipment, storage medium and computer program product in the field of intelligent manufacturing. The method includes: in an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered in an interactive interface, obtaining a manufacturing process inquiry message; based on historical inquiry interaction messages, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information; from a data analysis platform in the field of intelligent manufacturing, searching for manufacturing indicators that meet the intent information and match the entity information; according to the intent information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator, and displaying the obtained visual display data as a response result of the manufacturing process inquiry message in the interactive interface. The above method simplifies the interactive process in the field of intelligent manufacturing, shortens the time required for interactive operations, and effectively improves the interactive efficiency in the field of intelligent manufacturing.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to an interactive processing method, apparatus, computer equipment, storage medium and computer program product in the field of intelligent manufacturing. Background Art

[0002] With the continuous development of various technical fields, the industry is becoming more and more professional. In a specific industry, there are generally various professional terms and industry-specific languages.

[0003] Taking the manufacturing industry as an example, the manufacturing production line will generate a large amount of raw data, which needs to be professionally analyzed to understand its specific meaning. Data query for specific fields such as manufacturing usually requires a series of operation steps such as user login to a specific system, data query, and data analysis to obtain the data results required by the user, which leads to low interaction efficiency in interactive scenarios such as data query in the field of intelligent manufacturing. Summary of the invention

[0004] Based on this, it is necessary to provide an interactive processing method, device, computer equipment, computer-readable storage medium and computer program product in the field of intelligent manufacturing that can improve the interaction efficiency in response to the above-mentioned technical problems.

[0005] In a first aspect, the present application provides an interactive processing method in the field of intelligent manufacturing. The method comprises:

[0006] In an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained;

[0007] Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information;

[0008] Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing indicators that conform to the intention information and match the entity information;

[0009] According to the intention information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data representing the manufacturing process characteristics;

[0010] The visual display data is used as a response result of the manufacturing process inquiry message, and the visual display data is displayed in the interactive interface.

[0011] In some of the embodiments, when the manufacturing process inquiry message is a voice message, the method further includes:

[0012] Based on a non-autoregressive end-to-end speech recognition model, feature extraction is performed on the voice message to obtain a speech feature vector; the non-autoregressive end-to-end speech recognition model is trained based on sample voice messages, and the sample voice messages contain professional terms in the field of intelligent manufacturing;

[0013] By using the self-attention mechanism of the non-autoregressive end-to-end speech recognition model, the speech feature vector is sequentially feature encoded and context modeled to obtain an encoded feature;

[0014] Perform feature integration based on the coding features to obtain integrated features;

[0015] Performing text transcription on the integrated features to obtain a message text representing the voice message;

[0016] The method of performing intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information includes:

[0017] Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, the message text is subjected to intent recognition and entity extraction to obtain intent information and entity information.

[0018] In some embodiments, the performing of intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information includes:

[0019] Based on a natural language understanding model, performing intent recognition and entity extraction on the message text to obtain original intent information and original entity words;

[0020] Acquire historical status information corresponding to the historical query interaction message corresponding to the manufacturing process query message, and construct current status information based on the original intent information and the original entity words;

[0021] Performing intention prediction based on the historical state information and the current state information to obtain intention information matching the manufacturing process inquiry message;

[0022] The original entity words are standardized to obtain entity information matching the manufacturing process inquiry message.

[0023] In some embodiments, the standardization of the original entity words to obtain entity information matching the manufacturing process query message includes:

[0024] Obtain standardized expressions of entities in the field of intelligent manufacturing;

[0025] According to the entity standardized expression, the original entity word is entity mapped to obtain standardized entity information.

[0026] In some embodiments, the manufacturing indicator is a manufacturing indicator combination including a manufacturing process intention indicator and a manufacturing process entity indicator; searching the data analysis platform in the field of intelligent manufacturing for a manufacturing indicator that meets the intention information and matches the entity information includes:

[0027] Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing process intention indicators that match the intention information and manufacturing process entity indicators that match the entity information;

[0028] The manufacturing process intention indicator is combined with the manufacturing process entity indicator to obtain a manufacturing indicator.

[0029] In some of the embodiments, the data analysis platform provides a query interface, a data analysis interface, and a visualization interface;

[0030] The step of performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator according to the intention information to obtain visual display data representing the manufacturing process characteristics includes:

[0031] Calling a query interface matching the manufacturing index to query manufacturing process data corresponding to the manufacturing index;

[0032] Calling a data analysis interface that matches the intention information, performing a manufacturing process characteristic analysis on the manufacturing process data, and obtaining a manufacturing process characteristic analysis result;

[0033] A visualization interface matching the data analysis interface is called to perform data conversion on the manufacturing process characteristic analysis result to obtain visualization display data.

[0034] In a second aspect, the present application also provides an interactive processing device in the field of intelligent manufacturing. The device comprises:

[0035] An input processing module is used to obtain an input manufacturing process inquiry message in response to an inquiry message input operation triggered on an interactive interface in an interactive scenario in the field of intelligent manufacturing;

[0036] An intent recognition and entity extraction module, configured to perform intent recognition and entity extraction on the manufacturing process inquiry message based on a historical inquiry interaction message corresponding to the manufacturing process inquiry message, to obtain intent information and entity information;

[0037] An indicator search module, used to search for manufacturing indicators that meet the intention information and match the entity information from the data analysis platform in the field of intelligent manufacturing;

[0038] A manufacturing process characteristic analysis module, used to perform manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator according to the intention information, and obtain visual display data representing the manufacturing process characteristics;

[0039] A visualization display module is used to use the visualization display data as a response result of the manufacturing process inquiry message and display the visualization display data in the interactive interface.

[0040] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0041] In an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained;

[0042] Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information;

[0043] Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing indicators that conform to the intention information and match the entity information;

[0044] According to the intention information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data representing the manufacturing process characteristics;

[0045] The visual display data is used as a response result of the manufacturing process inquiry message, and the visual display data is displayed in the interactive interface.

[0046] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0047] In an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained;

[0048] Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information;

[0049] Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing indicators that conform to the intention information and match the entity information;

[0050] According to the intention information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data representing the manufacturing process characteristics;

[0051] The visual display data is used as a response result of the manufacturing process inquiry message, and the visual display data is displayed in the interactive interface.

[0052] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0053] In an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained;

[0054] Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information;

[0055] Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing indicators that conform to the intention information and match the entity information;

[0056] According to the intention information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data representing the manufacturing process characteristics;

[0057] The visual display data is used as a response result of the manufacturing process inquiry message, and the visual display data is displayed in the interactive interface.

[0058] The above-mentioned interactive processing method, device, computer equipment, storage medium and computer program product in the field of intelligent manufacturing, in an interactive scenario in the field of intelligent manufacturing, obtains an input manufacturing process inquiry message by responding to an inquiry message input operation triggered on an interactive interface, and then performs intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, so as to accurately understand the specific intention information of the user's interactive operation and the entity information related to the interaction, so as to facilitate accurate acquisition of matching manufacturing process data for data analysis. In the data acquisition process, manufacturing indicators that meet the intention information and match the entity information are found from the data analysis platform in the field of intelligent manufacturing. It can realize accurate search of manufacturing process data, and then analyze the manufacturing process characteristics of the manufacturing process data corresponding to the manufacturing indicators according to the intention information, and obtain visual display data that characterizes the manufacturing process characteristics. Finally, the visual display data is used as the response result of the manufacturing process query message, and the visual display data is displayed in the interactive interface. Only simple interactive operations are required to obtain data interaction results in the field of intelligent manufacturing in a visual way. It can not only obtain manufacturing process data through intent recognition and information extraction, perform manufacturing process characteristics analysis to obtain intuitively displayed data results, but also simplifies the interaction process and shortens the time required for interactive operations, thereby effectively improving the interaction efficiency in the field of intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 An application environment diagram of an interactive processing method in the field of intelligent manufacturing in one embodiment;

[0060] Figure 2 A flowchart of an interactive processing method in the field of intelligent manufacturing in one embodiment;

[0061] Figure 3 A schematic flow chart of an interactive processing method in the field of intelligent manufacturing in another embodiment;

[0062] Figure 4 A schematic diagram of a text transcription process of a voice message in one embodiment;

[0063] Figure 5 A flowchart of an intent recognition and entity extraction process in one embodiment;

[0064] Figure 6 is a schematic diagram of the interaction between a computer device and a data analysis platform in one embodiment;

[0065] Figure 7 It is a structural block diagram of an interactive processing device in the field of intelligent manufacturing in one embodiment;

[0066] Figure 8is an internal structure diagram of a computer device in one embodiment;

[0067] Fig. 9 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0068] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0069] The interactive processing method in the field of intelligent manufacturing provided by the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102, the server 104, and the data analysis platform 106 communicate through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other servers. In the interactive scenario in the field of intelligent manufacturing, the server 104 responds to the query message input operation triggered by the user in the interactive interface of the terminal 102, and determines the input manufacturing process query message; based on the historical query interaction message corresponding to the manufacturing process query message, the manufacturing process query message is intentionally identified and the entity is extracted to obtain the intent information and the entity information; from the data analysis platform 106 in the field of intelligent manufacturing, the manufacturing indicators that meet the intent information and match the entity information are searched; according to the intent information, the manufacturing process data corresponding to the manufacturing indicators are analyzed for the manufacturing process characteristics, and the visual display data that characterizes the manufacturing process characteristics is obtained; the visual display data is used as the response result of the manufacturing process query message, and the visual display data is displayed in the interactive interface of the terminal 102.

[0070] The terminal 102 may be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, IoT devices, and portable wearable devices, and the portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0071] In one embodiment, Figure 2As shown, an interactive processing method in the field of intelligent manufacturing is provided, and the method is applied to a computer device as an example for explanation. The computer device can specifically be a server with interactive functions, and can complete the interaction and data analysis process based on the server. The computer device can also be a system composed of a terminal and a server. The terminal provides the function of interacting with the user, and reports the signal corresponding to the interactive operation to the server. The server performs background data analysis and processing, and feeds back the final result obtained by the analysis to the terminal to display it to the user who triggered the interactive operation. In addition, the computer device can also be a terminal with strong computing power, which can not only respond to the user's interactive operation at the terminal, but also obtain the final result through data analysis and display it directly on the terminal. Specifically, the computer device implements the interactive processing method in the field of intelligent manufacturing, which specifically includes the following steps:

[0072] Step 202, in an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained.

[0073] Among them, the field of intelligent manufacturing is based on the deep integration of new-generation information technology and advanced manufacturing technology, and integrates artificial intelligence technology into part or all of the processes in the design, production, management, and service of manufacturing activities. It is a new production method with the characteristics of self-perception, self-decision-making, self-execution, self-adaptation, and self-learning.

[0074] Interaction scenarios in the field of intelligent manufacturing refer to scenarios that can provide interactive interfaces to achieve human-computer interaction in a specific industry environment. For example, in the manufacturing industry, computer equipment can provide interactive interfaces for manufacturing scenarios, and users can trigger interactive operations on the interactive interfaces. Computer equipment analyzes the corresponding manufacturing process characteristics in response to the interactive operations and feeds back the interactive results to users, thereby achieving human-computer interaction.

[0075] The interactive interface is a display page that provides the user with a human-computer interaction interface. Specifically, the interactive interface can display the specific operation mode of human-computer interaction, which is convenient for users in the field of intelligent manufacturing or non-intelligent manufacturing to understand the specific human-computer interaction mode and trigger the interactive operation. In addition, the interactive interface can also be used to display the results of data analysis after responding to the interactive operation, so that the user can trigger the interactive operation on the same interactive interface and view the response result of the interactive operation.

[0076] The inquiry message input operation is a specific implementation method of human-computer interaction. The inquiry message input operation can specifically include text input operation and voice input operation. The computer device can provide input interfaces for multiple different inquiry message input methods. Each input interface can be integrated into an input control. The user can input the manufacturing process inquiry message by selecting the corresponding input control.

[0077] The manufacturing process inquiry message is an input message that the user inputs through the interactive interface and can be used to query the manufacturing process data to directly obtain a visual display result. The message types of the manufacturing process inquiry message can specifically include voice messages and text messages. For example, in the interactive interface, text input controls and voice input controls can be displayed to facilitate users to select the desired interactive method to input the manufacturing process inquiry message. In one embodiment, after the user clicks the text input control, the text message can be input through an input device such as a physical keyboard or a virtual keyboard. In another embodiment, after the user clicks the voice input control, the computer device can evoke a voice collection device such as a microphone to collect the user's voice and obtain a voice message.

[0078] In a specific embodiment, in an interactive scenario in the field of intelligent manufacturing, a text input control and a voice input control may be displayed on the interactive interface for the user to select an interactive mode. If the user selects the text input control on the interactive interface to invoke the input panel, the computer device will respond to the text input operation triggered on the input panel and obtain a manufacturing process inquiry message with a message category of text. If the user selects the voice input control on the interactive interface and triggers the voice input operation, the computer device will respond to the voice input operation and obtain a manufacturing process inquiry message with a message category of voice.

[0079] Step 204 , based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, perform intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information.

[0080] Among them, the historical inquiry interaction message is the historical message corresponding to the interaction that occurred before the manufacturing process inquiry message was input. The computer device can search through historical data to obtain the historical inquiry interaction message corresponding to the manufacturing process inquiry message. Specifically, the historical inquiry interaction message of the manufacturing process inquiry message can be the historical message corresponding to the previous interaction or the previous N interactions of the same logged-in user, or it can be the historical message corresponding to the interaction that occurred within a preset historical time period. The message type of the historical inquiry interaction message can be the same as the manufacturing process inquiry message, or it can be different from the manufacturing process inquiry message. For example, the manufacturing process inquiry message is a text message, and the historical inquiry interaction message can be a text message or a voice message. Specifically, the acquisition strategy of the historical inquiry interaction message of the manufacturing process inquiry message can be set according to the needs of the specific application scenario, and is not limited here.

[0081] Intent recognition is a process used to determine the intention or purpose expressed in the sentence input by the user. The result obtained through the intention recognition process is the intention information. Simply put, intent recognition is to understand the user's words semantically in order to better answer the user's questions or provide related services. In multiple consecutive conversations, according to the user's usual language expression habits, the manufacturing process inquiry message that may be input is not a complete sentence. Therefore, it is necessary to perform intent recognition on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, so as to accurately understand the true intention of the manufacturing process inquiry message input by the user this time. Entity recognition is a process used to identify entity words with specific meanings in the manufacturing process inquiry message. The result obtained through entity recognition is entity information.

[0082] In some specific embodiments, taking the manufacturing process inquiry message as a text message as an example, after the computer device obtains the historical inquiry interaction message corresponding to the text message, it can directly use the text in the text message as the processing object, combine the text corresponding to the historical inquiry interaction message to perform intent recognition to obtain intent information, and perform entity recognition processing on the text to obtain entity information.

[0083] In other specific embodiments, taking the manufacturing process inquiry message as a voice message as an example, after obtaining the historical inquiry interaction message corresponding to the text message, the computer device can first transcribe the voice message into text, and based on the transcribed text, perform intent recognition in combination with the text corresponding to the historical inquiry interaction message to obtain intent information, and perform entity recognition processing on the text to obtain entity information.

[0084] Step 206, searching for manufacturing indicators that meet the intent information and match the entity information from the data analysis platform in the field of intelligent manufacturing.

[0085] Among them, the data analysis platform is a manufacturing process characteristic analysis platform specifically used to manage and analyze data in the field of intelligent manufacturing. Taking the manufacturing industry as an example, the data analysis platform of the manufacturing industry can be built based on the manufacturing execution system (MES) or other industrial systems to collect and analyze data related to the production line. Enterprises use the data analysis platform of the manufacturing industry to help monitor the performance of the production line, optimize the production process, and improve overall efficiency. Computer equipment is connected to the data analysis platform through the network, and can realize data interaction with the data analysis platform, such as data query, data acquisition, etc.

[0086] Data indicators are the measurement standards used to match data of different dimensions in the field of intelligent manufacturing. Different intent information corresponds to different data indicators, and different entity information may also correspond to different data indicators. Manufacturing indicators can be a set of indicators of 2 or more manufacturing data indicators. In some embodiments, the manufacturing indicators include at least one data indicator corresponding to intent information and one data indicator corresponding to entity information. The data analysis platform can find the manufacturing process data that meets the intent information and matches the entity information based on the manufacturing indicators, and feed the manufacturing process data back to the computer device.

[0087] In some embodiments, after the computer device performs intent recognition and entity extraction on the manufacturing process inquiry message and obtains intent information and entity information, the intent information and entity information can be sent to the data analysis platform, and the data analysis platform searches for manufacturing indicators that meet the intent information and match the entity information, and then performs the manufacturing process data according to the manufacturing indicators. If the computer device itself stores various data indicators of the data analysis platform, the computer device can also obtain manufacturing indicators that meet the intent information and match the entity information through data indicator matching processing, and send the manufacturing indicators to the data analysis platform, search for the manufacturing indicators from the data analysis platform, and then obtain the manufacturing process data corresponding to the manufacturing indicators.

[0088] Step 208 , performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicators according to the intention information, and obtaining visual display data representing the manufacturing process characteristics.

[0089] Among them, manufacturing process data refers to various data generated in the manufacturing process. Manufacturing process data covers all aspects from design to production execution, including but not limited to information on production tasks, personnel, equipment, materials, process quality, safety and environmental protection. For example, manufacturing process data may include at least one of the following types of data: production task data, workshop personnel data, production equipment data, production material data, process quality data, production safety and environmental protection data, etc.

[0090] Among them, production task data includes production plan and execution data, work order information (such as production start time, intermediate pause time, production end time), labor time consumption information, and the completion volume of the production plan. These data support statistical analysis of key production management indicators such as beat, production output, planned delivery completion accuracy, and order delivery cycle. Workshop personnel data include information such as who has produced what product model at which workstation, start time, and completion time records. This type of data helps locate the cause when quality problems occur, and can also statistically analyze the work efficiency of personnel. Production equipment data involves equipment status and tooling data, equipment power on and off status, alarm data, operating pressure, speed, temperature and other real-time data. These data support statistical equipment downtime, calculation of equipment utilization efficiency, and problem solving and optimization of equipment parameter settings based on alarm data analysis. Production material data includes the name, quantity, batch, and time of acceptance and use of the material at the workstation. This type of data supports precise control of material demand and inventory levels to ensure quality traceability. Process quality data covers the quality management and control data of the entire process from raw material inspection to product delivery, such as quality accidents, customer complaint handling, etc. Production safety and environmental protection data focuses on safety and environmental protection data in the production process to ensure the compliance of production activities.

[0091] Manufacturing process characteristic analysis is a data processing method that optimizes the production process and product quality by analyzing and evaluating the key characteristics of the manufacturing system. The specific processing method of manufacturing process characteristic analysis can be determined based on the intent information, such as statistics on the production volume of a day, and then calculating the product failure rate of a certain process procedure. The computer device can first obtain the manufacturing process data corresponding to the manufacturing indicator, and then perform manufacturing process characteristic analysis on the manufacturing process data through the manufacturing process characteristic analysis method indicated by the intent data to obtain the manufacturing process characteristic analysis results.

[0092] Visual display data refers to data that can intuitively display the trend of indicators and the presentation form of key information. For different categories of manufacturing process characteristic analysis results, corresponding visualization methods can be pre-configured in computer equipment to display data charts matching different categories of data sampling or directly display the manufacturing process characteristic analysis results. Specifically, the visualization display method can include bar charts, line charts, pie charts, scatter plots and other charts, or direct numerical display and text description.

[0093] In some embodiments, a series of pre-developed general data presentation components may be configured in the computer device. Each data presentation component can receive data input of different indicators, automatically select a suitable presentation method according to the structure and configuration of the data, and automatically generate a corresponding chart or text description. The general component has the characteristics of flexible design and high maintainability, so that it can adapt to different data structures and presentation requirements. Through the above processing method, the processing efficiency of converting the results of manufacturing process characteristic analysis into visual display data can be further improved. Furthermore, in order to ensure that users can easily understand and use the data presentation components, the computer device can provide clear legends, tooltips, and operation guides for the data presentation components, and by pre-configuring the display adaptation strategy, ensure that the charts can be well displayed on different devices and screen sizes.

[0094] Step 210: Use the visual display data as a response result of the manufacturing process inquiry message, and display the visual display data in an interactive interface.

[0095] The response result of the manufacturing process inquiry message is the interactive data used as the answer to the manufacturing process inquiry message. For example, after the user inputs the manufacturing process inquiry message through the interactive interface, the visual display result obtained after data analysis and processing will be used as the response result of the manufacturing process inquiry message and displayed in the interactive interface, so that the user can obtain the data display result by inputting a simple message.

[0096] In some embodiments, in the interactive interface of the computer device, the manufacturing process inquiry message and the visual display data can be displayed in a dialog manner. For example, the manufacturing process inquiry message input by the user is displayed in the interactive interface through a message bubble pointing in a first direction, and the visual display data obtained by data analysis is displayed through a message bubble pointing in a second direction. The first direction and the second direction are different directions. For example, the first direction is the direction pointing to the right side of the interactive interface, and the second direction is the direction pointing to the left side of the interactive interface, so as to match the interaction habits of general interaction scenarios.

[0097] In other embodiments, in the interactive interface of a computer device, manufacturing process inquiry messages and visual display data may be displayed separately in chronological order. For example, when a user inputs a manufacturing process inquiry message, a target display is displayed. After the manufacturing process characteristics are analyzed based on the manufacturing process inquiry message to obtain visual display data, the manufacturing process inquiry message is cancelled and the visual display data is displayed, so that there is sufficient space in the display page to display the visual display data.

[0098] In addition, for the visual display data displayed on the display page, the user can interact with the visual display data by clicking, filtering or entering data, and the computer device can respond to the operation and update the chart in real time to reflect the latest data status.

[0099] The above-mentioned interactive processing method in the field of intelligent manufacturing, in the interactive scenario in the field of intelligent manufacturing, obtains the input manufacturing process inquiry message by responding to the inquiry message input operation triggered on the interactive interface, and then performs intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, so as to accurately understand the specific intention information of the user's interactive operation and the entity information related to the interaction, so as to facilitate the accurate acquisition of matching manufacturing process data for data analysis. In the process of data acquisition, by searching for manufacturing indicators that meet the intention information and match the entity information from the data analysis platform in the field of intelligent manufacturing, it is possible to realize the manufacturing process. The manufacturing process data can be accurately searched, and then the manufacturing process characteristics are analyzed for the manufacturing process data corresponding to the manufacturing indicators according to the intent information, so as to obtain the visual display data representing the manufacturing process characteristics. Finally, the visual display data is used as the response result of the manufacturing process query message, and the visual display data is displayed in the interactive interface. Only simple interactive operations are required to obtain the data interaction results in the field of intelligent manufacturing in a visual way. Not only can the manufacturing process data be obtained through intent recognition and information extraction for analysis to obtain intuitively displayed data results, but also the interaction process is simplified and the time required for interactive operations is shortened, thereby effectively improving the interaction efficiency in the field of intelligent manufacturing.

[0100] In some embodiments, such as Figure 3 As shown, in the case where the manufacturing process inquiry message is a voice message, the interactive processing method in the field of intelligent manufacturing further includes steps 302 to 308:

[0101] Step 302: extract features from the voice message based on a non-autoregressive end-to-end speech recognition model to obtain a speech feature vector; the non-autoregressive end-to-end speech recognition model is trained based on sample voice messages, and the sample voice messages contain professional terms in the field of intelligent manufacturing.

[0102] Step 304, through the self-attention mechanism of the non-autoregressive end-to-end speech recognition model, the speech feature vector is sequentially feature encoded and context modeled to obtain the encoded feature.

[0103] Step 306: perform feature integration based on the coding features to obtain integrated features.

[0104] Step 308: transcribe the integrated features to obtain a message text representing the voice message.

[0105] Further, based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, the manufacturing process inquiry message is subjected to intent recognition and entity extraction to obtain intent information and entity information, that is, step 204 includes step 310:

[0106] Step 310 , based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, perform intent recognition and entity extraction on the message text to obtain intent information and entity information.

[0107] Among them, the non-autoregressive end-to-end speech recognition model can specifically be a Paraformer model. For ease of description, the Paraformer model is used for illustration in the following embodiments. The Paraformer model adopts a non-autoregressive structure, which can output all target texts in a single decoding, thereby significantly improving the computational efficiency. Compared with the traditional autoregressive model, the Paraformer has a significant advantage in decoding speed because it avoids the process of generating target texts one by one, reducing the computational complexity.

[0108] Specifically, Paraformer uses a deeper model structure, including multiple encoder and decoder layers, as well as a specific attention mechanism, which helps improve the model's expressiveness. Through the non-autoregressive structural design, Paraformer can generate all target texts in a single decoding, greatly improving the decoding speed.

[0109] During the training process, taking the manufacturing industry in the field of intelligent manufacturing as an example, the Paraformer model was specially adjusted and trained in order to optimize for the unique data set in the manufacturing production process. Specifically, the Paraformer model was fine-tuned using a large number of on-site recordings of the manufacturing industry to improve the Paraformer model's ability to recognize specific industry terms, technical language, and production environment noise. And by introducing manufacturing-related voice data during the training process, the Paraformer model can better adapt to the unique speech patterns and contexts in this field. Taking into account the various noise interferences that may exist at the production site, the Paraformer model was trained for model reinforcement to improve its recognition accuracy in noisy environments.

[0110] In a specific embodiment, for voice messages, it is necessary to transcribe the voice messages into message text, so as to facilitate the subsequent intent recognition and entity extraction processing. Specifically, taking the voice transcription processing through the Paraformer model as an example, Figure 4As shown, the processing process of the Paraformer model specifically includes speech input 402, feature extraction 404, feature encoding 406, context modeling 408, feature integration 410, text transcription 412 and text output 414.

[0111] Specifically, Paraformer performs in-depth feature extraction on the input speech data to obtain speech feature vectors, converting the sound wave data in the original speech data into feature vectors that can represent the essence of speech. These feature vectors capture the spectrum, pitch, energy and other potential acoustic features of speech. The extracted feature vectors are sent to the Paraformer model for encoding to obtain the original encoding features. The Paraformer model utilizes the self-attention mechanism in the Transformer model, which can effectively capture the dependencies between different time steps in the sequence and facilitate the understanding of the context of speech. Through the self-attention mechanism, Paraformer can perform in-depth context modeling on the original encoding features to obtain updated encoding features, so that the Paraformer model can understand the long-distance dependencies in speech and transcribe the continuous speech stream more accurately. The original encoding features are further integrated in the Paraformer model to form a comprehensive feature representation, namely, integrated features. Feature integration can convert complex acoustic features into a unified internal representation. Finally, the Paraformer model passes the integrated feature representation to the output layer for text transcription, generates the message text corresponding to the voice message and outputs it.

[0112] In this embodiment, a non-autoregressive end-to-end speech recognition model is used to optimize the model using a unique speech data set generated in the field of intelligent manufacturing as a training sample, and a non-autoregressive end-to-end speech recognition model is used to achieve transcription of voice messages to message texts, which greatly improves the applicability and accuracy of the text transcription results, ensuring that the model can understand and process terms and concepts specific to the field of intelligent manufacturing, facilitating subsequent accurate intent recognition and entity extraction, and then obtaining manufacturing process data for data analysis.

[0113] In some embodiments, based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, intent recognition and entity extraction are performed on the manufacturing process inquiry message to obtain intent information and entity information, including:

[0114] Based on the natural language understanding model, the message text of the manufacturing process inquiry message is respectively subjected to intent recognition and entity extraction to obtain the original intent information and original entity words.

[0115] Obtain the historical status information corresponding to the historical query interaction message corresponding to the manufacturing process query message, and construct the current status information based on the original intent information and the original entity words;

[0116] Intent prediction is performed based on historical state information and current state information to obtain intent information that matches the manufacturing process query message;

[0117] The original entity words are standardized to obtain entity information that matches the manufacturing process query message.

[0118] Among them, the natural speech understanding model is a model that extracts information from text through specific algorithms, understands the meaning of the text, and can perform corresponding operations or generate responses based on these understandings. Through the natural language understanding model, the message text of the manufacturing process inquiry message can be respectively subjected to intent recognition and entity extraction, and the original intent information and original entity words can be directly obtained. In the process of intent recognition, the natural speech understanding model can use the historical state information corresponding to the historical inquiry interaction message corresponding to the original intent information directly obtained from the manufacturing process inquiry message to construct the current state information to realize the combination of the historical state and the current state, and then perform intent prediction through the combination of the historical state information and the current state information to obtain intent information, so that in the case of a lack of words that directly represent the intent in the message text of the manufacturing process inquiry message, the user's intent can be clarified in combination with the historical messages.

[0119] In some of the embodiments, the original entity words are standardized to obtain entity information that matches the manufacturing process query message, including: obtaining standardized entity expressions in the field of intelligent manufacturing; and performing entity mapping on the original entity words according to the standardized entity expressions to obtain standardized entity information.

[0120] Since the message text of the manufacturing process inquiry message may contain abbreviations or aliases, the extracted original entity words can be standardized to obtain entity information matching the manufacturing process inquiry message. Different intelligent manufacturing fields may correspond to different standardized entity expressions. In computer equipment, standardized entity expressions matching the intelligent manufacturing field can be pre-configured, so that in the process of entity extraction, the standardized entity expressions in the intelligent manufacturing field can be used to obtain entity information that can accurately express the meaning of the entity.

[0121] Among them, standardization processing can be achieved through entity mapping processing of industry standard words. For example, for the original entity word, in the defined synonym mapping table, find out whether the original entity word has a corresponding synonym. If a synonym is found, the original entity word is replaced with the synonym, thereby achieving standardization processing of the original entity word and obtaining entity information.

[0122] In a specific application, taking the natural speech understanding model Rasa as an example, the process of Rasa performing intent recognition and entity extraction on the message text of the manufacturing process inquiry message is as follows: Figure 5 As shown. Rasa can specifically include a semantic understanding component Interpreter (502), an entity mapping component EntitySynonymMapper (504), a state tracker component Tracker (506), a dialogue strategy component Policy (508), and an intent analysis component Action (510).

[0123] Specifically, the process of realizing intent recognition and entity extraction through the Rasa model includes:

[0124] The message text of the manufacturing process inquiry message is first passed to the Interpreter component of Rasa, which is usually composed of an NLU (Natural Language Understanding) model and is responsible for understanding the user's message; the Interpreter component uses a pre-trained model to analyze the text, identify the user's intent, and extract the entity in the message. Among them, the intent is the basic operation that the user wants to complete, and the entity is the specific data in the message.

[0125] In order to improve the accuracy of the processing results, it is necessary to standardize the abbreviations or aliases in the message text. Specifically, if the message text contains abbreviations or aliases, the entity mapping component EntitySynonymMapper will convert the abbreviations or aliases into a standardized full name format. The specific steps are as follows: First, the entity extraction component identifies the entity words in the message text and gives the original values. Secondly, the entity mapping component will look up whether the original value of the entity word has a corresponding synonym in the defined synonym mapping table. If a synonym is found, the original value is replaced with the synonym. The entity value after synonym replacement is filled into the corresponding entity category so that it can be used as input for subsequent dialogue processing, which can improve the accuracy and consistency of the dialogue system, enable the system to better understand the user's intentions, and the rational use of entity mapping can effectively handle the user's diverse expressions and improve the user experience.

[0126] Furthermore, the extracted intent and entity information are then passed to the dialog state tracker Tracker. Tracker is responsible for maintaining the current state of the dialog, which includes the user intent, entity values, and historical messages of the current message text obtained from the current message text. Correspondingly, each historical message also has a corresponding historical state, including the user intent, entity values, and historical messages that occurred before the historical message. Tracker passes the current state and historical state to the dialog policy component Policy.

[0127] The Policy component is essentially a machine learning model that can decide what action to take next based on the current state and history of the conversation. The Policy component predicts the next most appropriate intent by analyzing the current state and historical state, and sends it to the Action component. The Action component can reply to the user's message, perform an operation, or request more information. Furthermore, after the predicted intent is executed, the Action component can obtain the execution result and return the execution result to the Tracker component, so that the execution result is added to the historical state of the conversation for future decision-making.

[0128] In this embodiment, the natural speech understanding model can perform intent recognition and entity extraction on the message text of the manufacturing process inquiry message, and can use the natural speech understanding model in combination with historical messages to perform intent analysis to obtain more accurate intent information. At the same time, by standardizing entity words, entity information that more accurately expresses the meaning of the entity can be obtained, thereby improving the accuracy of subsequent data search processes.

[0129] In some of the embodiments, the computer device may send the intent information and the entity information to the data analysis platform to search for manufacturing indicators that are consistent with the intent information and match the entity information. Before sending, the computer device may format the intent information and the entity information according to the data format requirements of the data analysis platform, so that the data analysis platform can directly search for data indicators according to the intent information and the entity information to obtain the manufacturing indicators.

[0130] In some specific applications, the intent information extracted by Rasa includes intent and the corresponding intent-value, and the entity information includes entity and the corresponding entity-value. Taking the format requirement of the manufacturing process characteristic analysis platform as JSON format, the computer device sets the intent and entity as the key in the JSON format, and the corresponding value is set as value. The conversion result is {"key1": "value1", "key2": "value2"...}, where the type of value can be customized according to actual conditions, supporting string, array, Boolean and other types.

[0131] In this embodiment, by converting the data format, errors caused by incompatibility of the data formats between the computer device and the data analysis platform can be avoided, and effective interaction between the computer device and the data analysis platform can be ensured, ensuring smooth execution of data search.

[0132] In some embodiments, the manufacturing indicator is a manufacturing indicator combination including a manufacturing process intention indicator and a manufacturing process entity indicator. From the data analysis platform in the field of intelligent manufacturing, the manufacturing indicator that meets the intention information and matches the entity information is searched, specifically including:

[0133] From the data analysis platform in the field of intelligent manufacturing, find the manufacturing process intent indicators that match the intent information and the manufacturing process entity indicators that match the entity information;

[0134] The manufacturing process intention index is combined with the manufacturing process entity index to obtain the manufacturing index.

[0135] Among them, in the data analysis platform, there are many different categories of data indicators, which are used to refer to different data composition structures. Specifically, the data analysis platform may include data indicators that conform to the intent information, data indicators that match the entity information, etc. After the computer device sends the intent information and the entity information to the data analysis platform, the data analysis platform can search for the manufacturing process intent indicators that conform to the intent information and the manufacturing process entity indicators that match the entity information through information matching, wherein the number of manufacturing process intent indicators that conform to the intent information is generally 1, and the number of manufacturing process entity indicators that match the entity information can be 1 or more.

[0136] Specifically, the data analysis platform can directly feed back the found manufacturing process intention indicators and manufacturing process entity indicators to the computer device, and the computer device combines the manufacturing process intention indicators and manufacturing process entity indicators to obtain manufacturing indicators. The data analysis platform can also combine the found manufacturing process intention indicators and manufacturing process entity indicators to obtain manufacturing indicators and then send them to the computer device.

[0137] In some of the embodiments, the data analysis platform provides a query interface, a data analysis interface, and a visualization interface; further, according to the intent information, the manufacturing process data corresponding to the manufacturing indicator is analyzed for manufacturing process characteristics to obtain visualization display data representing the manufacturing process characteristics, including:

[0138] Call the query interface that matches the manufacturing index to query the manufacturing process data corresponding to the manufacturing index;

[0139] Calling the data analysis interface that matches the intent information, performing manufacturing process characteristic analysis on the manufacturing process data, and obtaining the manufacturing process characteristic analysis results;

[0140] The visualization interface that matches the data analysis interface is called to convert the manufacturing process characteristic analysis results to obtain visualization display data.

[0141] Among them, in order to meet the needs of users to query the corresponding data of data indicators in a specific intelligent manufacturing field, the digital analysis platform can pre-configure the query interface, data analysis interface and visualization interface. Among them, the query interface can receive the conversation state consisting of intent information, entity information and historical messages. The query interface allows users to query specific data based on specific data indicators. Each data indicator has a corresponding query interface so that the digital analysis platform can provide accurate data based on the user's selection. The data analysis interface is used to process the query conditions provided by the user, and call the corresponding manufacturing process characteristic analysis and analysis logic according to the query conditions. The visualization interface is used to intuitively display the analysis results to the user in the form of graphics or tables, so that users can easily understand complex data.

[0142] In a specific implementation, the computer device calls a query interface that matches the manufacturing indicators so that the digital analysis platform queries the manufacturing process data corresponding to the manufacturing indicators and feeds back the data to the computer device, thereby acquiring the manufacturing process data. The computer device then calls a data analysis interface that matches the intent information so that the digital analysis platform performs a manufacturing process characteristic analysis on the manufacturing process data, obtains the manufacturing process characteristic analysis results and feeds back the results to the computer device. Finally, the visualization interface that matches the data analysis interface is called so that the digital analysis platform performs data conversion on the manufacturing process characteristic analysis results, obtains visual display data and feeds back the data to the computer device, thereby achieving effective interaction between the computer device and the digital analysis platform.

[0143] In a specific embodiment, taking the manufacturing industry as an example in the field of intelligent manufacturing, a computer device inputs formatted data (602) to a data analysis platform (604). The data analysis platform is a key tool for data management and analysis in the manufacturing industry, which can help enterprises monitor the performance of production lines, optimize production processes, and improve overall efficiency. The data analysis platform can be built based on a manufacturing execution system (MES) or other industrial systems to collect and analyze data related to the production line.

[0144] The implementation process of the specific interaction between the data analysis platform and the computer equipment is as follows:

[0145] In order to call the interface of the digital analysis platform, the following data needs to be collected during the intelligent question-answering process: Intent information, including the specific indicators that the user wants to query, such as production capacity, YU, FTY, etc. Entity information, including specific objects related to the query, such as specific production lines, product models, time ranges, etc. Dialogue state, whose data composition includes intent information and entity information, as well as the user's previous query history, so that the system can provide coherent context-related information.

[0146] The formatted data obtained by the computer device from Rasa needs to be integrated into the digital analysis platform. This usually involves passing Rasa's output data to the platform's underlying service interface. Rasa's output data may include intent information, entity information, or conversation status containing intent information and entity information. These output data are important indicators for analyzing the status of the production line.

[0147] In order to meet the needs of users to query data corresponding to data indicators in specific intelligent manufacturing fields, the digital analysis platform can pre-configure query interfaces, data analysis interfaces, and visualization interfaces.

[0148] Among them, the query interface can receive the conversation state composed of intent information, entity information and historical messages. The query interface allows users to query specific data based on specific data indicators such as production capacity, YU, FTY, etc. Each data indicator has a corresponding query interface so that the digital analysis platform can determine which indicator's data analysis interface to call based on the intent information and entity information, so as to provide accurate data according to the user's choice. The data analysis interface is used to process the query conditions provided by the user and call the corresponding manufacturing process characteristic analysis and analysis logic according to the query conditions. Specifically, it can be based on the data provided by the query interface and combined with the conversation state to perform necessary manufacturing process characteristic analysis and analysis, such as calculating average values, trend analysis, etc. Once the data is passed in, the data analysis platform will process the data according to the predefined indicator algorithm. The indicator algorithm is designed based on the best practices of the manufacturing industry and the specific needs of the enterprise to ensure that the performance of the production line can be accurately reflected. The visualization interface is used to receive the results of the data analysis interface and convert them into charts or tables so that the user interface can display this information and intuitively display the analysis results to the user in the form of graphics or tables, so that the user can easily understand complex data.

[0149] like Figure 6As shown, the computer device inputs the formatted data (602) into the data analysis platform (604). When the data analysis platform interface service is called, the system checks whether the response is successful (606). If successful, the digital analysis platform will return the analyzed data information (608), which may include key performance indicators (KPIs) such as production efficiency, failure rate, and capacity. If the interface service call fails, the system can ensure the integrity and accuracy of the data by retrying to call the interface service or guiding the user to enter more effective information when necessary. For example, in some cases, if the data analysis platform needs more contextual information to accurately interpret certain indicators, the computer device can prompt the user to enter additional data or clarify certain information.

[0150] A series of pre-developed general data presentation components can be configured in the digital analysis platform or computer device. Each data presentation component can receive data input of different indicators, automatically select the appropriate presentation method according to the structure and configuration of the data, and automatically generate corresponding charts or text descriptions. The general components have the characteristics of flexible design and high maintainability, so that they can adapt to different data structures and presentation requirements. Through the above processing method, the processing efficiency of converting the results of manufacturing process characteristic analysis into visual display data can be further improved. Furthermore, in order to ensure that users can easily understand and use the data presentation components, the computer device can provide clear legends, tooltips, and operation guides for the data presentation components, and pre-configure the display adaptation strategy to ensure that the charts can be well displayed on different devices and screen sizes.

[0151] In this embodiment, the powerful data management and data analysis processing capabilities of the data analysis platform can be utilized to realize effective interaction between the data analysis platform and the computer device by calling various interfaces configured on the data analysis platform. The manufacturing process inquiry message input by the user is analyzed by the computer device to obtain the intention information and entity information, and the information is provided to the data analysis platform through interface calls to obtain the analysis and processing results of the data analysis platform, which are finally presented to the user in the form of visual display data, thereby realizing efficient interactive response.

[0152] The present application also provides an application scenario, which applies the above-mentioned interactive processing method in the field of intelligent manufacturing. Specifically, the application of the interactive processing method in the field of intelligent manufacturing in the application scenario is as follows:

[0153] Most of the existing question-answering systems on the market currently use technologies such as information retrieval and knowledge reasoning to help users obtain the required knowledge more conveniently. However, this type of question-answering system still has certain limitations in accurately understanding questions and providing effective feedback in specific fields. To this end, this application proposes an interactive processing in the field of intelligent manufacturing, which specifically includes the following contents:

[0154] The computer device provides an interactive interface for querying characteristic data of the manufacturing process. The interactive interface displays text input controls and voice input controls for the user to select an interaction method. If the user selects the voice input control in the interactive interface and triggers a voice input operation, the computer device will respond to the voice input operation and obtain input voice data, which is a manufacturing process inquiry message.

[0155] The computer device transmits the input voice data to the Paraformer model. The Paraformer model performs in-depth feature extraction on the input voice data to obtain voice feature vectors, and converts the sound wave data in the original voice data into feature vectors that can represent the essence of voice. These feature vectors capture the spectrum, pitch, energy and other potential acoustic features of the voice. The extracted feature vectors are sent to the Paraformer model for encoding to obtain the original encoding features. The Paraformer model utilizes the self-attention mechanism in the Transformer model, which can effectively capture the dependencies between different time steps in the sequence and facilitate the understanding of the context of the voice. Through the self-attention mechanism, the Paraformer can perform in-depth context modeling on the original encoding features to obtain updated encoding features, so that the Paraformer model can understand the long-distance dependencies in the voice, thereby more accurately transcribing the continuous voice stream. The original encoding features are further integrated in the Paraformer model to form a comprehensive feature representation, namely, integrated features. Feature integration can convert complex acoustic features into a unified internal representation. Finally, the Paraformer model passes the integrated feature representation to the output layer for text transcription to generate the message text corresponding to the voice message.

[0156] Next, the message text of the manufacturing process inquiry message is passed to the Interpreter component of Rasa, which uses a pre-trained model to analyze the text, identify the user's intent, and extract entities in the message text. If the message text contains abbreviations or aliases, the entity mapping component EntitySynonymMapper converts the abbreviations or aliases into a standardized full name format. Further, the extracted intent information and standardized entity information are then passed to the conversation state tracker Tracker. Tracker is responsible for maintaining the current state of the conversation based on the user's intent, entity values, and historical messages of the current message text obtained from the current message text. Subsequently, Tracker passes the current state and historical state to the conversation policy component Policy. The Policy component predicts the next most appropriate intent by analyzing the current state and historical state, and sends it to the Action component. The Action component can be a reply to the user's message, perform an operation, or request more information. Further, after the predicted intent is executed, the Action component can obtain the execution result and return the execution result to the Tracker component.

[0157] Thus, Rasa extracts the intent information and entity information. The intent information includes intent and the corresponding intent-value, and the entity information includes entity and the corresponding entity-value. Taking the format requirement of the manufacturing process characteristic analysis platform as JSON format, the computer device sets the intent and entity as the key in the JSON format, and the corresponding value is set as value. The conversion result is {"key1": "value1", "key2": "value2"...}, where the type of value can be customized according to the actual situation, supporting string, array, Boolean and other types, so that the data analysis platform can directly match the intent information and entity information in the format, find the data indicators to obtain the manufacturing indicators, and perform data analysis and other processing based on the manufacturing process data corresponding to the manufacturing indicators, and finally obtain the visual display data for display to the user.

[0158] Rasa's output data may include intent information, entity information, or a dialog state that includes intent information and entity information. These output data are important indicators for analyzing the status of the production line. The formatted data obtained by the computer device from Rasa needs to be integrated into the digital analysis platform, which specifically involves passing Rasa's output data to the platform's underlying service interface.

[0159] In order to meet the needs of users to query the corresponding data of data indicators in a specific intelligent manufacturing field, the digital analysis platform can pre-configure the query interface, data analysis interface and visualization interface. Among them, the query interface can receive the conversation state composed of intent information, entity information and historical messages. The query interface allows users to query specific data based on specific data indicators such as production capacity, YU, FTY, etc. Each data indicator has a corresponding query interface so that the digital analysis platform can determine which indicator's data analysis interface to call based on the intent information and entity information, so as to provide accurate data according to the user's choice. The data analysis interface is used to process the query conditions provided by the user and call the corresponding data processing and analysis logic according to the query conditions. Specifically, it can be based on the data provided by the query interface and combined with the conversation state to perform necessary data processing and analysis, such as calculating the average value, trend analysis, etc. Once the data is passed in, the data analysis platform will process the data according to the predefined indicator algorithm. The indicator algorithm is designed according to the best practices of the manufacturing industry and the specific needs of the enterprise to ensure that the performance of the production line can be accurately reflected. The visualization interface is used to receive the results of the data analysis interface and convert them into charts or tables so that the user interface can display this information and intuitively display the analysis results to the user in the form of graphics or tables, so that users can easily understand complex data.

[0160] When the interface service is called, the system checks whether the response is successful. If successful, the digital analysis platform will return analyzed data information, which may include key performance indicators such as production efficiency, failure rate, and capacity. If the interface service call fails, the system can ensure the integrity and accuracy of the data by retrying to call the interface service or guiding the user to enter more effective information when necessary. For example, in some cases, if the data analysis platform needs more contextual information to accurately interpret certain indicators, the computer device can prompt the user to enter additional data or clarify certain information.

[0161] Through the above processing method, the Paraforme model is used to optimize the model with the unique voice data set generated by the manufacturing industry as the training sample, and the Paraforme model is used to realize the transcription of voice messages to message text, which greatly improves the applicability and accuracy of the text transcription results, ensures that the model can understand and process the terms and concepts specific to the field of intelligent manufacturing, and facilitates the subsequent accurate intent recognition and entity extraction, and then obtains the manufacturing process data for data analysis. And by integrating Rasa technology and combining training based on the Paraformer model, it can efficiently perform deep feature extraction and accurate entity matching on the text input by the user. The bidirectional training mechanism and the ability to understand contextual relationships of the Paraformer model enable it to perform well in the complex environment of the manufacturing industry and better capture user intentions and key information. In addition, the solution fully utilizes the efficiency and functionality of the underlying interface services by tightly integrating natural language processing, speech recognition and underlying interface services. It not only improves the overall performance of the system, but also ensures the smoothness and accuracy of data processing. Finally, by providing highly customized visual presentation, it meets the specific functional requirements of the manufacturing industry for data display. Through the configurable user interface, production data and management indicators can be presented in an intuitive and easy-to-understand way, helping management and operators make more appropriate decisions.

[0162] It should be understood that, although the steps in the flowcharts involved in the above embodiments 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 part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0163] Based on the same inventive concept, the embodiment of the present application also provides an interactive processing device in the field of intelligent manufacturing for implementing the interactive processing method in the field of intelligent manufacturing involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in the embodiments of one or more interactive processing devices in the field of intelligent manufacturing provided below can refer to the limitations of the interactive processing method in the field of intelligent manufacturing above, and will not be repeated here.

[0164] In one embodiment, Figure 7As shown, an interactive processing device in the field of intelligent manufacturing is provided, including: an input processing module 702, an intention recognition and entity extraction module 704, an indicator search module 706, a manufacturing process characteristic analysis module 708 and a visualization display module 710, wherein:

[0165] An input processing module 702 is used to obtain an input manufacturing process inquiry message in response to an inquiry message input operation triggered on an interactive interface in an interactive scenario in the field of intelligent manufacturing;

[0166] An intent recognition and entity extraction module 704 is used to perform intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information;

[0167] An indicator search module 706 is used to search for manufacturing indicators that meet the intention information and match the entity information from the data analysis platform in the intelligent manufacturing field;

[0168] A manufacturing process characteristic analysis module 708 is used to perform manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator according to the intention information to obtain visual display data representing the manufacturing process characteristics;

[0169] The visualization display module 710 is used to use the visualization display data as a response result of the manufacturing process inquiry message and display the visualization display data in the interactive interface.

[0170] In some embodiments, an interactive processing device in the field of intelligent manufacturing includes a text transcription module, which is used to extract features from the voice message based on a non-autoregressive end-to-end speech recognition model to obtain a voice feature vector; the non-autoregressive end-to-end speech recognition model is trained based on a sample voice message, and the sample voice message contains professional terms in the field of intelligent manufacturing; through the self-attention mechanism of the non-autoregressive end-to-end speech recognition model, the voice feature vector is sequentially feature encoded and context modeled to obtain a coded feature; based on the coded feature, feature integration is performed to obtain an integrated feature; the integrated feature is transcribed into text to obtain a message text representing the voice message;

[0171] The intent recognition and entity extraction module 704 is specifically used to perform intent recognition and entity extraction on the message text based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information.

[0172] In some embodiments, the intent recognition and entity extraction module 704 is specifically used to perform intent recognition and entity extraction on the message text of the manufacturing process inquiry message based on a natural language understanding model to obtain original intent information and original entity words; obtain historical status information corresponding to the historical inquiry interaction message corresponding to the manufacturing process inquiry message, and construct current status information based on the original intent information and original entity words; perform intent prediction based on the historical status information and the current status information to obtain intent information that matches the manufacturing process inquiry message; and standardize the original entity words to obtain entity information that matches the manufacturing process inquiry message.

[0173] In some embodiments, the intent recognition and entity extraction module 704 is specifically used to obtain standardized entity expressions in the field of intelligent manufacturing; according to the standardized entity expressions, the original entity words are entity mapped to obtain standardized entity information.

[0174] In some embodiments, the manufacturing indicator is a manufacturing indicator combination including a manufacturing process intention indicator and a manufacturing process entity indicator; the indicator search module 706 is specifically used to search for manufacturing process intention indicators that meet the intention information and manufacturing process entity indicators that match the entity information from the data analysis platform in the intelligent manufacturing field; the manufacturing process intention indicator is combined with the manufacturing process entity indicator to obtain a manufacturing indicator.

[0175] In some embodiments, the data analysis platform provides a query interface, a data analysis interface and a visualization interface; the manufacturing process characteristic analysis module 708 is used to call the query interface that matches the manufacturing indicator to query the manufacturing process data corresponding to the manufacturing indicator; call the data analysis interface that matches the intention information to perform a manufacturing process characteristic analysis on the manufacturing process data to obtain a manufacturing process characteristic analysis result; call the visualization interface that matches the data analysis interface to perform data conversion on the manufacturing process characteristic analysis result to obtain visual display data.

[0176] The above-mentioned interactive processing device in the field of intelligent manufacturing, in an interactive scenario in the field of intelligent manufacturing, obtains an input manufacturing process inquiry message by responding to an inquiry message input operation triggered on an interactive interface, and then performs intent recognition and entity extraction on the manufacturing process inquiry message based on a historical inquiry interaction message corresponding to the manufacturing process inquiry message, thereby being able to accurately understand the user's specific intention information and interaction-related entity information of this interactive operation, and facilitating accurate acquisition of matching manufacturing process data for data analysis. In the data acquisition process, by searching for manufacturing indicators that meet the intention information and match the entity information from the data analysis platform in the field of intelligent manufacturing, it is possible to implement the manufacturing process Accurately search for data, and then according to the intention information, perform manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data characterizing the manufacturing process characteristics, and finally use the visual display data as the response result of the manufacturing process inquiry message, and display the visual display data in the interactive interface. Only simple interactive operations are required to obtain data interaction results in the field of intelligent manufacturing in a visual way. Not only can the manufacturing process data be obtained through intent recognition and information extraction for analysis to obtain intuitively displayed data results, but the interaction process is also simplified, and the time required for interactive operations is shortened, thereby effectively improving the interaction efficiency in the field of intelligent manufacturing.

[0177] Each module in the interactive processing device in the field of intelligent manufacturing can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0178] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, an interactive processing method in the field of intelligent manufacturing is realized.

[0179] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an interactive processing method in the field of intelligent manufacturing is implemented. The display unit of the computer device is used to form a visually visible image, and can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device casing, or an external keyboard, touchpad or mouse, etc.

[0180] Those skilled in the art will understand that Figure 8 or Fig. 9The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0181] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0182] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0183] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0184] 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 relevant regulations.

[0185] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a manufacturing process characteristic analysis logic device based on quantum computing, etc., but are not limited to this.

[0186] The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0187] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. An interactive processing method in the field of intelligent manufacturing, characterized in that: The method comprises: In an interactive scenario in the field of intelligent manufacturing, in response to an inquiry message input operation triggered on an interactive interface, an input manufacturing process inquiry message is obtained; the manufacturing process inquiry message is an input message that a user inputs through the interactive interface and can be used to directly obtain a visual display result by inquiring about manufacturing process data; Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, performing intent recognition and entity extraction on the manufacturing process inquiry message to obtain intent information and entity information; Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing indicators that conform to the intention information and match the entity information; According to the intention information, performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator to obtain visual display data representing the manufacturing process characteristics; Using the visual display data as a response result of the manufacturing process inquiry message, and displaying the visual display data in the interactive interface; In the case where the manufacturing process inquiry message is a voice message, the method further includes: Based on a non-autoregressive end-to-end speech recognition model, feature extraction is performed on the voice message to obtain a speech feature vector; the non-autoregressive end-to-end speech recognition model is trained based on sample voice messages, and the sample voice messages contain professional terms in the field of intelligent manufacturing; By using the self-attention mechanism of the non-autoregressive end-to-end speech recognition model, the speech feature vector is sequentially feature encoded and context modeled to obtain an encoded feature; Perform feature integration based on the coding features to obtain integrated features; Performing text transcription on the integrated features to obtain a message text representing the voice message; The method of performing intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information includes: Based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message, the message text is subjected to intent recognition and entity extraction to obtain intent information and entity information; The method of performing intent recognition and entity extraction on the manufacturing process inquiry message based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intent information and entity information includes: Based on the natural language understanding model, the message text of the manufacturing process inquiry message is respectively subjected to intent recognition and entity extraction to obtain original intent information and original entity words; Acquire historical status information corresponding to the historical query interaction message corresponding to the manufacturing process query message, and construct current status information based on the original intent information and the original entity words; Performing intention prediction based on the historical state information and the current state information to obtain intention information matching the manufacturing process inquiry message; Standardizing the original entity words to obtain entity information matching the manufacturing process inquiry message; The manufacturing index is a manufacturing index combination including a manufacturing process intention index and a manufacturing process entity index; The step of searching, from the data analysis platform in the field of intelligent manufacturing, a manufacturing indicator that meets the intention information and matches the entity information comprises: Searching, from the data analysis platform in the field of intelligent manufacturing, manufacturing process intention indicators that match the intention information and manufacturing process entity indicators that match the entity information; The manufacturing process intention indicator is combined with the manufacturing process entity indicator to obtain a manufacturing indicator combination.

2. The method according to claim 1, characterized in that The step of performing standardization processing on the original entity words to obtain entity information matching the manufacturing process inquiry message includes: Obtain standardized expressions of entities in the field of intelligent manufacturing; According to the entity standardized expression, the original entity word is entity mapped to obtain standardized entity information.

3. The method according to claim 1 or 2, characterized in that: The data analysis platform provides a query interface, a data analysis interface and a visualization interface; The step of performing manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator according to the intention information to obtain visual display data representing the manufacturing process characteristics includes: Calling a query interface matching the manufacturing index to query manufacturing process data corresponding to the manufacturing index; Calling a data analysis interface that matches the intention information, performing a manufacturing process characteristic analysis on the manufacturing process data, and obtaining a manufacturing process characteristic analysis result; A visualization interface matching the data analysis interface is called to perform data conversion on the manufacturing process characteristic analysis result to obtain visualization display data.

4. An interactive processing device in the field of intelligent manufacturing, characterized in that: The device comprises: An input processing module is used to obtain an input manufacturing process inquiry message in response to an inquiry message input operation triggered on an interactive interface in an interactive scenario in the field of intelligent manufacturing; the manufacturing process inquiry message is an input message that a user inputs through the interactive interface and can be used to directly obtain a visual display result by inquiring about manufacturing process data; An intent recognition and entity extraction module, configured to perform intent recognition and entity extraction on the manufacturing process inquiry message based on a historical inquiry interaction message corresponding to the manufacturing process inquiry message, to obtain intent information and entity information; An indicator search module, used to search for manufacturing indicators that meet the intention information and match the entity information from the data analysis platform in the field of intelligent manufacturing; A manufacturing process characteristic analysis module, used to perform manufacturing process characteristic analysis on the manufacturing process data corresponding to the manufacturing indicator according to the intention information to obtain visual display data; A visualization display module, used to use the visualization display data as a response result of the manufacturing process inquiry message, and display the visualization display data in the interactive interface; In the case where the manufacturing process inquiry message is a voice message, the device also includes a text transcription module, which is used to extract features from the voice message based on a non-autoregressive end-to-end speech recognition model to obtain a voice feature vector; through the self-attention mechanism of the non-autoregressive end-to-end speech recognition model, the voice feature vector is sequentially feature encoded and context modeled to obtain a coded feature; based on the coded feature, feature integration is performed to obtain an integrated feature; the integrated feature is transcribed into text to obtain a message text representing the voice message; the non-autoregressive end-to-end speech recognition model is trained based on a sample voice message, and the sample voice message contains professional terms in the field of intelligent manufacturing; The intention recognition and entity extraction module is specifically used to perform intention recognition and entity extraction on the message text based on the historical inquiry interaction message corresponding to the manufacturing process inquiry message to obtain intention information and entity information; The intention recognition and entity extraction module is specifically used to perform intention recognition and entity extraction on the message text of the manufacturing process inquiry message based on a natural language understanding model to obtain original intention information and original entity words; obtain historical status information corresponding to the historical inquiry interaction message corresponding to the manufacturing process inquiry message, and construct current status information based on the original intention information and original entity words; perform intention prediction based on the historical status information and the current status information to obtain intention information matching the manufacturing process inquiry message; perform standardization on the original entity words to obtain entity information matching the manufacturing process inquiry message; The manufacturing indicator is a manufacturing indicator combination including a manufacturing process intention indicator and a manufacturing process entity indicator; the indicator search module is specifically used to search for manufacturing process intention indicators that meet the intention information and manufacturing process entity indicators that match the entity information from the data analysis platform in the intelligent manufacturing field; the manufacturing process intention indicator is combined with the manufacturing process entity indicator to obtain a manufacturing indicator.

5. The device according to claim 4, characterized in that The intention recognition and entity extraction module is specifically used to obtain standardized entity expressions in the field of intelligent manufacturing; according to the standardized entity expressions, the original entity words are entity mapped to obtain standardized entity information.

6. The device according to claim 4 or 5, characterized in that The data analysis platform provides a query interface, a data analysis interface and a visualization interface; the manufacturing process characteristic analysis module is specifically used to call the query interface matching the manufacturing index to query the manufacturing process data corresponding to the manufacturing index; call the data analysis interface matching the intention information to perform manufacturing process characteristic analysis on the manufacturing process data to obtain the manufacturing process characteristic analysis results; A visualization interface matching the data analysis interface is called to perform data conversion on the manufacturing process characteristic analysis result to obtain visualization display data.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 3 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 3 are implemented.

Images