Power work order abstract generation method and device, terminal equipment and storage medium

Abstract

The invention discloses an abstract generation method and device for an electric power work order, terminal equipment and a storage medium. The method comprises the steps of obtaining a to-be-processed electric power work order; inputting the to-be-processed power work order into a power work order abstract generation model so as to enable the power work order abstract generation model to determine the business type of the to-be-processed power work order, and inputting the to-be-processed power work order into a target power work order abstract generation sub-model, enabling the target electric power work order abstract generation sub-model to output the key information of the to-be-processed electric power work order, and outputting the abstract content of the to-be-processed electric power work order according to the key information and the to-be-processed electric power work order; wherein the electric power work order abstract generation model comprises a plurality of electric power work order abstract generation sub-models, and each electric power work order abstract generation sub-model corresponds to a business type; the business type of the target electric power work order abstract generation sub-model is the same as the business type of the to-be-processed electric power work order. According to the invention, the generation accuracy of the power work order abstract can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of natural language processing, and in particular to a method, device, terminal equipment and storage medium for generating a summary of an electric power work order. Background Art

[0002] As for the summary generation method of power work orders in the power industry, the existing technology mainly inputs the power work orders into general natural language processing models, such as GPT-3, GPT4 or ChatGPT, so that these models output the summary content in the power work orders. However, these general natural language processing models use large-scale general text data during training to understand and generate various natural language texts. These models are not specific to any particular industry or field, but are designed to handle a wide range of NLP tasks. Therefore, when these general natural language processing models process power work orders and generate summaries, it is difficult to accurately identify content related to the power industry, resulting in the lack of accuracy and credibility of the final generated summary content. Therefore, there is an urgent need for a summary generation method that can be used to process power work orders to improve the accuracy of power work order summary generation. Summary of the invention

[0003] The embodiments of the present invention provide a method, an apparatus, a terminal device and a storage medium for generating a summary of an electric power work order, which can improve the accuracy of generating a summary of an electric power work order.

[0004] An embodiment of the present invention provides a method for generating a summary of an electric power work order, comprising:

[0005] Get pending power work orders;

[0006] The power work order to be processed is input into the power work order summary generation model so that the power work order summary generation model determines the business type of the power work order to be processed, and the power work order to be processed is input into the target power work order summary generation sub-model so that the target power work order summary generation sub-model outputs the key information of the power work order to be processed, and the summary content of the power work order to be processed is output based on the key information and the power work order to be processed; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the power work order to be processed.

[0007] Furthermore, the power work order summary generation model further includes: a classifier;

[0008] The step of inputting the to-be-processed power work order into the power work order summary generation model so that the power work order summary generation model determines the business type of the to-be-processed power work order includes:

[0009] Inputting the pending power work order into a power work order summary generation model so that the power work order summary generation model extracts feature data of the pending power work order;

[0010] The characteristic data is input into the classifier, so that the classifier outputs the business type of the power work order to be processed according to the characteristic data.

[0011] Furthermore, the construction of the power work order summary generation model includes:

[0012] Obtaining a power work order summary sample training set; wherein the power work order summary sample training set includes a plurality of power work order summary sample training subsets, each power work order summary sample training subset includes a plurality of power work order summary samples, and each power work order summary sample training subset corresponds to a business type;

[0013] Constructing an initial power work order summary generation model; wherein the initial power work order summary generation model includes an initial classifier and a plurality of initial power work order summary generation sub-models, each initial power work order summary generation sub-model corresponding to a business type;

[0014] The initial power work order summary generation model is trained with the power work order summary sample training set until the initial power work order summary generation model converges, thereby obtaining the power work order summary generation model.

[0015] Further, the initial power work order summary generation model is trained with the power work order summary sample training set until the initial power work order summary generation model converges to obtain the power work order summary generation model, including:

[0016] Taking the power work order summary sample as the input of the initial classifier and the business type of the power work order summary sample as the output of the initial classifier, training the initial classifier until the initial classifier converges, freezing the first network parameter of the initial classifier, and generating a classifier;

[0017] For each initial power work order summary generation sub-model, the initial power work order summary generation sub-model is trained with the power work order summary sample training subset corresponding to the business type of the current initial power work order summary generation sub-model until each initial power work order summary generation sub-model converges, and each power work order summary generation sub-model to be adjusted is generated;

[0018] The initial power work order summary generation model and each power work order summary generation sub-model to be adjusted are jointly trained with the power work order summary sample training set until the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted jointly converge, thereby generating a power work order summary generation model.

[0019] Furthermore, in each iteration of the joint training, the second network parameters of each power work order summary generation sub-model to be adjusted and the third network parameters of the initial power work order summary generation model are adjusted according to the feedback mechanism.

[0020] Furthermore, before training the initial power work order summary generation model with the power work order summary sample training set, the method further includes:

[0021] Acquire a power industry semantic data sample set; wherein the power industry semantic data sample set includes a number of power industry semantic data samples representing professional terms in the power industry;

[0022] The training of the initial power work order summary generation model with the power work order summary sample training set includes:

[0023] The initial power work order summary generation model is trained with the power industry semantic data sample set and the power work order summary sample training set.

[0024] Furthermore, the key information includes: prompt words and text prefixes;

[0025] Outputting summary content of the pending power work order according to the key information and the pending power work order includes:

[0026] The power work order summary generation model outputs the summary content of the power work order to be processed based on the prompt words of the power work order to be processed output by the target power work order summary generation submodel, the text prefix of the power work order to be processed output by the target power work order summary generation submodel, and the summary content of the power work order to be processed.

[0027] Based on the above method embodiment, the present invention provides a corresponding device embodiment;

[0028] An embodiment of the present invention provides a summary generation device for an electric power work order, including: a work order acquisition module and a summary generation module;

[0029] The work order acquisition module is used to acquire the power work order to be processed;

[0030] The summary generation module is used to input the pending power work order into the power work order summary generation model so that the power work order summary generation model determines the business type of the pending power work order, inputs the pending power work order into the target power work order summary generation sub-model, so that the target power work order summary generation sub-model outputs the key information of the pending power work order, and outputs the summary content of the pending power work order based on the key information and the pending power work order; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the pending power work order.

[0031] Another embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the method for generating a summary of an electric power work order described in the above-mentioned embodiment of the invention is implemented.

[0032] Another embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute the summary generation method of the power work order described in the above-mentioned embodiment of the invention.

[0033] The following beneficial effects are achieved by implementing the present invention:

[0034] The present invention provides a summary generation method, device, terminal device and storage medium for an electric power work order. The summary generation method for an electric power work order inputs the acquired electric power work order to be processed into an electric power work order summary generation model, and then after the business type of the electric power work order to be processed is determined by the electric power work order summary generation model according to the electric power work order to be processed, the electric power work order to be processed is input into a target electric power work order summary generation sub-model with the same business type based on the business type of the electric power work order. The target electric power work order summary generation sub-model outputs key information according to the electric power work order to be processed, and the electric power work order summary generation model then outputs corresponding summary content according to the key information of the electric power work order and the electric power work order to be processed. After the business type of the electric power work order to be processed is determined by the electric power work order summary generation model, the key information is extracted by the electric power work order summary sub-model corresponding to the business type, which can avoid the situation in which the general large model in the prior art directly processes the electric power work order, resulting in the key information in the electric power work order being ignored, and improves the accuracy of the summary generation of the electric power work order. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 The present invention is a flowchart of a method for generating a summary of an electric power work order provided in one embodiment of the present invention.

[0036] Figure 2 It is a structural schematic diagram of a summary generation device for an electric power work order provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0039] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0040] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0041] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0042] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0043] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0044] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0045] like Figure 1 As shown, a method for generating a summary of an electric power work order provided by an embodiment of the present invention includes:

[0046] Step S1: Obtain the power work order to be processed;

[0047] Step S2: Input the pending power work order into the power work order summary generation model so that the power work order summary generation model determines the business type of the pending power work order, input the pending power work order into the target power work order summary generation sub-model so that the target power work order summary generation sub-model outputs the key information of the pending power work order, and outputs the summary content of the pending power work order based on the key information and the pending power work order; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the pending power work order.

[0048] For step S1, obtain the power work order to be processed.

[0049] For step S2, the power work order to be processed is input into the power work order summary generation model, so that the power work order summary generation model first determines the business type to which the input power work order to be processed belongs according to the business type to which the power work order to be processed belongs, and secondly, obtains the corresponding target power work order summary generation sub-model according to the business type to which the power work order to be processed belongs, and inputs the power work order to be processed into the target power work order summary generation sub-model, so that the target power work order summary generation sub-model extracts and outputs the key information of the power work order to be processed, and transmits the key information of the power work order to be processed to the power work order summary generation model, and the power work order summary generation model outputs the summary content of the power work order to be processed according to the key information and the power work order to be processed.

[0050] It should be noted that the power work order summary generation model includes multiple power work order summary generation sub-models, and each power work order summary generation sub-model corresponds to a business type. Preferably, when the power work order to be processed contains content of multiple business types, the content of the power work order to be processed is segmented according to the business type to obtain a number of segmented texts, each segmented text corresponds to a business type, and then the target power work order summary generation sub-model corresponding to each segmented text is selected according to the business type of the different segmented texts to process each segmented text separately, and after obtaining the key information of each segmented text, the summary content of the power work order to be processed is output through the power work order summary generation model and each key information.

[0051] In a preferred embodiment, the power work order summary generation model also includes: a classifier; inputting the power work order to be processed into the power work order summary generation model so that the power work order summary generation model determines the business type of the power work order to be processed, including: inputting the power work order to be processed into the power work order summary generation model so that the power work order summary generation model extracts feature data of the power work order to be processed; inputting the feature data into the classifier so that the classifier outputs the business type of the power work order to be processed according to the feature data.

[0052] Specifically, the power work order summary generation model also includes a classifier. When the power work order to be processed is input into the power work order summary generation model, the first thing it passes through is the classifier, which extracts the feature data in the power work order to be processed to determine the business type of the power work order to be processed. Preferably, in the present invention, the business types mainly include the following ten categories: power consumption and electricity fee inquiry, basic information of customer files, fault power outage, power outage in a belt, power outage for one household, electricity price, pre-arranged power outage, new installation and capacity increase, name change and transfer, and emergency repair efficiency.

[0053] In a preferred embodiment, the key information includes: prompt words and text prefixes; the output of the summary content of the power work order to be processed based on the key information and the power work order to be processed includes: the power work order summary generation model outputs the summary content of the power work order to be processed based on the prompt words of the power work order to be processed output by the target power work order summary generation submodel, the text prefix of the power work order to be processed output by the target power work order summary generation submodel, and the summary content of the power work order to be processed.

[0054] Specifically, the target power work order summary generation sub-model extracts key information from the power work order to be processed based on the input power work order to be processed, and extracts the prompt words and text prefixes corresponding to the power work order to be processed. Among them, the prompt words are mainly composed of instructions and input text. The instructions are the specific tasks that need to be performed by the power work order summary generation sub-model, which is the key information extraction task in the summary in the present invention; the input text corresponds to the input power work order to be processed in the present invention, which contains specific information that needs to be summarized, such as time, unit, person's name, and event name. Preferably, the prompt words can also include an output indicator. If the power work order to be processed needs to output a specified format, the output indicator guides the subsequent power work order summary generation model to output the summary content according to the specified format.

[0055] In a preferred embodiment, the construction of the electricity work order summary generation model includes: obtaining an electricity work order summary sample training set; wherein the electricity work order summary sample training set includes several electricity work order summary sample training subsets, each electricity work order summary sample training subset contains several electricity work order summary samples, and each electricity work order summary sample training subset corresponds to a business type; constructing an initial electricity work order summary generation model; wherein the initial electricity work order summary generation model includes an initial classifier and several initial electricity work order summary generation sub-models, and each initial electricity work order summary generation sub-model corresponds to a business type; training the initial electricity work order summary generation model with the electricity work order summary sample training set until the initial electricity work order summary generation model converges, thereby obtaining the electricity work order summary generation model.

[0056] Specifically, a power work order summary sample training set is obtained, and the power work order summary sample training set includes several power summary sample training subsets, and each power summary sample training subset corresponds to a business type. Based on the above business types specifically including ten categories, it can be known that the number of power summary sample training subsets to be obtained here is ten. Each power summary sample training subset contains multiple power summary samples, and each power summary sample is annotated with its business type, key information and real summary content.

[0057] An initial electricity work order summary generation model is constructed, which includes an initial classifier and several initial electricity work order summary generation sub-models. Based on the ten business types obtained above, ten initial electricity work order summary generation sub-models are constructed here, and each initial electricity work order summary generation sub-model corresponds to a business type.

[0058] In a preferred embodiment, before training the initial power work order summary generation model with the power work order summary sample training set, it also includes: obtaining a power industry semantic data sample set; wherein the power industry semantic data sample set includes a number of power industry semantic data samples representing professional terms in the power industry; the training of the initial power work order summary generation model with the power work order summary sample training set includes: training the initial power work order summary generation model with the power industry semantic data sample set and the power work order summary sample training set.

[0059] Specifically, a power industry semantic data sample set is obtained, which contains a number of power industry semantic data samples representing professional terms in the power industry, and each power industry semantic data sample is marked with a professional term. Before training the initial classifier and each initial power order summary generation sub-model in the initial power work order summary generation model, the initial power work order summary generation model is first pre-trained according to the power industry semantic data sample set, so that the initial power work order summary generation model can acquire the language understanding and recognition ability of professional terminology in the power industry through pre-training, ensuring that it has sufficient expression ability when summarizing complex power work orders.

[0060] In a preferred embodiment, the initial power work order summary generation model is trained with the power work order summary sample training set until the initial power work order summary generation model converges to obtain the power work order summary generation model, including:

[0061] Taking the power work order summary sample as the input of the initial classifier and the business type of the power work order summary sample as the output of the initial classifier, training the initial classifier until the initial classifier converges, freezing the first network parameter of the initial classifier, and generating a classifier;

[0062] For each initial power work order summary generation sub-model, the initial power work order summary generation sub-model is trained with the power work order summary sample training subset corresponding to the business type of the current initial power work order summary generation sub-model until each initial power work order summary generation sub-model converges, and each power work order summary generation sub-model to be adjusted is generated;

[0063] The initial power work order summary generation model and each power work order summary generation sub-model to be adjusted are jointly trained with the power work order summary sample training set until the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted jointly converge, thereby generating a power work order summary generation model.

[0064] Specifically, the initial classifier is trained according to each power work order summary sample in the power work order summary sample training set, so that the initial classifier can identify the business type corresponding to each sample. In each iteration of the initial classifier, an power work order summary sample is used as the input of the initial classifier, and the business type corresponding to the power work order summary sample is used as the output of the initial classifier to train the initial classifier. Through continuous iterative training, until the initial classifier loss function is minimized, the initial classifier converges at this time, and the converged initial classifier is used as the classifier in the final power work order summary generation model, and the classifier parameters are frozen.

[0065] Furthermore, each initial electricity work order summary generation sub-model is trained. Since the training of each initial electricity work order summary generation sub-model does not interfere with each other, each initial electricity work order summary generation sub-model is synchronously trained in a multi-threaded manner. For the training of any initial electricity work order summary generation sub-model, a training subset of electricity work order summary samples corresponding to the business type of the current initial electricity work order summary generation sub-model is obtained, and the current initial electricity work order summary generation sub-model is trained with several electricity work order summary samples in the training subset of electricity work order summary samples. During the training process, the electricity work order summary sample is used as the input of the current initial electricity work order summary generation sub-model, and the key information corresponding to the electricity work order summary sample is used as the output, and the current initial electricity work order summary generation sub-model is trained until the current initial electricity work order summary generation sub-model converges under the current training task, and the electricity work order summary generation sub-model to be adjusted is generated. When all the initial electricity work order summary generation sub-models converge, each electricity work order summary generation sub-model to be adjusted is obtained.

[0066] In a preferred embodiment, in each iteration of the joint training, the second network parameters of each power work order summary generation sub-model to be adjusted and the third network parameters of the initial power work order summary generation model are adjusted according to the feedback mechanism.

[0067] Specifically, the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted are jointly trained again with the power work order summary sample training set, and in each joint training process, the training error is determined according to the predicted summary content of the initial power work order summary generation model and the actual summary content of the input sample, and the feedback mechanism is triggered according to the training error to fine-tune the network parameters of each power work order summary generation sub-model to be adjusted and the network parameters of the initial power work order summary generation model, until the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted converge jointly, at which time the power work order summary generation model is generated.

[0068] Preferably, in the above-mentioned joint training process, an active evaluation index is introduced to dynamically evaluate the quality of the predicted summary content output by the initial power work order summary generation model, and the second network parameter and the third network parameter are adjusted according to the evaluation result. When the evaluation result converges, the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted converge jointly.

[0069] It should be added that the role of active evaluation indicators is to provide feedback on model training by quantitatively analyzing the differences between the predicted summary content and the real summary content. The following are some commonly used automatic evaluation indicators: (1) ROUGE (Recall-Oriented Understudy for Gisting Evaluation): By comparing the overlapping parts of the predicted summary content and the real summary content, ROUGE can evaluate the coverage of the predicted summary content on the real summary content, and is often used for quality evaluation of generative tasks. (2) BLEU (Bilingual Evaluation Understudy): It is used to evaluate the accuracy of the predicted summary content, calculate the similarity between the words in the predicted summary content generated by the model and the words in the predicted summary content, and measure whether the generated predicted summary content is reasonable and accurate. These indicators can provide real-time feedback on the performance of the model. When the evaluation indicators reach a certain convergence condition (such as the ROUGE or BLEU score no longer increases significantly), it means that the model has reached the optimal state and training can be stopped. This active evaluation mechanism ensures the quality of the generated summary and reduces the risk of overfitting. In addition, after generating the power work order summary generation model, the power work order summary generation model is optimized and adjusted through manual verification.

[0070] Based on the above method embodiment, the present invention provides a corresponding device embodiment.

[0071] like Figure 2 As shown, an embodiment of the present invention provides a summary generation device for an electric power work order, comprising: a work order acquisition module and a summary generation module;

[0072] The work order acquisition module is used to acquire the power work order to be processed;

[0073] The summary generation module is used to input the pending power work order into the power work order summary generation model so that the power work order summary generation model determines the business type of the pending power work order, inputs the pending power work order into the target power work order summary generation sub-model, so that the target power work order summary generation sub-model outputs the key information of the pending power work order, and outputs the summary content of the pending power work order based on the key information and the pending power work order; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the pending power work order.

[0074] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0075] Those skilled in the art can clearly understand that for the sake of convenience and brevity, the specific working process of the device described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

[0076] Based on the above method item embodiments, the present invention provides corresponding terminal device item embodiments.

[0077] An embodiment of the present invention provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a summary generation method for an electric power work order as described in any one of the present invention is implemented.

[0078] The terminal device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0079] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0080] The memory can be used to store the computer program, and the processor realizes various functions of the terminal device by running or executing the computer program stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the mobile phone, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0081] Based on the above method item embodiments, the present invention provides a corresponding storage medium item embodiment.

[0082] An embodiment of the present invention provides a storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the storage medium is located is controlled to execute a summary generation method for an electric power work order as described in any one of the present inventions.

[0083] The storage medium is a computer-readable storage medium, and the computer program is stored in the computer-readable storage medium. When the computer program is executed by the processor, the steps of each of the above-mentioned method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium.

[0084] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for generating a summary of an electric power work order, characterized in that: include: Get pending power work orders; The power work order to be processed is input into the power work order summary generation model so that the power work order summary generation model determines the business type of the power work order to be processed, and the power work order to be processed is input into the target power work order summary generation sub-model so that the target power work order summary generation sub-model outputs the key information of the power work order to be processed, and the summary content of the power work order to be processed is output based on the key information and the power work order to be processed; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the power work order to be processed.

2. A method for generating a summary of an electric power work order according to claim 1, characterized in that: The power work order summary generation model further includes: a classifier; The step of inputting the to-be-processed power work order into a power work order summary generation model so that the power work order summary generation model determines the business type of the to-be-processed power work order includes: Inputting the pending power work order into a power work order summary generation model so that the power work order summary generation model extracts feature data of the pending power work order; The characteristic data is input into the classifier, so that the classifier outputs the business type of the power work order to be processed according to the characteristic data.

3. A method for generating a summary of an electric power work order according to claim 2, characterized in that: The construction of the power work order summary generation model includes: Obtaining a power work order summary sample training set; wherein the power work order summary sample training set includes a plurality of power work order summary sample training subsets, each power work order summary sample training subset includes a plurality of power work order summary samples, and each power work order summary sample training subset corresponds to a business type; Constructing an initial power work order summary generation model; wherein the initial power work order summary generation model includes an initial classifier and a plurality of initial power work order summary generation sub-models, each initial power work order summary generation sub-model corresponding to a business type; The initial power work order summary generation model is trained with the power work order summary sample training set until the initial power work order summary generation model converges, thereby obtaining the power work order summary generation model.

4. A method for generating a summary of an electric power work order according to claim 3, characterized in that: The initial power work order summary generation model is trained with the power work order summary sample training set until the initial power work order summary generation model converges to obtain the power work order summary generation model, including: Taking the power work order summary sample as the input of the initial classifier and the business type of the power work order summary sample as the output of the initial classifier, training the initial classifier until the initial classifier converges, freezing the first network parameter of the initial classifier, and generating a classifier; For each initial power work order summary generation sub-model, the initial power work order summary generation sub-model is trained with the power work order summary sample training subset corresponding to the business type of the current initial power work order summary generation sub-model until each initial power work order summary generation sub-model converges, and each power work order summary generation sub-model to be adjusted is generated; The initial power work order summary generation model and each power work order summary generation sub-model to be adjusted are jointly trained with the power work order summary sample training set until the initial power work order summary generation model and each power work order summary generation sub-model to be adjusted jointly converge, thereby generating a power work order summary generation model.

5. A method for generating a summary of an electric power work order according to claim 4, characterized in that: In each iteration of the joint training, the second network parameters of each power work order summary generation sub-model to be adjusted and the third network parameters of the initial power work order summary generation model are adjusted according to the feedback mechanism.

6. A method for generating a summary of an electric power work order as claimed in claim 3, characterized in that: Before training the initial power work order summary generation model with the power work order summary sample training set, the method further includes: Acquire a power industry semantic data sample set; wherein the power industry semantic data sample set includes a number of power industry semantic data samples representing professional terms in the power industry; The training of the initial power work order summary generation model with the power work order summary sample training set includes: The initial power work order summary generation model is trained with the power industry semantic data sample set and the power work order summary sample training set.

7. A method for generating a summary of an electric power work order according to claim 1, characterized in that: The key information includes: prompt words and text prefixes; Outputting summary content of the pending power work order according to the key information and the pending power work order includes: The power work order summary generation model outputs the summary content of the power work order to be processed based on the prompt words of the power work order to be processed output by the target power work order summary generation submodel, the text prefix of the power work order to be processed output by the target power work order summary generation submodel, and the summary content of the power work order to be processed.

8. A summary generation device for an electric power work order, characterized in that: include: Work order acquisition module and summary generation module; The work order acquisition module is used to acquire the power work order to be processed; The summary generation module is used to input the pending power work order into the power work order summary generation model so that the power work order summary generation model determines the business type of the pending power work order, inputs the pending power work order into the target power work order summary generation sub-model, so that the target power work order summary generation sub-model outputs the key information of the pending power work order, and outputs the summary content of the pending power work order based on the key information and the pending power work order; wherein the power work order summary generation model includes several power work order summary generation sub-models, each power work order summary generation sub-model corresponds to a business type; the business type of the target power work order summary generation sub-model is the same as the business type of the pending power work order.

9. A terminal device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a method for generating a summary of an electric power work order as described in any one of claims 1 to 7 is implemented.

10. A storage medium, characterized in that: The storage medium includes a stored computer program, wherein when the computer program is executed, the device where the storage medium is located is controlled to execute a summary generation method for an electric power work order as described in any one of claims 1 to 7.

Images