Schedule data processing method and apparatus, electronic device, and storage medium

Abstract

The application provides a schedule data processing method and device, an electronic device and a storage medium. The method comprises the following steps: in response to a natural language instruction initiated by a target object, performing intent recognition on the natural language instruction to obtain a target schedule processing intent; when the target schedule processing intent represents a schedule setting intent, calling a preset large language model to perform semantic analysis on the natural language instruction to obtain to-do event description information and time description information corresponding to a to-be-set schedule; calling the large language model to perform logical conflict checking on the time description information, and obtaining target time information according to the conflict checking result and the time description information; matching a target schedule reminding strategy corresponding to the to-be-set schedule based on the to-do event description information; generating structured label data according to the target schedule reminding strategy, the target time information and the to-do event description information, and generating target schedule data based on the structured label data. The application can improve the intelligent level of schedule management.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a method and apparatus for processing schedule data, an electronic device, and a storage medium. Background Technology

[0002] With the development of Internet technology, smart terminals have been deeply integrated into people's daily lives, becoming an important tool for information acquisition and business management, providing great convenience for people's daily lives, and making people increasingly dependent on smart terminals for work and life schedules.

[0003] In related technologies, scheduling methods that rely on graphical user interfaces typically require users to manually complete operations such as setting time and inputting content, which are tedious and time-consuming. In addition, although voice assistants on smart terminals have basic command recognition capabilities, their ability to understand spoken and semantically ambiguous natural language commands is limited, making it difficult to accurately grasp the user's actual scheduling needs.

[0004] Therefore, how to accurately understand users' natural language commands and generate accurate schedule data to improve the intelligence level of schedule management has become an urgent technical problem to be solved. Summary of the Invention

[0005] The main objective of this application is to provide a schedule data processing method, apparatus, electronic device, and storage medium, which aims to accurately understand the user's natural language instructions, generate accurate schedule data, and improve the intelligence level of schedule management.

[0006] To achieve the above objectives, a first aspect of this application proposes a schedule data processing method, the method comprising:

[0007] In response to a natural language instruction initiated by the target object, and by performing intent recognition on the natural language instruction, the target schedule processing intent is obtained; When the target schedule processing intent represents the schedule setting intent, a preset large language model is invoked to perform semantic parsing on the natural language instruction to obtain at least one to-do event description information and time description information corresponding to the schedule to be set. The large language model is invoked to perform logical conflict verification on the time description information to obtain the conflict verification result, and the target time information is obtained based on the conflict verification result and the time description information. Based on the description information of the to-do event, match the target schedule reminder strategy corresponding to the schedule to be set among a number of preset schedule reminder strategies; Based on the target schedule reminder strategy, the target time information, and the corresponding to-do event description information, structured tag data is generated for each schedule to be set, and corresponding target schedule data is generated based on the structured tag data.

[0008] In some embodiments, the step of invoking a preset large language model to perform semantic parsing on the natural language instruction to obtain at least one to-do event description and time description information corresponding to the schedule to be set includes: Determine at least one schedule to be set in the natural language instructions, and call a preset large language model to perform pending event feature recognition on the natural language instructions to obtain the pending event description information corresponding to each scheduled event to be set. The large language model is invoked to perform temporal semantic parsing on the natural language instruction to obtain the time entity information and time context information corresponding to each schedule to be set in the natural language instruction. The time description information corresponding to the schedule to be set is generated based on the time entity information and the time context information.

[0009] In some embodiments, generating time description information corresponding to the schedule to be set based on the time entity information and the time context information includes: When the start and end status identifier text corresponding to the schedule to be set is detected in the natural language instruction, the schedule start time and schedule end time associated with the start and end status identifier text are obtained. The start and end status identifier text is used to associate the event start status and event end status of the to-do event description information corresponding to the schedule to be set. The duration range corresponding to the schedule to be set is determined based on the schedule start time and the schedule end time. The time description information corresponding to the schedule to be set is generated based on the time entity information, the time context information, and the duration range.

[0010] In some embodiments, the step of invoking the large language model to perform logical conflict verification on the time description information, obtaining a conflict verification result, and obtaining target time information based on the conflict verification result and the time description information includes: Obtain the current time information, and standardize the time format according to the current time information, the time entity information, the time context information and the duration range to obtain the standard time information corresponding to the schedule to be set. The current time information is used to represent the time when the natural language instruction is received. The large language model is invoked to perform time validity verification on the standard time information based on the current time information, and the time validity verification result is obtained. When the time validity verification result indicates that there is a time conflict between the current time information and the standard time information, a time conflict prompt message is generated based on the current time information and the standard time information; Obtain the time correction information fed back by the target object based on the time conflict prompt information, and correct the standard time information based on the time correction information to obtain the target time information.

[0011] In some embodiments, matching the target schedule reminder strategy corresponding to the schedule to be set from a preset plurality of schedule reminder strategies based on the to-do event description information includes: Obtain the schedule setting scenario information corresponding to the to-do event description information, and determine the corresponding target reminder time offset value from a set of preset reminder time offset values ​​based on the schedule setting scenario information. The time corresponding to the target time information is later than the trigger execution time of the schedule reminder operation. The reminder time offset value represents the time interval between the trigger execution time of the schedule reminder operation and the target time information. The large language model is invoked to generate corresponding schedule reminder content based on the schedule setting scenario; The target schedule reminder strategy is generated based on the target reminder time offset value and the schedule reminder content.

[0012] In some embodiments, generating structured tag data corresponding to each schedule to be set based on the target schedule reminder strategy, the target time information, and the corresponding to-do event description information further includes: The large language model is invoked to identify the intent correction identifier text in the natural language instruction. When the intent correction identifier text is detected in the natural language instruction, the to-do event description information, time description information and target schedule reminder strategy of the corresponding schedule to be set are updated according to the intent correction identifier text, so as to obtain the updated to-do event description information, updated time description information and updated schedule reminder strategy. The update structured tag data corresponding to the schedule to be set is generated based on the update to-do event description information, the update time description information, and the update schedule reminder strategy.

[0013] In some embodiments, the schedule data processing method provided in this application further includes: When the target schedule processing intent represents a schedule query intent, at least one of the query time information and query event information corresponding to the schedule to be queried in the natural language instruction is extracted; The target query schedule data is determined from the created schedule data based on at least one of the query time information and the query event information.

[0014] To achieve the above objectives, a second aspect of this application provides a schedule data processing apparatus, the apparatus comprising: A schedule data processing system, the schedule data processing system being used to perform the schedule data processing method as described in the first aspect; The smart glasses are used to receive natural language commands initiated by a target object and send them to the schedule data processing system, and then receive and display the target schedule data generated by the schedule data processing system based on the natural language commands.

[0015] To achieve the above objectives, a third aspect of the present application provides an electronic device, the electronic device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method described in the first aspect.

[0016] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0017] The schedule data processing method, apparatus, electronic device, and storage medium proposed in this application accurately understand the target object's intention to process the schedule by responding to natural language commands initiated by the target object and performing intent recognition. Then, a large language model is used to perform deep semantic parsing of the natural language commands, accurately extracting the description information of the to-do events and time descriptions for the schedule to be set. Furthermore, the large language model is invoked to perform logical conflict verification on the time description information, avoiding logical fallacies and time conflicts in the schedule time logic of the natural language commands. Simultaneously, personalized schedule reminder strategies are matched based on the to-do event description information, and the target schedule data is generated by integrating the above information. Thus, this application can accurately understand natural language commands, generate accurate target schedule data, and improve the level of intelligence in schedule management. Attached Figure Description

[0018] Figure 1 This is a flowchart of the schedule data processing method provided in the embodiments of this application; Figure 2 This is provided by the embodiments of this application. Figure 1 The flowchart of step S102 in the document; Figure 3 This is provided by the embodiments of this application. Figure 2 The flowchart of step S203 in the process; Figure 4 This is provided by the embodiments of this application. Figure 1 The flowchart of step S103 in the process; Figure 5 This is provided by the embodiments of this application. Figure 1 The flowchart of step S104 in the process; Figure 6 This is provided by the embodiments of this application. Figure 1 The flowchart of step S105 in the process; Figure 7 This is a schematic diagram of the structure of the schedule data processing device provided in the embodiments of this application; Figure 8 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0020] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0022] With the development of Internet technology, smart terminals have been deeply integrated into people's daily lives, becoming an important tool for information acquisition and business management, providing great convenience for people's daily lives, and making people increasingly dependent on smart terminals for work and life schedules.

[0023] In related technologies, scheduling methods that rely on graphical user interfaces typically require users to manually complete operations such as setting time and inputting content. For example, they need to manually select the date, set a specific time, input the content of the to-do event, and manually select the reminder offset value (such as 5 minutes or 2 hours in advance). These operations are tedious and time-consuming. In addition, although the voice assistants of smart terminals have basic command recognition capabilities, their ability to understand spoken and semantically ambiguous natural language commands is limited, making it difficult to accurately grasp the user's actual scheduling needs.

[0024] Furthermore, the scheduling methods in related technologies often mechanically execute scheduling instructions, failing to determine whether the time mentioned by the user is valid, resulting in invalid schedule reminders. In addition, the reminder strategies are simplistic and cannot automatically generate appropriate reminder strategies based on the nature of the to-do event, leading to a poor user experience.

[0025] Therefore, how to accurately understand users' natural language commands and generate accurate schedule data to improve the intelligence level of schedule management has become an urgent technical problem to be solved.

[0026] Based on this, embodiments of this application provide a schedule data processing method and apparatus, electronic device and storage medium, which aim to accurately understand the user's natural language instructions, generate accurate schedule data, and improve the intelligence level of schedule management.

[0027] The schedule data processing method, apparatus, electronic device, and storage medium provided in this application are specifically described through the following embodiments. First, the schedule data processing method in this application is described.

[0028] The schedule data processing method provided in this application relates to the field of artificial intelligence technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the schedule data processing method, but is not limited to the above forms.

[0029] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0030] Figure 1 This is an optional flowchart of the schedule data processing method provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0031] Step S101: In response to the natural language instruction initiated by the target object, perform intent recognition on the natural language instruction to obtain the target schedule processing intent; Step S102: When the target schedule processing intent represents the schedule setting intent, the preset big language model is called to perform semantic parsing on the natural language instruction to obtain at least one to-do event description information and time description information corresponding to the schedule to be set. Step S103: Call the large language model to perform logical conflict verification on the time description information, obtain the conflict verification result, and obtain the target time information based on the conflict verification result and the time description information; Step S104: Based on the description information of the to-do event, match the target schedule reminder strategy corresponding to the schedule to be set from multiple preset schedule reminder strategies; Step S105: Generate structured tag data for each schedule to be set based on the target schedule reminder strategy, target time information and corresponding to-do event description information, and generate corresponding target schedule data based on the structured tag data.

[0032] Steps S101 to S105 of this embodiment accurately understand the target object's intention to process the target schedule by responding to natural language commands initiated by the target object and performing intent recognition. Then, a large language model is used to perform deep semantic parsing of the natural language commands, accurately extracting the description information of the to-do events and time descriptions for the schedule to be set. Furthermore, the large language model is invoked to perform logical conflict verification on the time description information, avoiding logical fallacies and time conflicts in the schedule time within the natural language commands. Simultaneously, a personalized schedule reminder strategy is matched based on the to-do event description information, and the target schedule data is generated by combining the above information. Thus, this application can accurately understand natural language commands, generate accurate target schedule data, and improve the intelligence level of schedule management.

[0033] In step S101 of some embodiments, in response to a natural language instruction initiated by the target object, intent recognition is performed on the natural language instruction to obtain the target schedule processing intent. This step is executed by the intent determination and triage module in the schedule data processing system. Specifically, the natural language instruction is used to represent an unstructured instruction issued by the target object in the form of voice or text, such as "Set an alarm for me to go to Shanghai tomorrow" input by the target object via voice. After receiving the natural language instruction, the intent determination and triage module analyzes it. The intent determination and triage module can perform intent recognition on the natural language instruction to determine the target schedule processing intent. Intent recognition can be performed using a large language model, or by fine-tuning a lightweight pre-trained model such as BERT or TextCNN (to achieve convergence on a classification task in a specific domain), or by using keywords or regular expressions. There are no limitations here.

[0034] The target schedule processing intent represents the type of operation the target object wants to perform. Target schedule processing intents can include schedule query intents and schedule setting intents. The schedule query intent represents the target object's intention to query or retrieve schedule data. Schedule query intents are determined through a preset query protocol, which is a set of logical rules used by the schedule data processing system to determine whether a natural language command aims to retrieve existing schedule information. The query protocol can distinguish between schedule queries and intelligent assistant function inquiries. The schedule setting intent represents the intention to create or modify schedule data. Schedule setting intents are determined through a preset setting protocol, which is a set of logical rules used by the schedule processing system to determine whether a natural language command creates or modifies a schedule. Through this step, the schedule data processing system can accurately filter out irrelevant chatter or query requests, ensuring that commands with a clear schedule setting intent enter the corresponding semantic parsing process.

[0035] Thus, by performing semantic parsing on natural language instructions, this application improves the accuracy of triggering the corresponding schedule data processing flow by identifying the semantic features in the natural language instructions and determining the target schedule processing intent.

[0036] In some embodiments, the schedule data processing method provided in this application further includes the following steps: When the target schedule processing intent represents the schedule query intent, extract at least one of the query time information and query event information corresponding to the schedule to be queried from the natural language instruction; The target query schedule data is determined from the existing schedule data based on at least one of the query time information and query event information.

[0037] In this embodiment, when the target schedule processing intent represents a schedule query intent, at least one of the query time information and query event information corresponding to the schedule to be queried can be extracted from the natural language instruction. Specifically, schedule query conditions can be extracted from the natural language instruction first. Schedule query conditions include at least one of query time information and query event information. The query time information includes the retrieval time range defined in the natural language instruction, such as "check tomorrow's schedule," where "tomorrow" is the query time information that the schedule data processing system needs to parse. The query event information can include keywords used to describe the content of the schedule event, such as "when is the meeting," where "meeting" is the query event information. The schedule data processing system can support single-dimensional queries (queries based only on query time information or only on query event information) and also supports combined-dimensional queries.

[0038] Then, based on at least one of the query time information and query event information, the target query schedule data can be determined from the existing schedule data. The schedule data processing system uses the parsed schedule query conditions to perform a matching operation in a preset database. The existing schedule data includes structured data that has been parsed, verified, and stored by the schedule processing system. It is usually stored in the database in the form of tags. For example, the structured tag data corresponding to the existing schedule can be... <date> XXXX-XX-XX< / date> <content> Project Review< / content> The schedule data processing system can convert the extracted unstructured query time information and query event information into a retrieval key consistent with the database format (e.g., XXXX-XX-XX). Then, it can compare the retrieved data with the created schedule data, locate and extract one or more schedule records that meet the query conditions, and obtain the target query schedule data.

[0039] This application embodiment can accurately understand the target object's schedule query needs by extracting query time information or query event information from natural language commands, thereby improving the real-time performance and accuracy of query results. It effectively solves the problem of insufficient understanding of voice assistants in related technologies when faced with complex combined queries, thus improving the retrieval efficiency and accuracy of schedule information.

[0040] In step S102 of some embodiments, when the target schedule processing intent represents the schedule setting intent, a preset large language model can be invoked to perform semantic parsing on the natural language instruction to obtain at least one to-do event description information and time description information corresponding to the schedule to be set. The to-do event description information may include event description text after data cleaning and removal of irrelevant interference such as interjections, such as "meeting" or "business trip to Shanghai"; while the time description information includes time-related text descriptions in the natural language instruction, such as "next Friday," "tomorrow morning," or "8:00 AM."

[0041] In some embodiments, please refer to Figure 2 The system invokes a pre-defined large language model to perform semantic parsing on natural language instructions, obtaining at least one to-do event description and time description information corresponding to the schedule to be set, including the following steps S201 to S203: Step S201: Determine at least one schedule to be set in the natural language command, and call the preset large language model to perform pending event feature recognition on the natural language command to obtain the pending event description information corresponding to each pending schedule. Step S202: Call the large language model to perform temporal semantic parsing on the natural language instructions to obtain the time entity information and time context information corresponding to each schedule to be set in the natural language instructions; Step S203: Generate time description information corresponding to the schedule to be set based on time entity information and time context information.

[0042] In step S201 of some embodiments, the natural language instruction may include one or more schedules to be set. At least one schedule to be set in the natural language instruction can be determined first. Specifically, the text corresponding to the natural language instruction can be divided into multiple independent semantic units based on semantic conjunctions (e.g., separators or logical connectors), punctuation, and pairing logic of time and events (e.g., combinations of time and action events or different action predicates), thereby calculating the number of valid to-do items, and then determining the number of schedules to be set based on the number of to-do items. For example, if the natural language instruction is "Remind me to attend an all-staff meeting at 10 AM tomorrow, and also attend a class reunion at 7 PM Friday," the schedule data processing system identifies the conjunction "also," and uses this as a boundary to divide the text into two parts. The first part extracts {Time: Tomorrow 10 AM, Event: All-staff meeting}, and the second part extracts {Time: Friday 7 PM, Event: Class reunion}, then two schedules to be set are determined.

[0043] Simultaneously, a large language model can be invoked to identify the features of natural language instructions for pending events, obtaining the description information of each pending event for each scheduled task. Specifically, the large language model can be used to identify meaningless trigger words or polite expressions in natural language instructions, extracting irrelevant content such as interjections from redundant natural language instruction text. Based on a pre-defined stop word library or the attention mechanism of the large model, these texts are marked as non-event features and removed. Then, the large language model can be invoked to perform syntactic analysis on the text after removing irrelevant content, identifying predicate verbs and objects, as well as specific proper nouns (such as names of people, places, and projects), determining the key text describing specific events or actions, and retaining it as the descriptive features of the event. In this way, the description information of each pending event for each scheduled task can be obtained through the above steps.

[0044] In step S202 of some embodiments, a large language model is invoked to perform temporal semantic parsing on the natural language instruction, obtaining the time entity information and time context information corresponding to each schedule to be set in the natural language instruction. Specifically, the attention mechanism of the large language model can be invoked to separate the time entity information and time context information from the natural language instruction. Among them, the time entity information represents all text content in the natural language instruction that refers to time. The time entity information may include absolute time information (e.g., a specific date XX month XX day), relative time information (e.g., next Wednesday), and periodically expressed time information (e.g., every two weeks on Friday), etc.

[0045] Temporal context information represents the external constraint parameters of a time entity. Specifically, it can include fuzzy time references (morning, evening, early morning, etc.) and time zone information (time context information includes time difference factors). Temporal context information can also include the contextual information of the current schedule to be set and its predecessors (schedules set before the current time). For example, if the previous text mentioned "I'm going to the United States," the "arrival time" in the following text needs to be predicted in conjunction with the flight duration context. Another example is a natural language instruction: "I want to book a post-meeting dinner." The schedule data processing system can identify the temporal context information "post-meeting," and then analyze it to determine the meeting's end time in the predecessors (assuming the meeting ends at 16:00), thus parsing "post-meeting" as "after 16:00." A search is then performed in the predecessors. If no corresponding predecessor is found, a prompt message can be sent to the target. Continuing with the above example, the following prompt message could be sent: "No corresponding meeting schedule found. Please provide the specific meeting end time."

[0046] In step S203 of some embodiments, the time entity information and time context information extracted in the preceding steps can be summarized to generate time description information corresponding to the schedule to be set. At this time, the obtained time description information is an unstructured text string and cannot be directly used to generate the corresponding schedule data. It is necessary to convert the time description information into the corresponding standard structured code.

[0047] This application's embodiments utilize a large language model to deeply analyze and understand natural language instructions, effectively filtering redundant information and accurately identifying all text content describing time and pending events within the natural language instructions. This improves the robustness of the schedule data processing system in setting schedules based on natural language. When faced with ambiguous references or implicit time logic, this application's embodiments can make the most logical judgment, avoiding logical errors caused by mechanically extracting time information without in-depth analysis of time semantics, thus improving the accuracy of schedule management.

[0048] In some embodiments, please refer to Figure 3 The process involves generating time description information for the schedule to be set based on time entity information and time context information, including the following steps S301 to S303: Step S301: When the start and end status identifier text corresponding to the schedule to be set is detected in the natural language instruction, the start time and end time of the schedule associated with the start and end status identifier text are obtained. Step S302: Determine the duration range of the schedule to be set based on the schedule start time and schedule end time; Step S303: Generate time description information corresponding to the schedule to be set based on time entity information, time context information, and duration range.

[0049] In step S301 of some embodiments, a large language model can be invoked to detect the start and end status identifier text corresponding to the schedule to be set in the natural language instruction. The start and end status identifier text represents the start and end states of the event, which are used to associate the description information of the to-do event corresponding to the schedule to be set. It represents the related phrases or sentence structures in the natural language instruction used to define when a to-do item starts and ends. For example, keywords describing the start and end states of the same event, such as "go...back..." or "from...to...". Thus, embodiments of this application can utilize the semantic analysis capabilities of the large language model to identify complex to-do events with a time span or a clear start and end logic.

[0050] When the start and end status identifier text corresponding to the schedule to be set is detected in the natural language instruction, the start time and end time of the schedule associated with the start and end status identifier text can be obtained. The start and end times of the schedule represent the start and end nodes of the pending event in its logical lifecycle. For example, if the natural language instruction is "Meeting from 3 PM to 5 PM", then the start time of the schedule to be set is 3 PM and the end time is 5 PM.

[0051] By recognizing the aforementioned identifier text, the schedule data processing system can understand that the schedule time associated with the start and end status identifier text is not isolated, but rather two time nodes describing the same to-do event. This avoids repeatedly setting to-do items at different times, prevents schedule conflicts, and improves the accuracy of schedule settings.

[0052] In step S302 of some embodiments, the duration range corresponding to the schedule to be set is determined based on the schedule start time and schedule end time. The duration range is used to represent the complete time span covered from the schedule start time to the schedule end time, and then the duration range can be represented in the form of a time interval. The purpose of this step is to convert discrete time points into continuous time period data, ensuring the time integrity and accuracy of the to-do items. If the schedule data processing system detects that there are overlapping intervals in the duration range of the schedule to be set, it can send prompt information to the target object, such as conflicts between the schedules to be set, affecting the normal execution of the plan, and whether modifications are needed.

[0053] In step S303 of some embodiments, time description information corresponding to the schedule to be set can be generated based on time entity information, time context information, and duration range. After extracting time elements such as time entity information, time context information, and duration range through the aforementioned steps, the schedule data processing system can combine all the extracted time elements to obtain a set of time semantic text corresponding to the schedule to be set, i.e., time description information.

[0054] This application's embodiments utilize a large language model to semantically aggregate time information within natural language instructions, effectively solving the technical problem of related technologies being unable to understand time-span events in natural language instructions, thus improving the accuracy of schedule settings. The schedule data processing system can automatically aggregate scattered time points into a single date range based on duration ranges, generating continuous and logically complete schedule records, avoiding the tedious operation of users separately setting schedule start and end times.

[0055] In step S103 of some embodiments, a large language model is invoked to perform logical conflict verification on the time description information, obtaining a conflict verification result. Based on the conflict verification result and the time description information, the target time information is obtained. Logical conflict verification means that the schedule data processing system automatically detects whether there are semantic contradictions or timeliness errors in the natural language instructions. The schedule data processing system obtains a conflict verification result by performing logical conflict verification on the time description information obtained in the aforementioned steps. When the conflict verification result indicates that the time description information has timeliness errors, logical contradictions, or semantic contradictions, or that the time description information is still unclear after semantic parsing, the time description information can be corrected to obtain accurate target time information. When the conflict verification result indicates that the time description information does not have the aforementioned logical conflicts, the time description information can be used as the target time information for the schedule to be set.

[0056] In some embodiments, please refer to Figure 4 The system calls a large language model to perform logical conflict verification on the time description information, obtains the conflict verification result, and obtains the target time information based on the conflict verification result and the time description information, including the following steps S401 to S404: Step S401: Obtain the current time information, and standardize the time format based on the current time information, time entity information, time context information, and duration range to obtain the standard time information corresponding to the schedule to be set. Step S402: Call the large language model to perform time validity verification on the standard time information based on the current time information, and obtain the time validity verification result; Step S403: When the time validity verification result indicates that there is a time conflict between the current time information and the standard time information, a time conflict prompt message is generated based on the current time information and the standard time information; Step S404: Obtain the time correction information fed back by the target object based on the time conflict prompt information, and correct the standard time information based on the time correction information to obtain the target time information.

[0057] In step S401 of some embodiments, current time information is obtained, and time format standardization is performed based on the current time information, time entity information, time context information, and duration range to obtain standard time information corresponding to the schedule to be set. The current time information is used to characterize the time of the schedule data processing system when the natural language instruction is received, and can serve as reference time information for validating the time description information. Time format standardization refers to the process of combining the time entity information, time context information, and duration range extracted in previous steps with the current time information to perform absolute time calculation and convert it into a standardized time format.

[0058] Specifically, the time description information can be converted using a 24-hour clock and a proximity principle can be introduced. For example, if the current time is 10:00 and the natural language command contains "8 o'clock," the schedule data processing system will automatically determine it as 08:00 the next day or 20:00 that evening, rather than the past 08:00 this morning. Furthermore, ambiguous or relative times can be converted into a specific YYYY-MM-DD format by combining the current time information. For example, the time entity information "next Friday" can be converted into a specific "YYYY-MM-DD." Standard time information represents the compliant time data generated after the above steps, conforming to computer storage standards (such as ISO 8601). Standardizing time description information can eliminate the ambiguity of natural language and improve the standardization of the time representation in the schedule to be set.

[0059] In step S402 of some embodiments, a large language model can be invoked to perform time validity verification on the standard time information based on the current time information, thereby obtaining a time validity verification result. Specifically, time validity verification refers to the schedule data processing system comparing the standard time information obtained from the aforementioned steps with the current time information in a time sequence. This comparison determines whether the standard time information is valid. Time validity can be judged based on dimensions such as whether the standard time information is earlier than the current time information, whether there is a conflict between the standard time information of multiple schedules to be set, the duration of the pending event, or whether the occurrence time is reasonable (e.g., holding a meeting between 2:00 AM and 3:00 AM is unreasonable). By performing validity verification on the standard time information through the aforementioned steps, a time validity verification result is obtained. The time validity verification result may include a Boolean value or status code generated after the time comparison to characterize whether the standard time information is valid.

[0060] In step S403 of some embodiments, when the time validity verification result indicates a time conflict between the current time information and the standard time information, a time conflict prompt message is generated based on the current time information and the standard time information. Time conflicts can include various conflict types. For example, when the time indicated by the standard time information is before the time indicated by the current time information, the time validity verification result indicates the time information is invalid; or when the duration ranges of two or more schedules to be set overlap, the time validity verification result indicates the time information is invalid; or, when the duration of the schedule to be set or the occurrence time of the pending event is unreasonable, the time validity verification result indicates the time information is invalid.

[0061] When the time validity verification result indicates a time conflict between the current time information and the standard time information, the schedule data processing system can execute corresponding time conflict handling strategies for different time conflict types. Specifically, the schedule data processing system can generate time conflict prompts based on the current time information and the standard time information. Specifically, when the time conflict type is that the standard time information is earlier than the current time information, the schedule data processing system can prevent the generation of the schedule creation protocol and simultaneously call the large language model to generate friendly time conflict prompts. These prompts may include reminding the target that the time has passed and asking if they need to change the time. When the time conflict type is that the current time information (e.g., 2:30 PM) is within or exceeds a vague time period in the natural language instruction (e.g., this morning), the schedule data processing system can automatically switch to follow-up mode and send a time conflict prompt providing a specific time point to the target, such as, "The time you mentioned seems to have passed; do you want to set it for tomorrow morning?".

[0062] In step S404 of some embodiments, after sending a time conflict notification to the target object, the schedule data processing system can obtain time correction information fed back by the target object based on the time conflict notification, and correct the standard time information based on the time correction information to obtain the target time information. The time correction information refers to the new instruction or confirmation response given by the target object after receiving the notification from the schedule data processing system; for example, the target object replies "Oh no, it's tomorrow" or "Yes, change it to evening." The target time information refers to the accurate time data that the schedule data processing system ultimately determines based on the time correction information, ensuring logical validity and meeting the target object's needs.

[0063] This application embodiment intelligently intercepts invalid schedule setting operations by verifying the validity of time, and guides the target object to clarify its intention and correct the time information through time conflict prompts. This ensures that the final generated target time information not only meets the actual needs but also conforms to objective and reasonable time logic, thus avoiding schedule conflicts. It solves the problem of mechanical execution of instructions in the schedule management method of related technologies, which leads to low accuracy of schedule setting, and improves the accuracy and intelligent experience of schedule setting.

[0064] In step S104 of some embodiments, a target schedule reminder strategy corresponding to the schedule to be set can be matched from a set of preset schedule reminder strategies based on the description information of the to-do event. The target schedule reminder strategy refers to the reminder rules predefined by the target schedule data processing system for different event urgency levels, event preparation complexity, and event categories. Specifically, the schedule data processing system can analyze keywords in the description information of the to-do event to match the corresponding schedule reminder strategy. For example, when keywords such as "airplane" and "high-speed rail" are identified, a schedule reminder strategy corresponding to a major transportation event is matched; when keywords such as "meeting" and "dinner" are identified, a schedule reminder strategy corresponding to a regular event is matched. Thus, the schedule reminder strategy of this application embodiment avoids the single reminder mode in related technologies and realizes the intelligence and contextualization of schedule reminder services.

[0065] In some embodiments, please refer to Figure 5 Based on the description information of the to-do event, the system matches the target schedule reminder strategy corresponding to the schedule to be set from multiple preset schedule reminder strategies, including the following steps S501 to S503: Step S501: Obtain the schedule setting scenario information corresponding to the to-do event description information, and determine the corresponding target reminder time offset value from a set of preset reminder time offset values ​​based on the schedule setting scenario information; Step S502: Call the large language model to generate corresponding schedule reminder content based on the schedule setting scenario; Step S503: Generate the target schedule reminder strategy corresponding to the schedule to be set based on the target reminder time offset value and the schedule reminder content.

[0066] In step S501 of some embodiments, in order to ensure that the schedule is not delayed, the schedule data processing system needs to remind the target object before the scheduled time, that is, the time corresponding to the target time information is delayed after the triggering time of the schedule reminder operation. Specifically, the schedule setting scenario information can be extracted from the to-do event description information first. The schedule setting scenario information is used to represent the execution environment of the to-do event, and may include the event execution scenario, the urgency of the event, the action attributes corresponding to the to-do event, and the action content, etc.

[0067] For example, going to the airport is a major transportation scenario, attending a weekly meeting is an office scenario, and setting an alarm clock is a regular daily scenario. For the urgency of the event, for example, the description of the event includes keywords such as urgent, immediate, and important. For the action attributes and action content, for example, the description of the event includes specific actions such as taking an international flight or giving a report, as well as the objects of those actions.

[0068] Then, based on the schedule setting scenario information, the corresponding target reminder time offset value can be determined from a set of preset reminder time offset values. Specifically, the corresponding basic reminder time offset value can be determined first from a set of preset reminder time offset values ​​based on the schedule setting scenario information. The reminder time offset value represents the time interval between the trigger execution time of the schedule reminder operation and the target time information. It refers to the duration preset by the schedule data processing system that the schedule reminder action should be earlier than the actual occurrence time of the to-do event, such as setting the schedule reminder to be issued 20 minutes in advance. The basic reminder time offset value refers to the general reminder time offset value preset for a specific schedule setting scenario. For example, the schedule data processing system presets the basic offset value for major traffic scenarios to be 120 minutes and the basic offset value for ordinary office scenarios to be 5 minutes.

[0069] Furthermore, the large language model can be invoked to generate corresponding weighted coefficients based on the urgency of the event, the action content corresponding to the pending event, and the action attributes, and the basic reminder time offset value determined in the aforementioned steps can be adjusted using the generated weighted coefficients. Specifically, the large language model can be invoked to generate event preparation complexity coefficients based on the action content and action attributes corresponding to the pending event. The event preparation complexity coefficient is a quantitative value used to characterize the cumbersomeness of the actions in the pending event. For example, for taking an international flight, the large language model can determine that customs clearance is required, and the preparation work is complex, generating an event preparation complexity coefficient of 1.1, while for domestic high-speed rail, the event preparation complexity coefficient may be 1.0.

[0070] In other embodiments, a large language model can be invoked to analyze the geographic location or physical movement requirements implicit in the description of the to-do event. Specifically, the large language model can be invoked to identify whether the description of the to-do event contains location terms (such as picking someone up at Pudong Airport or picking up a package downstairs). When a location term is identified in the description of the to-do event, the geographic location movement time can be estimated, and a reminder time buffer value can be automatically calculated. For example, if the large language model identifies an airport in a natural language command, and the airport is relatively remote, a reminder time buffer value, such as 40 minutes, can be generated. The large language model can also be invoked to generate corresponding reminder time buffer values ​​based on the urgency of the event. For example, for a to-do event involving a report or presentation, where it is necessary to check presentation documents and prepare mentally, a reminder time buffer value of 5-10 minutes can be generated.

[0071] In other embodiments, when the schedule data processing system receives a natural language instruction, it converts the instruction into a vector, retrieves the top-N historical schedules with the highest semantic similarity from the target object's historical schedules in the vector database, and then examines the reminder time offset values ​​of these historical schedules that were ultimately confirmed by the target object. The retrieved historical preferences can then be used as context input to the large language model to instruct the large language model to generate the reminder time offset values ​​corresponding to the schedule to be set based on the target object's historical preferences.

[0072] The target reminder time offset refers to the actual advance reminder duration written into the schedule data processing system after adjustments based on the base reminder time offset, the event preparation complexity coefficient, and the reminder time buffer. The target reminder time offset can be obtained by multiplying the base reminder time offset by the event preparation complexity coefficient and then adding the reminder time buffer. For example, if the base reminder time offset is 10 minutes, the event preparation complexity coefficient is 1.5, and the reminder time buffer is 5 minutes, then the calculated target reminder time offset is 20 minutes.

[0073] In step S502 of some embodiments, a large language model is invoked to generate corresponding schedule reminder content based on the schedule setting scenario information, and a target schedule reminder strategy corresponding to the schedule to be set is generated according to the target reminder time offset value and the schedule reminder content. The schedule reminder content refers to emotionally charged text or voice displayed to the target object when the schedule reminder is triggered; for example, for transportation-related schedules, a warm and gentle message such as "Have a safe journey, please be careful" is generated; for alarm clock-related schedules, a concise message such as "Wake up on time" is generated.

[0074] In step S503 of some embodiments, the target schedule reminder strategy corresponding to the schedule to be set can be determined based on the target reminder time offset value generated in the aforementioned steps and the schedule reminder content.

[0075] This application's embodiments achieve dynamic calculation of reminder time offset values ​​and differentiated schedule reminders through a progressive process of scene classification matching, complexity coefficient correction, and emotional expression synthesis. Compared to the mechanical schedule reminder mode in related technologies, this application's embodiments automatically reserve more preparation time based on the complexity and urgency of event preparation. Furthermore, by incorporating scenario-based schedule reminder content, the reminders possess emotional warmth and contextual awareness, enhancing users' sense of security and experience in complex schedule management.

[0076] In step S105 of some embodiments, the schedule data processing system can convert the target schedule reminder strategy, target time information, and corresponding to-do event description information generated in the aforementioned steps into machine-executable structured protocol instructions through a structured protocol mapping module. Here, structured tag data refers to a standardized data format used internally by the schedule data processing system, containing specific marker symbols. Structured tag data can be in XML, JSON, or YAML format, without limitation. Corresponding tags can be set for the information extracted in the above steps, such as those used to store dates. <date>Tags, used to store 24-hour time points or time periods. <time>Tags, storage, and calculated target reminder time offset values <prompt>Tags and storage of cleaned (removed interjections and obsolete content) descriptions of to-do items. <content>Tags, etc. Furthermore, the schedule data processing system can call the calendar API to convert structured tag data into target schedule data, which can represent schedule records in the calendar application of a smart terminal.

[0077] In some embodiments, please refer to Figure 6 Based on the target schedule reminder strategy, target time information, and corresponding to-do event description information, structured tag data corresponding to each schedule to be set is generated, and the process also includes the following steps S601 to S602: Step S601: Call the large language model to identify the intent correction identifier text in the natural language instruction. When the intent correction identifier text is detected in the natural language instruction, update the to-do event description information, time description information and target schedule reminder strategy of the corresponding schedule to be set according to the intent correction identifier text, and obtain the updated to-do event description information, updated time description information and updated schedule reminder strategy. Step S602: Generate updated structured tag data corresponding to the schedule to be set based on the update to-do event description information, update time description information, and update schedule reminder strategy.

[0078] In step S601 of some embodiments, a large language model can be invoked to identify intent correction marker text in natural language instructions. When intent correction marker text is detected in a natural language instruction, the to-do event description information, time description information, and target schedule reminder strategy of the corresponding schedule to be set can be updated according to the intent correction marker text, resulting in updated to-do event description information, updated time description information, and updated schedule reminder strategy. The schedule data processing system can perform self-correction judgment. Specifically, it can first identify intent correction marker text in natural language instructions. Intent correction marker text refers to text that negates a preceding instruction or indicates a change in schedule information. For example, if the natural language instruction is "Meeting at 9:00 AM tomorrow... no, change it to 10:00 AM," then "no," "change it to," or "cancel the previous statement," etc., are intent correction marker text. Upon recognizing an intent to modify the event description, the schedule data processing system can update either the event description or the time description. If the intent to modify indicates an adjustment to the event description, the system updates the event description; similarly, if it indicates an adjustment to the time description, the system updates the time description. For example, the system might mark the original "9:00" as obsolete and set "10:00" as the effective time based on the intent to modify. If the intent to modify indicates an adjustment to either the event description or the time description, the system can update the reminder policy accordingly, re-matching the reminder rules as the event or time changes. For instance, if a natural language instruction changes a regular meeting to a flight schedule, the system can not only update the event description but also update the reminder policy from 5 minutes to 2 hours in advance.

[0079] In step S602 of some embodiments, updating the structured tag data means that after the schedule setting intention is modified, the standard instruction code generated by remapping is used for calling the underlying API. Specifically, the schedule data processing system can update the structured tag data before the intention change according to the updated to-do event description information, the updated time description information, and the updated schedule reminder strategy to obtain the updated structured tag data corresponding to the schedule to be set.

[0080] The embodiments of this application can automatically filter out redundant information in the user's spoken language, accurately identify the text content in natural language commands that indicate the intention to change schedule settings, and automatically correct the corresponding tag data according to the intention change information, so as to ensure that the final generated schedule data accurately matches the user's true intention and improve the level of intelligence in schedule management.

[0081] Please see Figure 7 This application also provides a schedule data processing apparatus, which includes: A schedule data processing system, which is used to execute the aforementioned schedule data processing method; Smart glasses are used to receive natural language commands initiated by a target object and send them to a schedule data processing system. The system then receives and displays the target schedule data generated by the natural language commands.

[0082] In some embodiments, a schedule data processing system refers to a collection of software and hardware deployed on a server or high-performance computing device. It integrates the reasoning capabilities of a large language model and logical verification algorithms, and is responsible for executing the aforementioned schedule data processing methods. The schedule data processing system can be a distributed server cluster running in the cloud, internally deploying a large language model specifically for fine-tuning time semantics and a set of logical verification scripts written in Python. The schedule data processing system can receive raw signals from front-end devices, perform a series of schedule data processing operations such as intent recognition, time normalization, conflict verification, and policy matching, and finally output structured tagged data corresponding to the schedule to be set.

[0083] Smart glasses are used to receive natural language commands initiated by the target user and send them to the schedule data processing system. The system then receives and displays the target schedule data generated based on the natural language commands. Smart glasses are a front-end interactive terminal, a wearable head-mounted device with independent processing capabilities, a wireless communication module, and an optical display module. They free the target user's hands, enabling augmented reality (AR) interaction. For example, when wearing the glasses, the target user can utter commands, and the smart glasses' microphone array can collect the voice. For instance, the target user could say while walking, "Set an alarm for me to go to Shanghai tomorrow." The target schedule data refers to structured visual information returned by the back-end processing system and adapted to the glasses' display area. For example, within seconds of the target user uttering the command, a semi-transparent card will appear on the glasses lenses, displaying "Created: Heading to Hongqiao Airport tomorrow at 08:00 (120-minute advance reminder)," for the target user to confirm intuitively.

[0084] This application embodiment uses smart glasses as an input / output port, combined with the processing results of a schedule data processing system, to achieve a convenient schedule management experience. In complex scenarios such as driving, exercising, or when both hands are occupied, the target user can complete high-precision schedule management simply through voice. The schedule data processing system utilizes the near-eye display characteristics of smart glasses to provide the generated schedule data to the target user in the most intuitive visual form, greatly improving the efficiency of schedule management and solving the problems of cumbersome operation and unsafe operation while mobile associated with traditional apps.

[0085] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described schedule data processing method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0086] Please see Figure 8 , Figure 8 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 801 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 802 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 802 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 802 and is called and executed by the processor 801 using the schedule data processing method of the embodiments of this application. The 803 input / output interface is used to implement information input and output. The communication interface 804 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 805 transmits information between various components of the device (e.g., processor 801, memory 802, input / output interface 803, and communication interface 804); The processor 801, memory 802, input / output interface 803, and communication interface 804 are connected to each other within the device via bus 805.

[0087] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described schedule data processing method.

[0088] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0089] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0090] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0091] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0092] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0093] The terms "first," "second," "third," "fourth," etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0094] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0096] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0097] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0098] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0099] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.< / content> < / prompt> < / time> < / date>

Claims

1. A method for processing schedule data, characterized in that, The method includes: In response to a natural language instruction initiated by the target object, and by performing intent recognition on the natural language instruction, the target schedule processing intent is obtained; When the target schedule processing intent represents the schedule setting intent, a preset large language model is invoked to perform semantic parsing on the natural language instruction to obtain at least one to-do event description information and time description information corresponding to the schedule to be set. The large language model is invoked to perform logical conflict verification on the time description information to obtain the conflict verification result, and the target time information is obtained based on the conflict verification result and the time description information. Based on the description information of the to-do event, match the target schedule reminder strategy corresponding to the schedule to be set among a number of preset schedule reminder strategies; Based on the target schedule reminder strategy, the target time information, and the corresponding to-do event description information, structured tag data is generated for each schedule to be set, and corresponding target schedule data is generated based on the structured tag data.

2. The method according to claim 1, characterized in that, The process involves invoking a pre-defined large language model to perform semantic parsing on the natural language instructions, obtaining at least one to-do event description and time description information corresponding to the schedule to be set, including: Determine at least one schedule to be set in the natural language instructions, and call a preset large language model to perform pending event feature recognition on the natural language instructions to obtain the pending event description information corresponding to each scheduled event to be set. The large language model is invoked to perform temporal semantic parsing on the natural language instruction to obtain the time entity information and time context information corresponding to each schedule to be set in the natural language instruction. The time description information corresponding to the schedule to be set is generated based on the time entity information and the time context information.

3. The method according to claim 2, characterized in that, The step of generating time description information corresponding to the schedule to be set based on the time entity information and the time context information includes: When the start and end status identifier text corresponding to the schedule to be set is detected in the natural language instruction, the schedule start time and schedule end time associated with the start and end status identifier text are obtained. The start and end status identifier text is used to associate the event start status and event end status of the to-do event description information corresponding to the schedule to be set. The duration range corresponding to the schedule to be set is determined based on the schedule start time and the schedule end time. The time description information corresponding to the schedule to be set is generated based on the time entity information, the time context information, and the duration range.

4. The method according to claim 3, characterized in that, The process of calling the large language model to perform logical conflict verification on the time description information, obtaining conflict verification results, and obtaining target time information based on the conflict verification results and the time description information includes: Obtain the current time information, and standardize the time format according to the current time information, the time entity information, the time context information and the duration range to obtain the standard time information corresponding to the schedule to be set. The current time information is used to represent the time when the natural language instruction is received. The large language model is invoked to perform time validity verification on the standard time information based on the current time information, and the time validity verification result is obtained. When the time validity verification result indicates that there is a time conflict between the current time information and the standard time information, a time conflict prompt message is generated based on the current time information and the standard time information; Obtain the time correction information fed back by the target object based on the time conflict prompt information, and correct the standard time information based on the time correction information to obtain the target time information.

5. The method according to claim 1, characterized in that, The step of matching the target schedule reminder strategy corresponding to the schedule to be set from a preset set of multiple schedule reminder strategies based on the description information of the to-do event includes: Obtain the schedule setting scenario information corresponding to the to-do event description information, and determine the corresponding target reminder time offset value from a set of preset reminder time offset values ​​based on the schedule setting scenario information. The time corresponding to the target time information is later than the trigger execution time of the schedule reminder operation. The reminder time offset value represents the time interval between the trigger execution time of the schedule reminder operation and the target time information. The large language model is invoked to generate corresponding schedule reminder content based on the schedule setting scenario; The target schedule reminder strategy is generated based on the target reminder time offset value and the schedule reminder content.

6. The method according to claim 1, characterized in that, The step of generating structured tag data for each schedule to be set based on the target schedule reminder strategy, the target time information, and the corresponding to-do event description information further includes: The large language model is invoked to identify the intent correction identifier text in the natural language instruction. When the intent correction identifier text is detected in the natural language instruction, the to-do event description information, time description information and target schedule reminder strategy of the corresponding schedule to be set are updated according to the intent correction identifier text, so as to obtain the updated to-do event description information, updated time description information and updated schedule reminder strategy. The update structured tag data corresponding to the schedule to be set is generated based on the update to-do event description information, the update time description information, and the update schedule reminder strategy.

7. The method according to claim 1, characterized in that, The method further includes: When the target schedule processing intent represents a schedule query intent, at least one of the query time information and query event information corresponding to the schedule to be queried in the natural language instruction is extracted; The target query schedule data is determined from the created schedule data based on at least one of the query time information and the query event information.

8. A schedule data processing device, characterized in that, The device includes: A schedule data processing system, the schedule data processing system being used to perform the schedule data processing method as described in any one of claims 1 to 7; The smart glasses are used to receive natural language commands initiated by a target object and send them to the schedule data processing system, and then receive and display the target schedule data generated by the schedule data processing system based on the natural language commands.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the schedule data processing method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the schedule data processing method according to any one of claims 1 to 7.

Images