Document task parsing method, device, electronic device and storage medium
By performing part-of-speech annotation, dependency analysis and entity recognition of the requirement document, extracting key information and inputting the task analysis model, the problem of difficulty in parsing specific tasks from the requirement document in the existing technology is solved, and the effective analysis of the task list is realized and implement.
Patent Information
- Application Number
- CN202410447758.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-15
- Publication Date
- 2025-02-14
- Estimated Expiration
- 2044-04-15
AI Technical Summary
It is difficult for the prior art to parse specific tasks based on the requirements content in the requirements document, especially when only the requirements content is listed in the requirements document but not the specific tasks are clearly listed.
By performing part-of-speech annotation, dependency analysis and entity recognition of the task document to be parsed, key information is extracted, and structured data is converted into JSON format, and then converted into semantic vectors. Enter the pre-trained task analysis model to obtain the task list.
It can effectively extract the key information of the task document to be parsed, and parse a list of tasks that meet the needs based on this information to ensure the reasonable and orderly execution of tasks and meet the needs of parsing task documents.
Smart Images
Figure CN118246436B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of document parsing, and in particular relates to a document task parsing method, device, electronic device and storage medium. Background Art
[0002] For task parsing in requirement documents, the existing technologies mostly identify tasks in documents through semantic recognition. However, the semantic recognition method can only perform task parsing on documents that clearly list specific task contents.
[0003] However, in actual situations, the requirement document often only lists the requirement content, but does not clearly list the specific tasks. It is often impossible to parse the task corresponding to the requirement document based on the content of the requirement document through semantic recognition.
[0004] Therefore, how to provide an effective solution to parse out tasks that meet document requirements based on document content has become a difficult problem that needs to be solved urgently in the prior art. Summary of the invention
[0005] The purpose of the present invention is to provide a document task parsing method, device, electronic device and storage medium to solve the above-mentioned problems existing in the prior art.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] In a first aspect, the present invention provides a document task parsing method, comprising:
[0008] By performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed, key information of the task document to be parsed is extracted;
[0009] Convert the key information of the task document to be parsed into structured data in JSON format;
[0010] Converting the structured data into semantic vectors;
[0011] Inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask;
[0012] The task analysis model is trained by taking the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output.
[0013] Based on the above disclosed content, the present invention extracts the key information of the task document to be parsed by performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed; converts the key information of the task document to be parsed into structured data in JSON format; converts the structured data into semantic vectors; inputs the semantic vectors into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask; wherein the task analysis model is trained with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output. In this way, the key information of the task document to be parsed can be extracted, and then a task list that meets the requirements of the task document to be parsed can be parsed according to the key information of the task document to be parsed, so as to reasonably and orderly complete the tasks indicated in the task document to be parsed according to the task list, thereby meeting the requirements of the parsed task document.
[0014] In a possible design, inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed includes:
[0015] Normalizing the semantic vector;
[0016] The normalized semantic vector is input into a pre-trained task analysis model to obtain a task list corresponding to the task document to be analyzed.
[0017] In a possible design, the task list also includes computing model information and / or software tool information required for each task in the at least one task.
[0018] In a possible design, after obtaining the task list corresponding to the task document to be parsed, the method further includes:
[0019] According to the task type of each subtask in the task list, a task processing strategy for each subtask is determined.
[0020] In one possible design, the task analysis model is a BERT model.
[0021] In one possible design, the method further includes:
[0022] Get multiple sample task documents;
[0023] Extracting key information of each sample task document from the multiple sample task documents by performing part-of-speech tagging, dependency parsing, and entity recognition on the multiple sample task documents;
[0024] Convert the key information of each sample task document into structured data in JSON format;
[0025] Convert the structured data corresponding to each sample task document into a semantic vector to obtain the semantic vector corresponding to each sample task document;
[0026] The semantic vector corresponding to each sample task document is used as the input of the model, and the task list corresponding to each sample task document is used as the output of the model for training to obtain the task analysis model.
[0027] In a possible design, after converting the structured data corresponding to each sample task document into a semantic vector to obtain the semantic vector corresponding to each sample task document, the method further includes:
[0028] Calculate the similarity between the semantic vector corresponding to each sample task document and the semantic vector in the knowledge base;
[0029] Based on the similarity with the semantic vectors in the knowledge base, the semantic vector corresponding to each sample task document is expanded.
[0030] In a second aspect, the present invention provides a document task parsing device, comprising:
[0031] An extraction unit, used to extract key information of the task document to be parsed by performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed;
[0032] A first conversion unit, used to convert key information of the task document to be parsed into structured data in JSON format;
[0033] A second conversion unit, configured to convert the structured data into a semantic vector;
[0034] A computing unit, used for inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask;
[0035] The task analysis model is trained by taking the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output.
[0036] In a third aspect, the present invention provides an electronic device comprising a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the document task parsing method as described in the first aspect or any possible design of the first aspect.
[0037] In a fourth aspect, the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed on a computer, executes the document task parsing method described in the first aspect or any possible design of the first aspect.
[0038] In a fifth aspect, the present invention provides a computer program product comprising instructions, which, when executed on a computer, causes the computer to execute the document task parsing method as described in the first aspect or any possible design of the first aspect.
[0039] Beneficial effects:
[0040] The document task parsing method, device, electronic device and storage medium provided by the present invention can extract the key information of the task document to be parsed, and then parse out a task list that meets the requirements of the task document to be parsed based on the key information of the task document to be parsed, so as to reasonably and orderly complete the tasks indicated in the task document to be parsed according to the task list, thereby meeting the requirements of the task document to be parsed and facilitating practical application and promotion. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 A flowchart of a document task parsing method provided in an embodiment of the present application;
[0042] Figure 2 A schematic block diagram of a document task parsing device provided in an embodiment of the present application;
[0043] Figure 3 A schematic block diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in combination with the drawings and the description of the embodiments or the prior art. Obviously, the following description of the structure of the drawings is only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. It should be noted that the description of these embodiments is used to help understand the present invention, but does not constitute a limitation of the present invention.
[0045] It should be understood that although the terms first, second, etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another unit. For example, a first unit can be referred to as a second unit, and similarly, a second unit can be referred to as a first unit without departing from the scope of the exemplary embodiments of the present invention.
[0046] It should be understood that for the term "and / or" that may appear in this text, it is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, B exists alone, and both A and B exist simultaneously. For the term " / and" that may appear in this text, it describes another association object relationship, indicating that there can be two relationships. For example, A / and B can represent: A exists alone, and both A and B exist. Additionally, for the character " / " that may appear in this text, it generally indicates that the associated objects before and after are in an "or" relationship.
[0047] To facilitate the parsing of tasks in a document, an embodiment of the present application provides a document task parsing method, apparatus, electronic device, and storage medium. The document task parsing method, apparatus, electronic device, and storage medium can parse out a task list that meets the requirements of the document to be parsed.
[0048] The document task parsing method provided by the embodiment of the present application can be applied to a user terminal or a server. It can be understood that the execution subject does not constitute a limitation to the embodiment of the present application.
[0049] The document task parsing method provided by the embodiment of the present application will be described in detail below.
[0050] As Figure 1 shown, it is a flowchart of the document task parsing method provided by the first aspect of the embodiment of the present application. The document task parsing method can but is not limited to include the following steps S101 - S105.
[0051] Step S101. Extract the key information of the document to be parsed by performing part-of-speech tagging, dependency parsing, and entity recognition on the document to be parsed.
[0052] Specifically, the document to be parsed can first be split into words and symbols by a word segmenter, and each word can be assigned a part of speech, such as a noun, verb, adjective, etc., through natural language processing (NLP).
[0053] At the same time, a sentence dependency tree is constructed to determine the relationships between the words in the document. For example, for the sentence "I ate a bowl of rice", "I" is the subject and depends on the verb "ate", "ate" is the predicate and depends on "了", "了" is a particle and depends on "ate", and "的" is a modifier and depends on "a bowl of rice".
[0054] In addition, specific entities in the document to be parsed, such as task names, tools, skills, etc., can also be detected and labeled.
[0055] Finally, the words tagged with part of speech, the constructed sentence dependency tree and the identified entities are used as the key information of the task document to be parsed.
[0056] Step S102: Convert key information of the task document to be parsed into structured data in JSON format.
[0057] In an embodiment of the present application, structured data is formed by converting the key information of the task document to be parsed into JSON format to facilitate subsequent machine processing.
[0058] Step S103: Convert structured data into semantic vectors.
[0059] In an embodiment of the present application, structured data can be converted into high-dimensional semantic vectors based on a word embedding model.
[0060] Step S104. Input the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask.
[0061] In an embodiment of the present application, a task analysis model for analyzing tasks of a document is pre-trained, and the task analysis model can be obtained by training with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output. The task analysis model can be, but is not limited to, a BERT (Bidirectional Encoder Representation from Transformers) model or a Transformer model.
[0062] After the structured data is converted into a semantic vector, the semantic vector can be input into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask.
[0063] In one or more embodiments, the task list may also include computing model information and / or software tool information required for each task in the at least one task. Specifically, when executing a task in the task list, some external computing models and / or software tools (such as GPT, Llama2, Wenxinyiyan, etc.) may be called according to task requirements. Therefore, when generating the task list, the task list may also include computing model and / or software tool information required for each task in the at least one task, so as to prompt the user to use the corresponding computing model and / or software tool to ensure that the task can be executed smoothly.
[0064] In one or more embodiments, when inputting a semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, the semantic vector can be normalized first, and then the normalized semantic vector can be input into the pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed.
[0065] Step S105: Determine a task processing strategy for each subtask according to the task type of each subtask in the task list.
[0066] In the embodiment of the present application, the corresponding task processing strategy can be matched according to the task type of each subtask. The task processing strategy can refer to the thinking strategy method used to complete the task, such as directional thinking method, comparative thinking method and SWOT analysis method.
[0067] For example, for data analysis tasks such as analyzing hot topics, you can use SWOT analysis to identify the strengths, weaknesses, opportunities, and threats of hot topics. For selection tasks such as choosing a brand or product, you can use comparative thinking to evaluate the strengths and weaknesses of each brand or product option. For tasks such as writing about topics, you can use reverse thinking to analyze from different angles to generate novel and attractive content.
[0068] The training process of the task analysis model provided by the present application will be described in detail below. The training process of the task analysis model may include but is not limited to the following steps S201-S205.
[0069] Step S201: Obtain multiple sample task documents.
[0070] In the embodiment of the present application, the multiple sample task documents include but are not limited to documents in specific fields (such as medicine, finance, engineering, etc.), books and papers that study classic thinking skills and strategies, and API documents related to specific industrial standards and tools.
[0071] For example, you can use web crawlers to crawl articles, papers, and reports from experts in the industry, such as Nature, Aminer, IEEE Xplore, ScienceDirect, arXiv, Springer, Elsevier, JSTOR, Wiley Online Library, Stanford Research Institute, Google AI Research, OpenAI, Meta, Baidu AI Open Platform, YiQianwen, or iFlytek Spark. You can obtain API documentation related to specific industry standards and tools through open source platforms such as GitHub, GitLab, SourceForge, Bitbucket, etc.
[0072] Step S202: extract key information of each sample task document from the multiple sample task documents by performing part-of-speech tagging, dependency parsing and entity recognition on the multiple sample task documents.
[0073] By performing part-of-speech tagging, dependency parsing and entity recognition on the multiple sample task documents, key information of each of the multiple sample task documents is extracted, which is consistent with the process of the aforementioned step S101 and will not be described in detail here.
[0074] In one or more embodiments, before extracting key information of each sample task document from the multiple sample task documents, the multiple sample task documents may be preprocessed. The preprocessing of the multiple sample task documents may include denoising, data enhancement, and data balancing.
[0075] Among them, denoising can be to remove irrelevant or erroneous information from the data. For example, web pages and articles obtained from the Internet often contain many HTML tags, special characters, or other content that is useless for the purpose of this training. Therefore, after classification according to the actual situation, regular expression filtering, large language model semantic filtering or manual filtering are used.
[0076] Data enhancement can include text rotation, back translation (translating text into another language and then translating it back), word replacement (replacing the original word with a synonym), etc., in order to cover as many expressions as possible on the same knowledge and achieve an enhanced effect.
[0077] Data balancing means that in order to avoid the model being biased towards a certain category, it is necessary to ensure that the data of the sample task documents of multiple categories are evenly distributed. If there are fewer sample task documents in a certain category, the balance of sample task documents of various categories can be achieved by increasing sampling or expanding data, or reducing sample task documents of other categories.
[0078] Step S203: Convert key information of each sample task document into structured data in JSON format.
[0079] Step S204: Convert the structured data corresponding to each sample task document into a semantic vector to obtain the semantic vector corresponding to each sample task document.
[0080] In one or more embodiments, after converting the structured data corresponding to each sample task document into a semantic vector and obtaining the semantic vector corresponding to each sample task document, the similarity between the semantic vector corresponding to each sample task document and the semantic vector in the knowledge base can be calculated, and based on the similarity with the semantic vector in the knowledge base, the semantic vector corresponding to each sample task document can be expanded. In this way, the trained samples can be expanded to improve the accuracy of the task analysis model trained subsequently.
[0081] Step S205: Take the semantic vector corresponding to each sample task document as the input of the model, and the task list corresponding to each sample task document as the output of the model for training to obtain a task analysis model.
[0082] Specifically, a task list corresponding to each sample task document can be determined in advance according to the content corresponding to each sample task document, and the task list includes at least one task set in sequence, and each task includes at least one subtask. When training, the semantic vector corresponding to each sample task document can be used as the input of the model, and the task list corresponding to each sample task document can be used as the output of the model for training to obtain the task analysis model.
[0083] In an embodiment of the present application, cross-validation can be used during training, that is, a part of the sample task documents are selected as training samples for training, and another part of the sample task documents are selected as test samples for testing, so as to ensure that the model has good generalization ability, and use Bayesian optimization to adjust the model hyperparameters, and stop iteration when the overall accuracy reaches above 0.90.
[0084] The document task parsing method provided by the present invention extracts the key information of the task document to be parsed by performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed; converts the key information of the task document to be parsed into structured data in JSON format; converts the structured data into a semantic vector; inputs the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask; wherein the task analysis model is obtained by training with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output. In this way, the key information of the task document to be parsed can be extracted, and then a task list that meets the requirements of the task document to be parsed can be parsed according to the key information of the task document to be parsed, so as to reasonably and orderly complete the tasks indicated in the task document to be parsed according to the task list, thereby meeting the requirements of the task document to be parsed, facilitating practical application and promotion, and facilitating practical application and promotion.
[0085] See also Figure 2 The embodiment of the present application also provides a document task parsing device, the document task parsing device comprising:
[0086] An extraction unit, used to extract key information of the task document to be parsed by performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed;
[0087] A first conversion unit, used to convert key information of the task document to be parsed into structured data in JSON format;
[0088] A second conversion unit, configured to convert the structured data into a semantic vector;
[0089] A computing unit, used for inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask;
[0090] The task analysis model is obtained by training with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output.
[0091] The working process, working details and technical effects of the document task parsing device provided in the second aspect of this embodiment can be found in the first aspect of the embodiment and will not be described in detail here.
[0092] like Figure 3 As shown, the third aspect of an embodiment of the present application provides an electronic device, comprising a memory, a processor and a transceiver that are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program and execute the document task parsing method as described in the first aspect of the embodiment.
[0093] For specific examples, the memory may include but is not limited to random access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out memory (FIFO) and / or first-in-last-out memory (FILO), etc.; the processor may be but is not limited to a microprocessor of the STM32F105 series, an ARM (Advanced RISC-Machines), an X86 or other architecture processor, or a processor with an integrated NPU (neural-network processing units); the transceiver may be but is not limited to a WiFi (Wireless Fidelity) wireless transceiver, a Bluetooth wireless transceiver, a General Packet Radio Service (GPRS) wireless transceiver, a ZigBee protocol (a low-power local area network protocol based on the IEEE802.15.4 standard, ZigBee) wireless transceiver, a 3G transceiver, a 4G transceiver and / or a 5G transceiver, etc.
[0094] In a fourth aspect of this embodiment, there is provided a computer-readable storage medium storing instructions including the document task parsing method described in the first aspect of the embodiment, that is, the computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the document task parsing method described in the first aspect is executed. The computer-readable storage medium refers to a carrier for storing data, which may include but is not limited to a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive, and / or a memory stick, etc., and the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
[0095] The fifth aspect of this embodiment provides a computer program product comprising instructions, which, when executed on a computer, causes the computer to execute the document task parsing method as described in the first aspect of the embodiment, wherein the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
[0096] It should be understood that certain details are provided in the following description to facilitate a complete understanding of the example embodiments. However, it should be understood by those of ordinary skill in the art that the example embodiments can be implemented without these certain details. For example, the system can be shown in a block diagram to avoid obscuring the example with unnecessary details. In other examples, well-known processes, structures, and techniques may not be shown in unnecessary detail to avoid obscuring the example embodiments.
[0097] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.
Claims
1. A document task parsing method, characterized in that: include: By performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed, key information of the task document to be parsed is extracted, wherein the key information includes words after part-of-speech tagging, constructed sentence dependency trees and recognized entities; Convert the key information of the task document to be parsed into structured data in JSON format; Converting the structured data into semantic vectors; Inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask; The task analysis model is obtained by training with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output; The task list also includes computing model information and / or software tool information required for each task in the at least one task, so as to prompt the user of the computing model and / or software tool to be used; After obtaining the task list corresponding to the task document to be parsed, the method further includes: According to the task type of each subtask in the task list, determine the task processing strategy of each subtask; The task processing strategy is a thinking strategy method used to complete the task, and the thinking strategy method is a directional thinking method, a comparative thinking method or a SWOT analysis method.
2. The document task parsing method according to claim 1, characterized in that: The step of inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed includes: Normalizing the semantic vector; The normalized semantic vector is input into a pre-trained task analysis model to obtain a task list corresponding to the task document to be analyzed.
3. The document task parsing method according to claim 1, characterized in that: The task analysis model is a BERT model.
4. The document task parsing method according to claim 1, characterized in that: The method further comprises: Get multiple sample task documents; Extracting key information of each sample task document from the multiple sample task documents by performing part-of-speech tagging, dependency parsing, and entity recognition on the multiple sample task documents; Convert the key information of each sample task document into structured data in JSON format; Convert the structured data corresponding to each sample task document into a semantic vector to obtain the semantic vector corresponding to each sample task document; The semantic vector corresponding to each sample task document is used as the input of the model, and the task list corresponding to each sample task document is used as the output of the model for training to obtain the task analysis model.
5. The document task parsing method according to claim 4, characterized in that: After converting the structured data corresponding to each sample task document into a semantic vector to obtain the semantic vector corresponding to each sample task document, the method further includes: Calculate the similarity between the semantic vector corresponding to each sample task document and the semantic vector in the knowledge base; Based on the similarity with the semantic vectors in the knowledge base, the semantic vector corresponding to each sample task document is expanded.
6. A document task parsing device, characterized in that: include: An extraction unit is used to extract key information of the task document to be parsed by performing part-of-speech tagging, dependency parsing and entity recognition on the task document to be parsed, wherein the key information includes words after part-of-speech tagging, constructed sentence dependency trees and recognized entities; A first conversion unit, used to convert key information of the task document to be parsed into structured data in JSON format; A second conversion unit, configured to convert the structured data into a semantic vector; A computing unit, used for inputting the semantic vector into a pre-trained task analysis model to obtain a task list corresponding to the task document to be parsed, wherein the task list includes at least one task, and each task includes at least one subtask; The task analysis model is obtained by training with the semantic vector corresponding to the key information of the sample task document as the input of the model and the task list corresponding to the sample task document as the output; The task list also includes computing model information and / or software tool information required for each task in the at least one task, so as to prompt the user of the computing model and / or software tool to be used; The document task parsing device also includes: A determination unit, configured to determine a task processing strategy for each subtask according to a task type of each subtask in the task list after obtaining the task list corresponding to the task document to be parsed; The task processing strategy is a thinking strategy method used to complete the task, and the thinking strategy method is a directional thinking method, a comparative thinking method or a SWOT analysis method.
7. An electronic device, characterized in that: It comprises a memory, a processor and a transceiver which are communicatively connected in sequence, wherein the memory is used to store a computer program, the transceiver is used to send and receive messages, and the processor is used to read the computer program to execute the document task parsing method as described in any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the document task parsing method according to any one of claims 1 to 5 is executed.
Citation Information
Patent Citations
Entity relationship extraction method, device and equipment, and storage medium
CN111274394A