Job data processing method and device based on dialogue mode

Abstract

This disclosure provides a dialogue-based method and apparatus for processing job data, applicable to the fields of autonomous driving, natural language processing, and information retrieval. The method includes: acquiring initial requirement information input by a target object through dialogue; displaying data feedback results representing job data, whereby the job data represents the job behavior performed by the target object in a job scenario, and the data attributes of the job data match the initial requirement information; determining job description tags corresponding to the job data based on the target object's labeled requirement intent in response to the data feedback results; and determining target job data based on the job description tags and the job data.

Description

Technical Field

[0001] This disclosure relates to the fields of autonomous driving, natural language processing, and information retrieval, and more specifically, to a dialogue-based method and apparatus for processing job data. Background Technology

[0002] Autonomous driving scenario data is crucial for the training and evaluation of autonomous driving models in mining areas. For massive amounts of scenario data, the industry mainly uses keywords to label the data. However, when retrieving data representing operational behaviors, the semantics of the labels are difficult to accurately match with the interactive text entered by the user. Furthermore, the data labeling process is quite complex, making it difficult for the labels to accurately represent the data used in mining areas. Summary of the Invention

[0003] In view of this, this disclosure provides a method and apparatus for processing job data based on a dialogue mechanism. An electronic device, a storage medium, and a program product are also provided.

[0004] One aspect of this disclosure provides a dialogue-based method for processing job data, comprising: acquiring initial requirement information input by a target object through dialogue; displaying data feedback results representing job data, wherein the job data represents the job behavior performed by the job object in the job scenario, and the data attributes of the job data match the initial requirement information; determining job description tags corresponding to the job data based on the annotation requirement intent input by the target object in response to the data feedback results; and determining target job data based on the job description tags and the job data.

[0005] Another aspect of this disclosure provides a dialogue-based job data processing apparatus, comprising: an acquisition module for acquiring initial requirement information input by a target object through dialogue; a display module for displaying data feedback results representing job data, wherein the job data represents the job behavior performed by the job object in the job scenario, and the data attributes of the job data match the initial requirement information; a first determination module for determining job description tags corresponding to the job data based on the annotation requirement intent input by the target object in response to the data feedback results; and a second determination module for determining target job data based on the job description tags and the job data.

[0006] Another aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein, when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to perform the method as described above.

[0007] Another aspect of this disclosure provides a computer-readable storage medium storing computer-executable instructions, which, when executed, are used to implement the method described above.

[0008] Another aspect of this disclosure provides a computer program product including computer-executable instructions that, when executed, are used to implement the method as described above.

[0009] According to embodiments of this disclosure, by acquiring initial demand information and displaying data feedback results that match the initial demand information, the target object can obtain the data attributes of the task data to be labeled based on the data feedback results through natural language description in a dialogue manner. Thus, by inputting the labeling demand intent, the task description label is determined, so as to realize the labeling of the task data to be labeled quickly and accurately through dialogue, thereby improving the labeling accuracy and precision of the generated labeled target task data. Attached Figure Description

[0010] The above and other objects, features and advantages of this disclosure will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:

[0011] Figure 1 The present disclosure illustrates an application scenario of a dialogue-based job data processing method and apparatus according to embodiments thereof;

[0012] Figure 2 A flowchart of a dialogue-based job data processing method according to an embodiment of the present disclosure is shown;

[0013] Figure 3 A schematic diagram illustrating semantic understanding according to embodiments of the present disclosure is shown;

[0014] Figure 4 A schematic diagram illustrating the input of initial requirements information according to an embodiment of this disclosure is shown;

[0015] Figure 5 A schematic diagram of a job-related terminology according to an embodiment of the present disclosure is shown;

[0016] Figure 6 A schematic diagram of conversational scene retrieval according to an embodiment of the present disclosure is shown;

[0017] Figure 7 A structural block diagram of a dialogue-based job data processing apparatus according to an embodiment of the present disclosure is shown;

[0018] Figure 8 A block diagram of an electronic device suitable for a dialogue-based job data processing method according to an embodiment of the present disclosure is shown. Detailed Implementation

[0019] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0020] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0021] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0022] When using expressions such as "at least one of A, B and C", they should generally be interpreted in accordance with the meaning that is commonly understood by those skilled in the art (e.g., "a system having at least one of A, B and C" should include, but is not limited to, a system having A alone, a system having B alone, a system having C alone, a system having A and B, a system having A and C, a system having B and C, and / or a system having A, B and C, etc.).

[0023] Autonomous driving scenario data is crucial for the training and evaluation of autonomous driving models in mining areas. For massive amounts of scenario data, the industry primarily uses keyword tagging and vector features as the main methods for scenario retrieval. While keyword-based scenario filtering is relatively convenient, data keyword tags need to be pre-set and processed before use, resulting in a certain lag (untagged scenarios are unusable). The same scenario can be described using similar keywords from a natural language perspective, but existing retrieval systems cannot understand similar semantics. Furthermore, keyword-based retrieval requires pre-setting the available tag categories and number, and pre-processing the data, leading to lag; users can only filter existing keywords, unable to perform real-time, in-depth data processing (users are only passively using the data); tags are essentially a result of collecting users' highest needs, but the process of users deeply processing the data and related solutions cannot be effectively recorded, resulting in low information content in the scenario tagging system.

[0024] In view of this, this disclosure provides a method and apparatus for processing job data based on a dialogue method. The method includes: acquiring initial requirement information input by a target object through a dialogue method; displaying data feedback results representing job data, wherein the job data represents the job behavior performed by the job object in the job scenario, and the data attributes of the job data match the initial requirement information; determining job description tags corresponding to the job data based on the annotation requirement intent input by the target object in response to the data feedback results; and determining target job data based on the job description tags and the job data.

[0025] In the technical solutions disclosed herein, the collection, updating, analysis, processing, use, transmission, provision, disclosure, and storage of data (e.g., including but not limited to user personal information) comply with relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken to prevent unauthorized access to user personal information data and to safeguard user personal information security, network security, and national security. In the embodiments of this disclosure, user authorization or consent is obtained before acquiring or collecting user personal information.

[0026] Figure 1 The illustration shows an application scenario of the dialogue-based job data processing method and apparatus according to embodiments of the present disclosure. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0027] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 serves as a medium for providing communication links between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0028] Users can use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 via the network 104 to receive or send messages, etc. Various communication client applications can be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social media platform software, etc. (for example only).

[0029] The first terminal device 101, the second terminal device 102, and the third terminal device 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0030] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103 (this is just an example). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0031] It should be noted that the dialogue-based job data processing method provided in this embodiment can generally be executed by server 105. Correspondingly, the dialogue-based job data processing device provided in this embodiment can generally be located in server 105. The dialogue-based job data processing method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Correspondingly, the dialogue-based job data processing device provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with the first terminal device 101, the second terminal device 102, the third terminal device 103, and / or server 105. Alternatively, the dialogue-based job data processing method provided in this embodiment can also be executed by the first terminal device 101, the second terminal device 102, or the third terminal device 103, or by other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103. Accordingly, the dialogue-based job data processing device provided in this embodiment can also be located in the first terminal device 101, the second terminal device 102, or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102, or the third terminal device 103.

[0032] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0033] Figure 2 A flowchart illustrating a dialogue-based job data processing method according to an embodiment of the present disclosure is shown.

[0034] like Figure 2 As shown, the method includes operations S201 to S204.

[0035] In operation S201, the initial requirement information input by the target object through a dialogue is obtained.

[0036] In operation S202, the data feedback results representing the operation data are displayed.

[0037] According to embodiments of this disclosure, job data characterizes the job behavior performed by a job object in a job scenario, and the data attributes of the job data are matched with the initial requirement information.

[0038] In operation S203, based on the target object's annotation requirements for the data feedback results, determine the job description label corresponding to the job data.

[0039] In operation S204, the target job data is determined based on the job description label and job data.

[0040] According to embodiments of this disclosure, the target audience includes algorithm engineers, testers, and operations personnel, but is not limited to these; embodiments of this disclosure do not limit the target audience. Initial requirement information represents the first input information in the dialogue process, or multiple levels of requirement intent obtained through follow-up questions before clarifying detailed requirements.

[0041] According to embodiments of this disclosure, the work objects may include equipment such as bulldozers and excavators, but are not limited to these; embodiments of this disclosure do not limit the work objects. Work scenarios may include mining, road construction, etc., but are not limited to these; this disclosure does not limit the work scenarios. The data attributes of the work data may be inherent attributes of the data such as location, time, equipment, and weather at the time of data collection, or they may be determined by the annotator after annotating the work data according to actual needs.

[0042] The annotation intent may include, for example: 1. Whether the number of obstacles affecting travel exceeds 3?; 2. Input tags, such as daytime, sandstorm, crushing station, no risk, etc. However, it is not limited to these, and the embodiments of this disclosure do not limit the specific annotation intent. Based on the annotation intent input by the target object in response to the data feedback results, the operation description tags corresponding to the operation data can be determined.

[0043] The target job data is determined based on the job description tags and job data. The target job data may include associated stored job description tags and job data. The target job data can be used to train autonomous driving models, object detection models, or abnormal job behavior monitoring models; it can also be used to build map data and terrain data; or for job monitoring needs such as safety monitoring and job efficiency monitoring.

[0044] According to embodiments of this disclosure, by acquiring initial demand information and displaying data feedback results that match the initial demand information, the target object can obtain the data attributes of the task data to be labeled based on the data feedback results through natural language description in a dialogue manner. Thus, by inputting the labeling demand intent, the task description label is determined, so as to realize the labeling of the task data to be labeled quickly and accurately through dialogue, thereby improving the labeling accuracy and precision of the generated labeled target task data.

[0045] According to embodiments of this disclosure, the annotation requirement intent includes a content requirement intent, which describes the target object's requirement for the data content in the task data; based on the annotation requirement intent input by the target object in response to the data feedback result, the task description tag corresponding to the task data is determined, including: performing semantic understanding on the task data based on the content requirement intent to obtain the target task data in the task data and a content intent tag representing the content requirement intent, wherein the data content of the target task data matches the content requirement intent, and the task description tag includes the content intent tag.

[0046] Content requirement intent describes the target object's requirements for the data content in the operation data. Content requirement intent may include determining the number of obstacles, the type of obstacles, or weather conditions, but is not limited to these. This disclosure does not limit the specific content of content requirement intent.

[0047] Semantic understanding of task data can be performed based on content requirement intent to obtain the target task data and content intent tags representing the content requirement intent. For example, the content requirement intent might be to determine whether the number of obstacles affecting driving in the scene exceeds three. The data content of the target task data could include whether the number of obstacles exceeds three. The content intent tag could be an obstacle.

[0048] According to embodiments of this disclosure, semantic understanding of job data based on content requirement intent includes: performing semantic understanding of job data based on content requirement intent to obtain clarification prompt information, wherein the clarification prompt information is used to prompt the target object to clarify the missing or semantically ambiguous intent information in the content requirement intent; in response to obtaining the clarification content input by the target object in response to the clarification prompt information, performing semantic understanding of job data based on the clarification content to obtain a content intent tag, and target job data in the job data that matches the content intent tag.

[0049] By semantically understanding the task data based on the content requirements, clarification prompts can be obtained. These prompts are used to prompt the target audience to clarify any missing or semantically ambiguous intent information in the content requirements, thereby improving the accuracy of determining the actual intent.

[0050] For example, the content requirement intent could be "to filter data from mining trucks". By semantically understanding the operational data based on the content requirement intent, clarification prompts can be obtained. These clarification prompts could include "type of mining truck" or "data format", such as "does it refer to heavy mining trucks transporting ore or engineering command vehicles", "do you need video data, images, or sensor log data", or "what time period, which region, and whether it is fully loaded or empty is required".

[0051] By responding to the clarification content input by the target object in response to the clarification prompt, semantic understanding of the operation data can be performed based on the clarification content, thereby obtaining the content intent tag and the target operation data that matches the content intent tag. For example, the clarification content is "Video monitoring of large transport mining trucks is required, yesterday from 8:00 AM to 6:00 PM in the eastern area, under full load." By semantically understanding the operation data based on the clarification content, the content intent tag and the target operation data can be obtained.

[0052] According to embodiments of this disclosure, the actual intent of the target object to annotate the task data is determined by engaging in dialogue with a large model and asking follow-up questions, thereby improving the accuracy of content intent tags and the retrieval accuracy of the target task data.

[0053] Figure 3 A schematic diagram illustrating semantic understanding according to an embodiment of this disclosure is shown.

[0054] like Figure 3 As shown, conversational scenarios mainly consist of two parts: user input and result output. Users use natural language, images, and other information as input for interaction. The system provides feedback to the user with processing results through natural language analysis, retrieval, and processing, supporting multi-turn "topic-based conversational" interactions.

[0055] By engaging with the large model and asking follow-up questions, the target's actual intention in labeling the operational data can be determined, which may include multiple rounds of clarification prompts. First, the content requirement intention was determined to be filtering data 310 from mining trucks. Based on this content requirement intention, semantic understanding of the operational data was performed to obtain clarification prompts. The first clarification prompt 320 was: "Does this refer to heavy mining trucks transporting ore, or to the engineering command vehicle?"

[0056] The first clarification content (330) input by the target object in response to the clarification prompt is obtained. Further, based on this clarification content, semantic understanding is performed on the operational data again to obtain the second clarification prompt (340): "Do you need video data, images, or sensor log data?"

[0057] The system obtains the second clarification content 350 input by the target object in response to the clarification prompt, which is "screening video monitoring data of large transport mining trucks". Further, based on this second clarification content, the system performs semantic understanding on the operation data again, resulting in the third clarification prompt 340, which asks, "What time period, which area, and whether it is fully loaded or empty?"

[0058] The target object obtained the third clarification content 370 input in response to the third clarification prompt information 340. The third clarification content 370 is "Filter video surveillance of large transport mining trucks in the eastern region, from 8:00 am to 6:00 pm yesterday, in a fully loaded state".

[0059] Based on the clarification content in each round, semantic understanding of the task data can be performed to obtain content intent tags and target task data that match the content intent tags. By continuously questioning the large model through dialogue, the actual intent of the target object in labeling the task data can be determined, thereby improving the accuracy of content intent tags and the retrieval accuracy of target task data.

[0060] According to embodiments of this disclosure, determining the job description tag corresponding to the job data based on the annotation requirement intent input by the target object in response to the data feedback result includes: performing semantic understanding of the annotation requirement intent to determine target tag attribute and tag value data, wherein the target tag attribute indicates the tag type of the intent tag that the target object needs to annotate; and determining the job description tag related to the job data based on the target tag attribute and tag value data.

[0061] The target label attribute can represent the label type of the intent label that needs to be annotated on the target object. The target label attribute can represent time, weather, object, impact, operation object type, etc., but is not limited to these, and the embodiments disclosed herein do not limit this. Time can include year, month, day, and hour; weather can include sandstorms, heavy fog, etc.; impact can represent the collision safety risk attribute value corresponding to the impact attribute, etc. The operation object type can include graders, excavators, etc., and will not be elaborated further here.

[0062] The system performs semantic understanding of the annotation request intent to determine the target label attributes and label value data. For example, inputs could include "daytime," "rain / fog," and "crushing station." When semantically understanding this intent, "rain / fog" can be categorized as "weather" within the target label attributes. Then, "rain / fog" can be compared with the label value data contained in the weather section. Unlike "dense fog" within the target label attributes, "rain / fog" can be used as a new label value. Additionally, semantic understanding can be performed based on the context of the data feedback results related to the task data during the dialogue, outputting the determined target label attributes and label value data.

[0063] By analyzing and processing the annotation intent with explicit information such as time, the model's results can be processed to produce converted outputs of data text, image modal results, and scene representations. The text input by users during the dialogue process is organized into tags for data annotation.

[0064] According to embodiments of this disclosure, by semantically understanding the intent of the annotation requirements, the target tag attributes and tag value data can be determined. Then, based on the target tag attributes and tag value data, the job description tags related to the job data can be determined, which not only facilitates user retrieval but also improves the real-time performance of data processing and the convenience of interaction.

[0065] According to embodiments of this disclosure, obtaining initial requirement information input by a target object through a dialogue includes: receiving first initial requirement information input by the target object; performing semantic understanding on the first initial requirement information and outputting first feedback information that matches the requirement intent of the first initial requirement information, wherein the first feedback information describes the data attributes of initial job data and the initial job data is determined from a preset job dataset based on the first initial requirement information; displaying the first feedback information and receiving second initial requirement information input by the target object in response to the first feedback information.

[0066] Upon receiving initial demand information from the target object, the system can perform semantic understanding on this initial demand information and output initial feedback information that matches the intent of the demand. The initial feedback information describes the data attributes of the initial operation data, such as the mine name (site name), data acquisition location, and data modality type (e.g., image or point cloud). The initial operation data is determined from a preset operation dataset based on the initial demand information.

[0067] Display the first feedback information and receive the second initial requirement information input by the target object in response to the first feedback information.

[0068] Figure 4 A schematic diagram illustrating the input of initial requirements information according to an embodiment of this disclosure is shown.

[0069] like Figure 4 As shown, the system receives the first initial demand information input by the target object. The first initial demand information 410 is "Filter the data of the first week of September in xx mining area and filter out the scene data containing mining trucks and dust". The system performs semantic understanding on the first initial demand information 410 and outputs the first feedback information 420 that matches the demand intent of the first initial demand information 410. The first feedback information 420 is "A total of xx data points, covering xx mining areas...".

[0070] The system displays first feedback information 420 and receives second initial requirement information 430 input from the target object in response to the first feedback information 420. The second initial requirement information 430 is: "Display the image of the first frame of data from the first data point after 9 AM at xx mine, and analyze the scenario." Semantic understanding can be performed on the second initial requirement information to obtain the corresponding feedback result 440. The feedback result 440 includes text describing the image of the first frame of data, as well as the image of the first frame of data itself.

[0071] According to embodiments of this disclosure, semantic understanding is performed on the first initial demand information, and first feedback information matching the demand intent of the first initial demand information is output. This includes: understanding the intent of the first initial demand information based on a job-related vocabulary related to the job scenario to obtain the data attribute demand intent; calling a retrieval tool to perform a retrieval task on a preset job database based on tool parameters determined based on the data attribute demand intent to obtain initial job data; and performing semantic understanding on the initial job data to obtain the first feedback information.

[0072] The job-related terminology can be represented as a mining area-specific terminology, which may include time, job area, job object, or attributes. Attributes may include conditions, status, alarms, trajectories, or production output. By understanding the intent of the initial demand information based on the job-related terminology, the data attribute demand intent can be obtained. The retrieval tool is then invoked to perform a retrieval task on the preset job database based on the tool parameters determined by the data attribute demand intent, obtaining initial job data. For example, tool parameters may include time attributes, spatial attributes, or demand attributes. The initial job data may include the number of retrieved terms and multiple data entries corresponding to the data attribute demand intent. Semantic understanding of the initial job data yields the first feedback information.

[0073] Figure 5 A schematic diagram of a job-related terminology according to an embodiment of the present disclosure is shown.

[0074] like Figure 5 As shown, the operation-related vocabulary includes a mining area scenario lexicon, encompassing time (year, month, day, hour, etc.), weather (dust storm, fog, etc.), objects (excavators, mining trucks, etc.), behaviors (dumping, loading, etc.), quantity, type (grader, etc.), location (crushing station, No. 1 mine pit, etc.), and impacts (collision, safety risks). By combining the operation-related vocabulary with mining area scenarios, mining domain knowledge is preprocessed using natural language to guide the model to become a domain expert, ultimately facilitating interaction among domain experts through a shared domain-specific language.

[0075] According to embodiments of this disclosure, by refining the relevant content in the job-related thesaurus, the balance between domain expertise and the convenience of text-based language interaction is resolved, achieving professional abstraction of mining operations and unifying information interaction methods based on natural language.

[0076] According to embodiments of this disclosure, the second initial demand information represents the data processing intent for the initial job data, and the data feedback result is determined based on the following operations: according to the second initial demand information, the target task is executed on the job data in the initial related data that matches the data processing intent, and the execution result is obtained; the execution result is semantically understood to obtain the data feedback result.

[0077] The second initial requirement information can characterize the data processing intent for the initial task data. For example, the second initial requirement information could be an image of the first frame of the first data after 9:00 AM, allowing for analysis of the scene. Based on the second initial requirement information, the target task is executed on the task data in the initial related data that matches the data processing intent. For example, the execution result of analyzing the image in the scene based on the data processing intent. Semantic understanding is performed on the execution result to obtain the data feedback result. The data feedback result can be represented based on natural language description or structured data.

[0078] According to embodiments of this disclosure, the dialogue-based job data processing method further includes: retrieving specified target job data from the target job data based on data requirement information input by the target object, wherein the job description tag of the specified target job data matches the data requirement information.

[0079] In some embodiments, the target task corresponding to the specified target operation data includes at least one of a map data update task and a model training task.

[0080] The map data update task represents updating map data related to the operation scenario; the model training task represents training the algorithm model used to perform the operation.

[0081] Based on the data requirement information input by the target object, specified target task data is retrieved from the target task data. The specified target task data may include map data, model training data, etc., but is not limited to these; the embodiments of this disclosure do not limit the specified target task data.

[0082] Map data can be used to perform map data update tasks, and model training tasks can also be performed based on model training data.

[0083] According to embodiments of this disclosure, by using the data requirement information input by the target object, specified target operation data can be retrieved from the target operation data. Based on the specified target operation data, task updates can be performed, avoiding data silos and enabling data to flow to where it is needed, forming a closed loop of data feedback chain based on dialogue.

[0084] According to embodiments of this disclosure, the target job data includes job data, job description tags, and evaluation information. The evaluation information is used to evaluate the data quality of the target job data. Based on the data demand information input by the target object, a specified target job data is retrieved from the target job data, including: determining candidate job data from the job data based on the retrieval intent represented by the data demand information and the matching results between the retrieval intent and the job description tags; determining the specified target job data to be pushed to the target object from the multiple candidate job data based on the evaluation information corresponding to each of the multiple candidate job data; preferably, the evaluation information is determined based on at least one of the following methods: determining the evaluation information based on the interactive evaluation information input by the target object regarding the job data; determining the evaluation information based on the retrieval demand level of the job data, where the retrieval demand level represents: the push popularity of pushing the job data as the specified target job data to the target object; and determining the evaluation information based on semantic understanding of the contextual information related to the labeled demand intent.

[0085] Target task data can include task data, task description tags, and evaluation information. The evaluation information can be used to assess the data quality of the target task data. For example, the data quality of the target task data can be scored, with higher scores indicating better or worse data quality. Based on the search intent represented by the data demand information and the matching results between the data and the task description tags, candidate task data can be determined from the task data. Based on the evaluation information corresponding to each of the multiple candidate task data, the candidate task data with the highest evaluation score is selected and designated as the specified target task data to be pushed to the target audience.

[0086] Interactive evaluation information can represent scores directly entered by users through an interactive interface. The degree of retrieval demand can characterize the extent to which task data is pushed to the target audience as specified target task data; this can be determined based on popularity information statistically analyzed from the number of hits in relevant scenario dialogues. Evaluation information can be determined through semantic understanding of contextual information related to the labeled demand intent.

[0087] According to embodiments of this disclosure, the large model interprets the context entered by the user in the dialog box during the annotation process to understand the data quality evaluation of the job data with annotated job description labels, thereby realizing the automatic generation of quality assessment results representing the target object for job data based on the understanding of the large model.

[0088] This publication integrates domain knowledge to create a mining area scenario lexicon (i.e., a vocabulary related to operations). Through language alignment with this lexicon, the barrier to entry for models to understand domain knowledge is lowered. A multi-turn dialogue-based data processing method based on natural language is introduced into the data retrieval, analysis, and processing of autonomous driving scenarios in mining areas. This significantly improves the real-time performance of data processing (real-time tagging), the convenience of interaction, and the depth of natural language-based data processing. It transforms users from passively searching to actively engaging in multi-turn data processing, and converts the usage process (scoring the accuracy of search results and embedding dialogue content into a corpus) into data production and processing.

[0089] Figure 6 A schematic diagram of conversational scene retrieval according to an embodiment of the present disclosure is shown.

[0090] like Figure 6 As shown, based on Figure 4 The feedback result 440 obtained can be used to further retrieve specific target operation data from the target operation data based on the data requirement information input by the target object. For example, data requirement information 601 is "Does the number of obstacles affecting driving in the scene exceed 3 within 5 seconds?". Based on the retrieval intent represented by data requirement information 601 and the matching result between it and the operation description tags, candidate operation data is determined from the operation data; candidate operation data 602 is "No, all 3 obstacles are in non-driving areas". This candidate operation data 602 has corresponding evaluation information 603, which can be, for example, "The above result is inaccurate, give 3 points. Give the key tags for this scene". Based on the above, specific target operation data can be retrieved, and the operation description tags of the specific target operation data can be matched with the data requirement information. The operation description tag 604 is "Daytime, Sandstorm, Crushing Station, Reversing, Queuing, No Risk". The target task can then be executed based on the specific target operation data. The tag map prompt information 605, "Confirm the tag is correct, update the tag library", can also be displayed to prompt relevant personnel to update the tags of the operation data. This disclosure introduces a scoring function for search results and a real-time update function for scene tags, which can improve the accuracy and real-time performance of the system.

[0091] Figure 7 A schematic block diagram of a dialogue-based job data processing apparatus according to an embodiment of the present disclosure is shown.

[0092] like Figure 7 As shown, the dialogue-based job data processing device of this embodiment includes an acquisition module 710, a display module 720, a first determination module 730, and a second determination module 740.

[0093] The acquisition module 710 is used to acquire the initial requirement information input by the target object through a dialogue.

[0094] Display module 720 is used to display the data feedback results that represent the operation data. The operation data represents the operation behavior performed by the operation object in the operation scenario, and the data attributes of the operation data are matched with the initial requirement information.

[0095] The first determining module 730 is used to determine the job description label corresponding to the job data based on the annotation requirements of the target object in response to the data feedback results.

[0096] The second determination module 740 is used to determine the target job data based on the job description label and job data.

[0097] According to embodiments of this disclosure, by acquiring initial demand information and displaying data feedback results that match the initial demand information, the target object can obtain the data attributes of the task data to be labeled based on the data feedback results through natural language description in a dialogue manner. Thus, by inputting the labeling demand intent, the task description label is determined, so as to realize the labeling of the task data to be labeled quickly and accurately through dialogue, thereby improving the labeling accuracy and precision of the generated labeled target task data.

[0098] The annotation requirement intent includes the content requirement intent, which describes the target object's requirements for the data content in the task data.

[0099] The first determining module 730 includes: a semantic understanding unit.

[0100] The semantic understanding unit is used to perform semantic understanding on the task data according to the content requirement intent, and obtain the target task data and the content intent tag representing the content requirement intent in the task data. The data content of the target task data matches the content requirement intent, and the task description tag includes the content intent tag.

[0101] The semantic understanding unit includes: a clarification prompt information subunit and a content intent label subunit.

[0102] The clarification prompt information subunit is used to perform semantic understanding of the task data according to the content requirement intent to obtain clarification prompt information. The clarification prompt information is used to prompt the target object to clarify the missing or semantically ambiguous intent information in the content requirement intent.

[0103] The content intent tag subunit is used to respond to the clarification content input by the target object in response to the clarification prompt information, to perform semantic understanding on the task data based on the clarification content, and to obtain the content intent tag, as well as the target task data in the task data that matches the content intent tag.

[0104] The first determining module 730 includes: a semantic unit and a job description label unit.

[0105] Semantic units are used to semantically understand the intent of annotation requirements, determine the target label attributes and label value data, and the target label attributes indicate the label type of the intent label that the target object needs to be annotated.

[0106] The job description tag unit is used to determine the job description tags related to the job data based on the target tag attribute and tag value data.

[0107] The acquisition module 710 includes: a receiving unit, an output unit, and a second initial demand information unit.

[0108] The receiving unit is used to receive the first initial requirement information input by the target object.

[0109] The output unit is used to perform semantic understanding on the first initial demand information and output first feedback information that matches the demand intent of the first initial demand information. The first feedback information describes the data attributes of the initial job data, which is determined from the preset job dataset based on the first initial demand information.

[0110] The second initial requirement information unit is used to display the first feedback information and receive the second initial requirement information input by the target object in response to the first feedback information.

[0111] The output unit includes: intent understanding subunit, execution subunit, and semantic subunit.

[0112] The intent understanding subunit is used to understand the intent of the first initial requirement information based on a job-related vocabulary related to the job scenario, and to obtain the data attribute requirement intent.

[0113] The execution subunit is used to call the retrieval tool to perform a retrieval task on the preset job database according to the tool parameters determined based on the data attribute requirements, and obtain the initial job data.

[0114] The semantic subunit is used to perform semantic understanding on the initial task data to obtain the first feedback information.

[0115] The second initial requirement information represents the data processing intent for the initial operation data.

[0116] The output unit includes: the execution result acquisition subunit and the data feedback result subunit.

[0117] The execution result sub-unit is used to perform target tasks on the job data in the initial related data that matches the data processing intention, based on the second initial requirement information, and obtain the execution result.

[0118] The data feedback result subunit is used to perform semantic understanding of the execution results and obtain data feedback results.

[0119] The dialogue-based job data processing device of this embodiment includes a retrieval module.

[0120] The retrieval module is used to retrieve specified target job data from the target job data based on the data requirement information input by the target object, wherein the job description tags of the specified target job data are matched with the data requirement information.

[0121] The retrieval module includes an update unit and a training unit.

[0122] The update unit is used for map data update tasks and represents map data related to the update operation scenario.

[0123] The training unit is used for model training tasks, representing the training of an algorithm model for performing job behaviors.

[0124] The target task data includes task data, task description labels, and evaluation information. The evaluation information is used to assess the data quality of the target task data.

[0125] The retrieval module includes: a first determining unit and a second determining unit.

[0126] The first determining unit is used to determine candidate job data from the job data based on the retrieval intent represented by the data demand information and the matching results between the job description tags.

[0127] The second determining unit is used to determine the specified target job data to be pushed to the target object from the multiple candidate job data based on the evaluation information corresponding to each of the multiple candidate job data.

[0128] The second determining unit includes: a first evaluation subunit, a second evaluation subunit, and a semantic annotation subunit.

[0129] The first evaluation subunit is used to determine evaluation information based on the interactive evaluation information of the target object for the operation data input.

[0130] The second evaluation subunit is used to determine evaluation information based on the degree of retrieval demand for the task data. The degree of retrieval demand is characterized by the push popularity of the task data as the specified target task data to the target object.

[0131] The annotation semantic subunit is used to semantically understand and determine the contextual information related to the annotation requirements.

[0132] Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure, or at least part of the functions of any one or more of them, can be implemented in one module. Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be implemented by dividing them into multiple modules. Any one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be at least partially implemented as hardware circuitry, such as a Field-Programmable Gate Array (FPGA), a Programmable Logic Array (PLA), a System-on-Chip, a System-on-a-Substrate, a System-on-Package, an Application-Specific Integrated Circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, one or more of the modules, submodules, units, and subunits according to embodiments of the present disclosure can be at least partially implemented as computer program modules, which, when run, can perform corresponding functions.

[0133] For example, any plurality of the acquisition module 710, display module 720, first determination module 730, and second determination module 740 can be combined into one module / unit / subunit, or any one of these modules / units / subunits can be split into multiple modules / units / subunits. Alternatively, at least part of the functionality of one or more of these modules / units / subunits can be combined with at least part of the functionality of other modules / units / subunits and implemented in one module / unit / subunit. According to embodiments of this disclosure, at least one of the acquisition module 710, display module 720, first determination module 730, and second determination module 740 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or any other reasonable means of integrating or packaging the circuitry, or implemented in software, hardware, or firmware, or in any suitable combination of any of these three implementation methods. Alternatively, at least one of the acquisition module 710, display module 720, first determination module 730 and second determination module 740 may be implemented at least partially as a computer program module, which can perform corresponding functions when the computer program module is run.

[0134] Figure 8 A block diagram of an electronic device suitable for implementing the methods described above, according to embodiments of the present disclosure, is illustrated schematically. Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0135] like Figure 8 As shown, an electronic device 800 according to an embodiment of the present disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in ROM 802 (Read-Only Memory) or a program loaded from storage portion 808 into RAM 803 (Random Access Memory). The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0136] RAM 803 stores various programs and data required for the operation of electronic device 800. Processor 801, ROM 802, and RAM 803 are interconnected via bus 804. Processor 801 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 802 and / or RAM 803. It should be noted that the programs may also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.

[0137] According to embodiments of this disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to a bus 804. The electronic device 800 may also include one or more of the following components connected to the input / output (I / O) interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 810 as needed so that computer programs read from it can be installed into the storage section 808 as needed.

[0138] According to embodiments of this disclosure, the method flow according to embodiments of this disclosure can be implemented as a computer software program. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication section 809, and / or installed from removable medium 811. When the computer program is executed by processor 801, it performs the functions defined in the system of embodiments of this disclosure. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0139] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs that, when executed, implement the method according to the embodiments of this disclosure.

[0140] According to embodiments of this disclosure, the computer-readable storage medium can be a non-volatile computer-readable storage medium. Examples include, but are not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0141] For example, according to embodiments of this disclosure, a computer-readable storage medium may include the ROM 802 and / or RAM 803 described above and / or one or more memories other than ROM 802 and RAM 803.

[0142] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods provided in the embodiments of this disclosure. When the computer program product is run on an electronic device, the program code enables the electronic device to implement the dialog-based job data processing method provided in the embodiments of this disclosure.

[0143] When the computer program is executed by the processor 801, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.

[0144] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 809, and / or installed from a removable medium 811. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.

[0145] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on a user's computing device, partially on a user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0146] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions. Those skilled in the art will understand that the features described in the various embodiments of the present disclosure can be combined and / or combined in various ways, even if such combinations are not explicitly described in the present disclosure. In particular, the features described in the various embodiments of this disclosure may be combined and / or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.

[0147] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.

Claims

1. A dialogue-based method for processing job data, comprising: Obtain the initial requirements information input by the target object through a dialogue; The display shows the data feedback results representing the work data, which represents the work behavior performed by the work object in the work scenario, and the data attributes of the work data are matched with the initial requirement information; Based on the target object's annotation requirements in response to the data feedback results, determine the job description tags corresponding to the job data; and The target job data is determined based on the job description label and the job data.

2. The method according to claim 1, wherein, The annotation requirement intent includes the content requirement intent, which describes the target object's requirements for the data content in the task data; Specifically, based on the annotation requirements input by the target object in response to the data feedback results, determining the job description tags corresponding to the job data includes: Based on the content requirement intent, the task data is semantically understood to obtain the target task data and the content intent tag representing the content requirement intent. The data content of the target task data matches the content requirement intent, and the task description tag includes the content intent tag.

3. The method according to claim 2, wherein, The semantic understanding of the task data based on the content requirement intent includes: Based on the stated content requirement intent, the task data is semantically understood to obtain clarification prompts. These clarification prompts are used to prompt the target object to clarify any missing or semantically ambiguous intent information in the stated content requirement intent; and In response to obtaining the clarification content input by the target object in response to the clarification prompt information, the task data is semantically understood based on the clarification content to obtain the content intent tag and the target task data in the task data that matches the content intent tag.

4. The method according to claim 1, wherein, The step of determining the job description tag corresponding to the job data based on the annotation requirement intent input by the target object in response to the data feedback result includes: The semantic understanding of the stated annotation requirement intent is used to determine the target tag attributes and tag value data, wherein the target tag attributes represent the tag type of the intended tag that the target object needs to be annotated; and Based on the target tag attribute and the tag value data, determine the job description tag related to the job data.

5. The method according to claim 1, wherein, The process of obtaining the initial requirement information input by the target object through dialogue includes: Receive the first initial requirement information input by the target object; The first initial demand information is semantically understood, and a first feedback information matching the demand intent of the first initial demand information is output. The first feedback information describes the data attributes of the initial task data, which is determined from a preset task dataset based on the first initial demand information. Display the first feedback information and receive the second initial requirement information input by the target object in response to the first feedback information.

6. The method according to claim 5, wherein, The step of semantically understanding the first initial demand information and outputting first feedback information that matches the demand intent of the first initial demand information includes: Based on the job-related vocabulary list related to the job scenario, the first initial requirement information is subjected to intent understanding to obtain the data attribute requirement intent. The retrieval tool is invoked to perform a retrieval task on a preset job database based on tool parameters determined according to the data attribute requirements, thereby obtaining initial job data; and The initial task data is semantically understood to obtain the first feedback information.

7. The method according to claim 5, wherein, The second initial demand information represents the data processing intention for the initial job data, and the data feedback result is determined based on the following operations: Based on the second initial requirement information, the target task is executed on the job data in the initial related data that matches the data processing intention, and the execution result is obtained; The execution result is semantically understood to obtain the data feedback result.

8. The method according to claim 1, further comprising: Based on the data requirement information input by the target object, retrieve specified target job data from the target job data, wherein the job description tag of the specified target job data matches the data requirement information; Preferably, the specified target task data corresponds to at least one of the following target tasks: The map data update task represents updating the map data related to the aforementioned work scenario; The model training task represents the training of an algorithm model used to perform the job behavior.

9. The method according to claim 8, wherein, The target job data includes the job data, the job description tags, and evaluation information, wherein the evaluation information is used to evaluate the data quality of the target job data. The process of retrieving specified target job data from the target job data based on the data requirement information input by the target object includes: Based on the search intent represented by the data demand information and the matching results between the job description tags, candidate job data are determined from the job data; Based on the evaluation information corresponding to each of the multiple candidate job data, the specified target job data to be pushed to the target object is determined from the multiple candidate job data; Preferably, the evaluation information is determined based on at least one of the following methods: The evaluation information is determined based on the interactive evaluation information input by the target object in response to the operation data; The evaluation information is determined based on the degree of retrieval demand for the job data, wherein the degree of retrieval demand represents the push popularity of the job data as the specified target job data to the target object; The semantic understanding of the contextual information related to the stated annotation requirements is used to determine the requirements.

10. A dialogue-based job data processing device, comprising: The acquisition module is used to acquire the initial requirement information input by the target object through a dialogue. The display module is used to display the data feedback results that represent the operation data, which represents the operation behavior performed by the operation object in the operation scenario, and the data attributes of the operation data are matched with the initial requirement information. The first determining module is used to determine the job description tag corresponding to the job data based on the annotation requirement intention input by the target object in response to the data feedback result; as well as The second determining module is used to determine the target job data based on the job description label and the job data.

Images