CN117744620A

CN117744620A - Automatic generation method, device, equipment and storage medium of articles

Info

Publication number: CN117744620A
Application number: CN202311768121.6A
Authority: CN
Inventors: 龚健; 宋书青; 王雷; 王召
Original assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Current assignee: Beijing Baidu Netcom Science and Technology Co Ltd
Priority date: 2023-12-20
Filing date: 2023-12-20
Publication date: 2024-03-22

Abstract

The disclosure provides an automatic article generation method, device, equipment and storage medium, relates to the technical field of artificial intelligence, and particularly relates to the technical field of intelligent writing. The specific implementation scheme is as follows: acquiring a sketch outline and a sketch material of an article to be generated; the method comprises the steps of writing outline to represent outline of an article to be generated, and writing material to represent material of the article to be generated; determining a first prompt word according to the sketch outline, wherein the first prompt word characterizes an evaluation result of the sketch outline; according to the first prompting word, adjusting the sketch outline to obtain an adjusted sketch outline; and generating an article to be generated according to the adjusted sketch outline and the sketch material. Therefore, the efficiency and the automation degree of the article generation are improved.

Description

Automatic generation method, device, equipment and storage medium of articles

Technical Field

The disclosure relates to the technical field of intelligent writing in the technical field of artificial intelligence, in particular to an automatic article generating method, device, equipment and storage medium.

Background

With the rapid development of information technology, computer-aided writing has become an important tool for helping authors to improve writing efficiency and quality.

What is needed is a way to automatically and quickly complete the automatic collaboration of articles.

Disclosure of Invention

The disclosure provides an automatic article generating method, device, equipment and storage medium, which are used for realizing rapid and automatic article generating.

According to a first aspect of the present disclosure, there is provided an automatic article generating method, including:

acquiring a sketch outline and a sketch material of an article to be generated; the sketching outline characterizes outline of the article to be generated, and the sketching material characterizes material of the article to be generated; determining a first prompt word according to the sketch outline, wherein the first prompt word represents an evaluation result of the sketch outline;

according to the first prompting word, the sketch outline is adjusted, and the adjusted sketch outline is obtained;

and generating the article to be generated according to the adjusted sketch outline and the sketch material.

According to a second aspect of the present disclosure, there is provided an automatic generation apparatus of an article, including:

the acquisition unit is used for acquiring the sketch outline and the sketch material of the article to be generated; the sketching outline characterizes outline of the article to be generated, and the sketching material characterizes material of the article to be generated; determining a first prompt word according to the sketch outline, wherein the first prompt word represents an evaluation result of the sketch outline;

The adjusting unit is used for adjusting the sketch outline according to the first prompt word to obtain an adjusted sketch outline;

and the generating unit is used for generating the article to be generated according to the adjusted sketch outline and the sketch material.

According to a third aspect of the present disclosure, there is provided a computer program product comprising: a computer program stored in a readable storage medium, from which it can be read by at least one processor of an electronic device, the at least one processor executing the computer program causing the electronic device to perform the method of the first aspect.

According to a fourth aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the automatic generation method of the article of the first aspect.

According to a fifth aspect of the present disclosure, there is provided a computer program product comprising: a computer program stored in a readable storage medium, from which it can be read by at least one processor of an electronic device, the at least one processor executing the computer program causing the electronic device to perform the method of automatically generating articles according to the first aspect.

According to the technology disclosed by the invention, the quality of the sketch outline can be rapidly and effectively improved by representing the first prompt word of the evaluation result of the sketch outline and adjusting the sketch outline, and further, the generation efficiency and the automation degree of the article can be improved by generating the article to be generated through the adjusted sketch outline and the sketch material.

It should be understood that the description in this section is not intended to identify key or critical features of the embodiments of the disclosure, nor is it intended to be used to limit the scope of the disclosure. Other features of the present disclosure will become apparent from the following specification.

Drawings

The drawings are for a better understanding of the present solution and are not to be construed as limiting the present disclosure. Wherein:

FIG. 1 is a schematic diagram according to a first embodiment of the present disclosure;

FIG. 2 is a schematic diagram according to a second embodiment of the present disclosure;

FIG. 3 is a schematic diagram according to a third embodiment of the present disclosure;

FIG. 4 is a schematic diagram according to a fourth embodiment of the present disclosure;

FIG. 5 is a schematic diagram of a fifth embodiment of the present disclosure;

fig. 6 is a schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure.

Detailed Description

Exemplary embodiments of the present disclosure are described below in conjunction with the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding, and should be considered as merely exemplary. Accordingly, one of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.

With the development of artificial intelligence technology, computer aided writing has become a research hotspot. For example: the template-driven writing system can provide various templates for users to help the users write documents according to specific formats; a rule-based grammar checker and editor can pay attention to the grammar and style of texts, such as a built-in review function of Grammarly and Microsoft Word, and can help users to identify and correct text errors; the artificial intelligence authoring assistant is a large language model based on AI technology, such as a GPT series model of OpenAI, and can generate coherent texts according to the requirements of users and return the coherent texts to the users.

However, the template-driven writing system can only provide templates for users, and cannot automatically generate coherent texts, and the rule-based grammar checker and editor can only check text contents after text generation, so that an artificial intelligence writing assistant can generate coherent texts, but has strong randomness, poor text structure and insufficient contents.

In the prior art, firstly, a sketch outline is selected through a preset outline template library, and then the sketch outline, a sketch material and a large language model are combined to obtain an article to be generated. However, the outline templates in the outline template library are fixed, so that the requirements of all users on writing outline cannot be met, and the accuracy of the generated articles is low. After the article is generated according to the sketch outline, the article needs to be further manually modified, so that the automation degree of the article generating method is reduced, and the article generating efficiency is also reduced.

In order to solve the above problems, the present disclosure provides an automatic article generating method, apparatus, device and storage medium, which are applied to the technical field of intelligent writing in the technical field of artificial intelligence, and in the present disclosure, by characterizing a first prompting word for an evaluation result of a writing outline, the quality of the writing outline can be quickly and effectively improved, and further, the article to be generated is generated by the adjusted writing outline and writing material, so that the generating efficiency and automation degree of the article can be improved.

The following describes the technical scheme of the present disclosure and how the technical scheme of the present disclosure solves the above technical problems in detail with specific embodiments. The following embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments. Embodiments of the present disclosure will be described below with reference to the accompanying drawings.

Fig. 1 is a schematic diagram according to a first embodiment of the present disclosure. As shown in fig. 1, the method for automatically generating an article according to the first embodiment of the present disclosure includes:

step S101, acquiring a sketch outline and a sketch material of an article to be generated; the method comprises the steps of writing outline to represent outline of an article to be generated, and writing material to represent material of the article to be generated; and determining a first prompt word according to the sketch outline, wherein the first prompt word characterizes an evaluation result of the sketch outline.

The execution device of the embodiment of the disclosure may be a server including an automatic article generating device, and may be simply referred to as a server.

Specifically, the server may acquire a sketch outline and a sketch material of an article to be generated; wherein, the outline is a document which generally describes the outline and the key points, and does not write all the contents of the whole text, only writes the main contents, and the outline includes but is not limited to: title, outline, etc., wherein the sketch outline characterizes outline of the article to be generated, namely, the sketch outline is outline of the article to be generated and document of the main points. The material refers to original material for literature and artistic creation, wherein the written material characterizes the material of the article to be generated, namely, the written material is the original material for creating the article to be generated.

The method for acquiring the sketch outline of the article to be generated is not limited, and optionally, the sketch outline can be determined according to the reference article, the sketch main body and the sketch material. Alternatively, the authoring outline may be selected from the outline template according to the authoring intent of the user. Optionally, the sketch outline input by the user can be obtained through the article generation interface.

The method for obtaining the writing material of the article to be generated is not limited, and optionally, the initial writing material can be obtained from a material source of the article to be generated according to the writing subject, and the obtained initial writing material is subjected to material processing to obtain the writing material of the article to be generated. Optionally, the authoring material input by the user may be obtained through the article generation interface. Alternatively, the authoring material therein may be extracted from the reference file.

Specifically, after the sketch outline and the sketch material of the article to be generated are obtained, a first prompt word can be determined according to the sketch outline, wherein the prompt word prompt mainly has the effect of prompting the context of input information and the parameter information of the input model for the Al model, when the model with supervised learning or without supervised learning is trained, the prompt word prompt can help the model to better understand the input intention and respond correspondingly, the interpretive and accessibility of the model can be improved by the prompt word prompt, namely the prompt word prompt is used for providing a prompt or guide for the AI model to help the AI model to better understand and complete tasks, and the first prompt word characterizes the evaluation result of the sketch outline.

The method for determining the first prompting word according to the sketch outline is not limited, and optionally, the first prompting word corresponding to the sketch outline can be determined according to a preset first prompting word template and the sketch outline. Optionally, the preset evaluation condition for the sketch outline and the sketch outline may be combined, the prompt word for generating the first prompt word is determined, and the first prompt word is input into the preset large language model and output. Optionally, the sketch outline may be sent to an article generation interface to obtain a first prompt word input by the user based on the sketch outline.

And step S102, adjusting the sketch outline according to the first prompt word to obtain the adjusted sketch outline.

Specifically, the sketch outline of the article to be generated, which is obtained in step S101, may be adjusted according to the first prompt word determined in step S101, so as to obtain the adjusted sketch outline. The method for adjusting the sketch outline according to the first prompting word is not limited, and optionally, the first prompting word can be input into a preset large language model, the sketch outline evaluation value and the improvement information are output, and the sketch outline is adjusted according to the sketch outline evaluation value and the improvement information. Optionally, after determining the evaluation value and the improvement information of the sketch outline, the sketch outline may be combined with the improvement opinion, the prompting word for generating the adjusted sketch outline is determined and input into a preset large language model, and the adjusted sketch outline is output.

And step S103, generating an article to be generated according to the adjusted sketch outline and the adjusted sketch material.

Specifically, the article to be generated may be generated according to the adjusted sketch outline obtained in step S102 and the sketch material of the article to be generated obtained in step S101. The method for generating the article to be generated according to the adjusted sketch outline and the sketch material is not limited, and optionally, a third prompting word corresponding to the adjusted sketch outline can be determined according to a preset third prompting word template, the adjusted sketch outline and the sketch material, and the third prompting word is further input into a preset large language model to output the article to be generated. Optionally, the adjusted sketch outline and the sketch material can be combined to generate the prompting word of the third prompting word, the prompting word is input into a preset large language model, the third prompting word is output, the third prompting word is further input into the preset large language model, and the article to be generated is output. Optionally, the adjusted sketch and the adjusted sketch material can be sent to an article generating interface, so that a user can obtain a third prompt word based on the input of the adjusted sketch and the adjusted sketch material, and the third prompt word is further input into a preset large language model to output an article to be generated. Optionally, the adjusted sketch and the sketch material can be sent to an article generation interface to obtain the article to be generated, which is input by the user based on the adjusted sketch and the sketch material.

In the embodiment of the disclosure, by acquiring a sketch outline and a sketch material of an article to be generated, wherein the sketch outline represents the outline of the article to be generated, and the sketch material represents the material of the article to be generated; and determining a first prompt word according to the writing outline, wherein the first prompt word characterizes an evaluation result of the writing outline, the writing outline is adjusted according to the first prompt word, the adjusted writing outline is obtained, and an article to be generated is generated according to the adjusted writing outline and writing materials, wherein the quality of the writing outline can be improved rapidly and effectively by means of the first prompt word characterizing the evaluation result of the writing outline, and further, the article to be generated is generated by means of the adjusted writing outline and the writing materials, so that the generation efficiency and the automation degree of the article can be improved.

In some embodiments, in the process of determining the first prompting word according to the sketch outline, the first prompting word corresponding to the sketch outline may be determined according to a preset first prompting word template and the sketch outline; the preset first prompting word template comprises evaluation conditions of each sketch outline. Specifically, the first prompt word template may be set according to a preset evaluation condition of the sketch outline, where the evaluation condition of the sketch outline includes: whether the sketch outline meets one or more of a definition requirement, a logic requirement, an integrity requirement, a consistency requirement, a detail suitability requirement, an innovation requirement and a uniqueness requirement or not, wherein the definition requirement refers to the evaluation of whether the sketch outline clearly shows main views, arguments and the like of an article; logic requirements are that whether the sketch outline is arranged according to a logic sequence is evaluated, so that readers can easily follow ideas of articles and the like; integrity requirements, namely, evaluating whether the sketch outline covers all main aspects, important parts, key points and the like of the theme; consistency requirements, namely, evaluating whether each part of the sketch outline keeps consistent style and depth, whether each part of the sketch outline is closely related to a theme, and the like; detailed suitability requirements refer to evaluating whether information in the sketch outline is detailed and concise, whether key parts are properly emphasized, and the like; innovative and/or uniqueness requirements refer to assessing whether a sketch outline reveals a unique understanding or novel view of the subject matter, helps to stand out articles, etc. Specifically, after the first prompt word template is set according to the preset evaluation condition of the sketch outline, the first prompt word template with a preset value and the sketch outline can be combined to generate a first prompt word corresponding to the sketch outline, optionally, the preset first prompt word template includes a field to be filled of the sketch outline in addition to the evaluation condition of each sketch outline described above, and the acquired sketch outline can be filled into the field to be filled of the sketch outline to generate the first prompt word corresponding to the sketch outline. The preset first prompting word template comprises evaluation conditions of each writing outline, so that comprehensive evaluation of the writing outline is realized, and the first prompting word determined on the basis can realize more comprehensive evaluation and adjustment of the writing outline, so that the quality of the adjusted writing outline is further improved, and the efficiency and the degree of automation of article generation are improved.

In some embodiments, in the process of adjusting the sketch outline according to the first prompting word to obtain the adjusted sketch outline, the first prompting word can be input into a preset large language model, and the sketch outline evaluation value and the improvement information are output; wherein the improvement information characterizes a modification suggestion for the authoring outline; if the evaluation value of the writing outline is smaller than the preset evaluation threshold value, the writing outline is adjusted according to the improvement information, and the adjusted writing outline is obtained. The large language model LLM is a deep learning model trained by using a large amount of text data, can generate natural language text or understand meaning of the language text, can process various natural language tasks such as text classification, question-answering, dialogue and the like, is an important path leading to artificial intelligence, is a model capable of outputting output contents represented by the prompt words according to the corresponding prompt words, and in the embodiment of the present disclosure, the preset large language model can output evaluation results of written outline represented by the first prompt words according to the first prompt words, wherein the present disclosure does not limit how to determine the output contents of the model through the prompt words and the large language model.

Specifically, in the embodiment of the present disclosure, the evaluation result of the authoring outline includes: the evaluation value of the writing outline and the improvement information, wherein the evaluation value of the writing outline is calculated according to the evaluation condition of the writing outline described above and the preset evaluation condition weight, and optionally, the sub-evaluation value corresponding to the evaluation condition of each writing outline can be determined according to the evaluation condition of the writing outline, and then the evaluation value of the writing outline is calculated according to the preset evaluation condition weight, wherein the improvement information characterizes the modification suggestion of the writing outline, and the modification suggestion corresponds to the evaluation condition of the writing outline described above and the sub-evaluation value corresponding to the evaluation condition of each writing outline, for example: and if the sub-evaluation value corresponding to the integrity requirement is smaller than a preset sub-evaluation value threshold, generating a modification opinion corresponding to the integrity requirement and the sub-evaluation value. Specifically, if it is determined that the evaluation value of the writing outline is greater than or equal to the preset evaluation threshold, it is determined that the writing outline meets the requirement of the evaluation condition of the writing outline, and the writing outline can be used as the adjusted writing outline. Specifically, if it is determined that the evaluation value of the writing outline is smaller than the preset evaluation threshold, the writing outline is adjusted according to the improvement information to obtain an adjusted writing outline, wherein in the embodiment of the disclosure, the manner of adjusting the writing outline according to the improvement information to obtain the adjusted writing outline is not limited. The method comprises the steps of selecting a large language model, and selecting a large language model, wherein the evaluation efficiency of the sketch outline can be effectively improved by using the large language model, wherein the sketch outline to be adjusted can be screened by using the evaluation threshold, the adjustment quantity of the sketch outline is reduced, the generation efficiency of articles is further improved, the generated modification opinion can be used for guiding a user to conduct modification in a targeted manner, the quality of the sketch outline after adjustment is further improved, and therefore the generation efficiency and the automation degree of the articles are improved.

In some embodiments, if it is determined that the evaluation value of the sketch outline is smaller than the preset evaluation threshold, the sketch outline is adjusted according to the improvement information, and in the process of obtaining the adjusted sketch outline, the sketch outline can be adjusted according to the modification content indicated by the modification instruction and the modification instruction input by the user based on the improvement information, so that the adjusted sketch outline is obtained; wherein the modification instruction indicates modification content for the authoring outline. Specifically, the embodiment of the disclosure does not limit the manner of adjusting the sketch outline based on the modification content indicated by the modification instruction, and optionally, the content to be modified in the sketch outline indicated by the modification instruction may be determined, and the content to be modified is replaced by the modification content to obtain the adjusted sketch outline. The writing outline is adjusted according to the modification instruction of the user, so that the writing outline meets the user requirement better on the basis of improving the quality, the generated article meets the user requirement better, and the article generation efficiency and the automation degree are improved.

In some embodiments, in the process of generating the article to be generated according to the adjusted sketch outline and the sketch material, a third prompting word corresponding to the adjusted sketch outline may be determined according to a preset third prompting word template, the adjusted sketch outline and the sketch material; the preset third prompting word template comprises generating conditions of each article to be generated; the third prompt word characterizes the generation result of the article to be generated; and then inputting the third prompt word into a preset large language model, and outputting the article to be generated. The generating conditions of the article to be generated include, but are not limited to: the matching requirement of the writing outline and the requirement of writing materials contained in the article, and optionally, the generating conditions of the article to be generated can further include: in the process of determining the third prompting word corresponding to the adjusted writing outline according to the preset third prompting word template, the adjusted writing outline and the writing material, the preset third prompting word template, the adjusted writing outline and the writing material can be combined to generate the third prompting word corresponding to the adjusted writing outline, and optionally, the preset third prompting word template comprises the field to be filled of the adjusted writing outline and the field to be filled of the writing material besides the generation condition of each article to be generated described above, and the adjusted writing outline and the writing material can be respectively filled into the corresponding field to be filled to generate the third prompting word corresponding to the adjusted writing outline. The description of the large language model and the preset large language model may refer to the description above, which is not repeated herein, and in the embodiment of the present disclosure, the preset large language model may output, according to the third prompting word, a generating result of the article to be generated, that is, the article to be generated, which is characterized by the third prompting word. The method for automatically generating the articles can improve the efficiency of the automatic article generating method by generating articles to be generated through a preset large language model. Based on the third prompt word generated according to the adjusted sketch outline, the article to be generated is generated, the accuracy of the article can be improved, and the generation efficiency and the automation degree of the article are further improved.

Fig. 2 is a schematic diagram according to a second embodiment of the present disclosure. As shown in fig. 2, a process for obtaining a composition outline and a composition material of an article to be generated according to a second embodiment of the present disclosure includes:

step S201, acquiring a writing theme and a reference article of an article to be generated; the writing theme characterizes the theme of the article to be generated; wherein the reference article characterizes a reference outline of the article to be generated.

Specifically, the server may acquire a sketching subject and a reference article of an article to be generated; the writing theme characterizes the theme of the article to be generated; wherein the reference article characterizes a reference outline of the article to be generated. The method for acquiring the sketching theme of the article to be generated is not limited, and optionally, the sketching theme input by the user in the theme input box of the article generation interface can be acquired, and optionally, the sketching theme selected by the user from the candidate theme can be acquired based on the candidate theme provided in the article generation interface. The method for obtaining the reference is not limited in this disclosure, and alternatively, the reference article uploaded by the user by clicking the file upload button in the article generation interface may be obtained.

And step S202, acquiring the authoring material according to the authoring subject.

Specifically, the authoring material may be acquired according to the authoring subject of the article to be generated acquired in step S201. The method for acquiring the authoring material according to the authoring subject is not limited, and optionally, a material source of an article to be generated can be determined first; wherein the material source characterizes the source of the initial authoring material, in particular, the present disclosure does not limit the material source, wherein the material source includes, but is not limited to: the method comprises the steps of pre-constructing a material library and retrieving the material library, wherein the retrieving the material library comprises but is not limited to: news search stories, encyclopedia search stories, internet search stories, or a combination thereof. The process of constructing the pre-constructed material library may include: determining a theme of a material library, a material source and a material screening standard; acquiring material data according to a material library theme, a material source and a material screening standard based on modes such as internet searching, field data acquisition and web crawlers; and cleaning, classifying and storing the material data to obtain a material library corresponding to the material library theme. The method for determining the material source of the article to be generated is not limited in the present disclosure, and alternatively, the material source selected by the user from the material sources to be selected may be obtained based on the material sources to be selected provided in the article generation interface.

Optionally, after determining the material source of the article to be generated, initial authoring material may be obtained from the material source according to the authoring subject; the method comprises the steps of firstly, writing materials to represent materials directly obtained from a material source; specifically, the method for obtaining the initial writing materials from the material sources according to the writing subjects is not limited, and optionally, the correlation strength of each initial writing material in the writing subjects and the material sources can be determined, and the initial writing materials are obtained from the material sources according to a preset correlation strength threshold.

Optionally, after the initial authoring material is obtained, a type of the initial authoring material may be determined; the type of the initial writing material characterizes the text structural characteristics of the initial writing material; and processing the initial writing material according to the type of the initial writing material to obtain the writing material. If the division basis is different, the types of the initial writing materials are different, and in the disclosure, the types of the initial writing materials represent the text structural features of the initial writing materials, so that the text result features can be used as the division basis, and the initial writing materials are divided into strict structure materials and/or non-strict structure materials, wherein the strict structure materials refer to materials which have high requirements on the text structure and need to be cited in original text when being cited, for example: legal provision and other materials, wherein the material with a non-strict structure refers to a material with low requirements on a text structure and no need of original text reference during reference, for example: story cases, and the like. The material processing is used for simplifying the acquired initial writing material, and is because the preset large language model has a certain number of character limits on the input prompt words, and the writing material is used for determining the generation of the outline prompt words and the generation of the prompt words of the articles to be generated, so that the material processing can be performed on the initial writing material after the initial writing material is acquired. Specifically, according to the description of the type of the initial writing material above, in the process of processing the initial writing material, different modes of material processing are required to be performed on the initial writing material of different types, so that the accuracy of material processing is improved, and the accuracy of the acquired writing material is further improved.

Optionally, the types of the initial authoring material include: in the process of processing the initial writing material according to the type of the initial writing material, if the type of the initial writing material is the structure-strict material, performing fragment extraction processing on the initial writing material to obtain the writing material, wherein the method for performing fragment extraction processing on the initial writing material is not limited, and optionally, the process for performing fragment extraction processing on the initial writing material comprises the following steps: performing paragraph segmentation processing on the initial writing materials to obtain a plurality of initial paragraph materials; according to the preset character segment length, carrying out character segment segmentation processing on each initial paragraph material to obtain a plurality of initial character segment materials; determining a vector of each initial character segment material; calculating the correlation strength of the theme and each initial character segment material according to the vector of the written theme and the vector of each initial character segment material; determining character segment materials according to a preset correlation strength threshold; determining and recalling corresponding initial paragraph materials according to the character segment materials; the recalled initial paragraph material is taken as the authoring material. Optionally, if multiple initial paragraph materials are recalled, the ranking may be performed according to the correlation strength of each initial paragraph material, so that the authoring material is selected from the initial paragraph material sequence according to a screening instruction determined by the user based on the initial paragraph material sequence. And if the type of the initial writing material is a structural non-strict material, carrying out induction and summarization treatment on the initial writing material according to a preset large language model to obtain the writing material. The method for carrying out the induction and summarization processing on the initial writing materials according to the preset large language model is not limited, and optionally, the material induction and summarization prompt words can be determined according to the initial writing materials and the preset induction and summarization template, and the material induction and summarization prompt words are input into the preset large language model to obtain the writing materials subjected to induction and summarization processing. The accuracy of the material processing can be improved by carrying out fragment extraction processing on the structural strict type material, and the efficiency of the material processing can be improved by carrying out fragment extraction on the structural non-strict type material through a preset large language model.

In a possible embodiment, the initial writing material of the article to be generated may be obtained through other manners, optionally, the initial writing material input by the user in the material input box of the article generation interface may be obtained, optionally, the writing material in the article to be referred to obtained in step S201 may be extracted and used as the initial writing material of the article to be generated. The material processing may be performed on the initial authoring material acquired by other manners according to the material processing procedure described above, which is not described herein.

And step 203, determining a sketch outline according to the reference article, the sketch theme and the sketch material.

Specifically, the authoring outline may be determined based on the reference article and the authoring subject acquired in step S201, and the authoring material acquired in step S202. The method and the device have the advantages that the process of determining the sketch outline according to the reference article, the sketch theme and the sketch material is not limited, and optionally, the reference outline can be extracted from the reference article; the method for extracting the reference outline from the reference article is not limited, and optionally, the reference outline prompting word can be generated according to a preset reference outline prompting word template and the reference article, the reference outline prompting word is input into a preset large language model, and the reference outline corresponding to the reference article is output. After the reference outline is extracted from the reference article, determining a second prompting word corresponding to the writing theme according to a preset second prompting word template, the reference outline, the writing theme and the writing material; the preset second prompting word template comprises a generation condition of each sketch outline; the second prompting word represents a generation result of the writing outline; and inputting the second prompt word into a preset large language model, and outputting the sketch outline. The generation conditions of the sketch outline include, but are not limited to: referring to the outline matching requirement, the writing theme matching requirement and the requirement on writing materials contained in the writing outline, optionally, the generating conditions of the writing outline may further include: in the process of determining the second prompting word corresponding to the sketching theme according to the preset second prompting word template, the reference outline, the sketching theme and the sketching material, the preset second prompting word template, the reference outline, the sketching theme and the sketching material can be combined to generate the second prompting word corresponding to the sketching theme, optionally, the preset second prompting word template comprises the generating condition of each sketching outline described above, and further comprises a field to be filled in of the reference outline, a field to be filled in of the sketching theme and a field to be filled in of the sketching material, and the reference outline, the sketching theme and the sketching material can be respectively filled in the corresponding fields to be filled in to generate the second prompting word corresponding to the sketching theme. The description of the large language model and the preset large language model may refer to the description above, which is not repeated herein, and in the embodiment of the present disclosure, the preset large language model may output, according to the second prompting word, a result of generating the writing outline represented by the second prompting word, that is, the writing outline to be generated. The second prompt word is generated by referring to the outline, so that the generated sketch outline meets the requirements of users, and the to-be-generated sketch outline is generated by a preset large language model, so that the efficiency of the automatic generation method of the articles can be improved.

In the embodiment of the disclosure, a writing theme and a reference article of an article to be generated are acquired; the writing theme characterizes the theme of the article to be generated; the reference article characterizes a reference outline of the article to be generated, the writing material is obtained according to the writing theme, the writing outline is determined according to the reference article, the writing theme and the writing material, wherein the writing material is obtained according to the writing theme, the correlation between the material and the theme is improved, the accuracy of the writing outline is further improved, and the writing outline is determined according to the reference article, the writing theme and the writing material, so that the writing outline meets the requirements of users.

Fig. 3 is a schematic diagram according to a third embodiment of the present disclosure. As shown in fig. 3, a process of generating an article to be generated after a color rendering process performed on the article to be generated according to a third embodiment of the present disclosure includes:

and step S301, acquiring the type of the rendering processing, and acquiring the writing segment to be rendered from the article to be generated.

Specifically, the type of the rendering process may be obtained, and the writing segment to be rendered may be obtained from the article to be generated, where the type of the rendering process includes: generating an oil poem for a writing clip, generating a mark-up word for the writing clip, summarizing the writing clip, expanding the writing clip, renewing the writing clip, and sensitive word checking the writing clip, wherein the method is not limited to the type of the obtained color, alternatively, the type of the color selected by a user from the types of the selected color can be obtained based on the types of the selected color provided in an article generating interface, wherein the method is not limited to the type of the color selected by the user from the types of the selected color, alternatively, the method can obtain the writing clip of the color selected by the user from the articles to be generated based on the articles to be generated provided in the article generating interface, wherein the color selected by the user from the articles to be generated comprises one or more writing clips

Step S302, determining a fourth prompting word corresponding to the type of the color rendering according to a preset fourth prompting word template, the type of the color rendering and the segment to be color rendering; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word characterizes the color rendering processing result of the writing fragment.

Specifically, the fourth prompting word corresponding to the type of the color rendering process may be determined according to the preset fourth prompting word template, the type of the color rendering process obtained in step S301, and the segment to be color rendering process; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word characterizes the color rendering processing result of the writing fragment. Wherein, the coloring treatment conditions of the writing fragment include but are not limited to: the requirements of the type of the rendering process and the requirements of the fragments to be rendered, specifically, in the process of determining the fourth prompting word corresponding to the type of the rendering process according to the preset fourth prompting word template, the type of the rendering process and the fragments to be rendered, the preset fourth prompting word template, the type of the rendering process and the fragments to be rendered can be combined to generate the fourth prompting word corresponding to the rendering type, and optionally, the preset fourth prompting word template comprises the field to be filled of the type of the rendering process and the field to be filled of the fragments to be rendered besides the rendering process condition of each writing fragment described above, and the type of the rendering process and the fragments to be rendered can be respectively filled into the corresponding fields to be filled to generate the fourth prompting word corresponding to the rendering type. The fourth prompting word is generated through the type of the color rendering processing and the writing segment to be subjected to the color rendering processing, so that the generated segment after the color rendering processing can better meet the requirements of users.

Step S303, inputting the fourth prompt word into a preset large language model, and outputting the segment after the color rendering processing.

Specifically, the fourth prompting word determined in step S302 may be input into a preset large language model, and the segment after the color rendering processing is output, where the description of the large language model and the preset large language model may refer to the description above, which is not repeated herein. The method comprises the steps of generating a fragment after the color rendering processing through a preset large language model, and improving the efficiency of an automatic article generating method.

And step S304, generating the article to be generated after the color rendering treatment according to the article to be generated and the segment after the color rendering treatment.

Specifically, after obtaining the post-conditioning segment, the post-conditioning to-be-generated article may be generated according to the to-be-generated article and the post-conditioning segment, where the process of generating the post-conditioning to-be-generated article according to the to-be-generated article and the post-conditioning segment is not limited in the present disclosure, and optionally, if the type of the conditioning is one or more of generating an oil poem for the writing segment, generating a skin tone for the writing segment, and writing the writing segment, the post-conditioning segment may be inserted into a corresponding position in the to-be-conditioning segment, and generating the post-conditioning to-be-generated article may be selected, and if the type of the conditioning is one or more of summarizing the writing segment, expanding the writing segment, and performing a sensitive word inspection for the writing segment, the post-conditioning segment may be replaced with the corresponding to-be-conditioning segment, and the post-conditioning to-be generated article may be generated.

In the embodiment of the disclosure, by acquiring the type of the color rendering processing, acquiring a writing segment to be color rendering processing from an article to be generated, and determining a fourth prompting word corresponding to the type of the color rendering processing according to a preset fourth prompting word template, the type of the color rendering processing and the segment to be color rendering processing; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word represents a result of the color rendering processing of the writing segment, the fourth prompting word is input into a preset large language model, the segment after the color rendering processing is output, the article to be generated after the color rendering processing is generated according to the article to be generated and the segment after the color rendering processing, the article to be generated is subjected to the color rendering processing, various optional color rendering processing types are set, the richness of the article to be generated can be improved, and the efficiency of the color rendering processing can be improved by combining the preset large language model in the color rendering processing process.

Fig. 4 is a schematic diagram according to a fourth embodiment of the present disclosure. As shown in fig. 4, the processing for inserting an article image into an article to be generated according to the fourth embodiment of the present disclosure includes:

Step S401, generating a to-be-selected article image list according to the sketch outline and/or the to-be-generated articles; the article image list to be selected characterizes article pictures and/or article videos selected by a user.

Specifically, a list of images of articles to be selected may be generated according to the sketch and/or the articles to be generated; the method comprises the steps that a to-be-selected article image list represents article pictures and/or article videos for users to select, wherein the method for generating the to-be-selected article image list according to the sketch and/or the to-be-generated articles is not limited, and optionally, a to-be-selected article image set can be obtained according to the sketch and/or the to-be-generated articles; the material source of the article image to be selected is not limited in the disclosure, and optionally may include: the pre-constructed material library and/or the search material library can be used for acquiring the vector of each article image to be selected optionally; calculating the correlation strength between the article image to be selected according to the vector of each title and/or each outline in the sketch outline or the vector of each paragraph in the article to be generated; and acquiring an image set of the article to be selected according to a preset correlation strength threshold. Optionally, the image set of the article to be selected may be obtained according to the sketching theme, and the specific process may refer to the above-described process of obtaining the image set of the article to be selected according to the sketching outline and/or the article to be generated, which is not described herein. Optionally, after the article image set to be selected is obtained, the reference frequency of each article image to be selected in the article image set to be selected can be obtained; wherein, the reference frequency includes: download frequency, click view frequency. Optionally, after the reference frequency of each article image to be selected is obtained in the article image set to be selected, each article image to be selected may be ranked according to the reference frequency, so as to generate the article image list to be selected. On the basis of acquiring the article image collection to be selected, a list of article images to be selected is generated according to the reference frequency of each article image to be selected, so that the user can select conveniently, and the user experience is improved.

Optionally, an update period of the article image list to be selected may be preset, so that the server updates the article image list to be selected according to the preset update period, thereby ensuring timeliness of the article image.

Step S402, responding to a selection instruction input by a user based on a to-be-selected article image list, inserting an article image indicated by the selection instruction into an article to be generated, and generating the article to be generated comprising the article image; wherein the selection instruction indicates the article image selected by the user.

Specifically, after the candidate article image list is generated, the article images indicated by the selection instruction can be inserted into the articles to be generated in response to the selection instruction input by the user based on the candidate article image list, and the articles to be generated including the article images can be generated; wherein the selection instruction indicates the article image selected by the user. The method for inserting the article image indicated by the selection instruction into the article to be generated is not limited, and alternatively, the article image indicated by the selection instruction can be obtained; determining titles and/or outlines in the corresponding sketch outline or paragraphs in the article to be generated according to the article image; the article image is inserted into the title and/or outline written in the sketch or at the insertion position of the article to be generated indicated by the paragraph in the article to be generated. Alternatively, an article image indicated by the selection instruction may be acquired; acquiring a position selection instruction input by a user based on an article image; and inserting the article image into the insertion position of the article to be generated indicated by the position selection instruction. Optionally, before inserting the article image into the article to be generated, the authority information of the article image may be detected, so as to ensure that the article picture can be cited.

In the embodiment of the disclosure, the image list of the articles to be selected is generated according to the sketch outline and/or the articles to be generated; the article image list to be selected characterizes an article picture and/or an article video for the user to select, and an article image indicated by a selection instruction is inserted into an article to be generated to generate the article to be generated, wherein the article image to be selected comprises the article image to be selected and the article video; the selection instruction indicates an article image selected by a user, wherein the article image insertion processing is performed on the article to be generated, and the selectable article image list to be selected is set to improve the richness and the expression form of the article to be generated, wherein the introduction frequency of the article image is combined in the generation process of the article image list to be selected, so that the user experience of the article image insertion processing can be improved.

In some embodiments, after the adjusted sketch outline is generated, feedback parameters of the user on the adjusted sketch outline may be obtained, where the feedback parameters include: praise frequency and click step frequency; and secondly training the preset large language model according to the feedback parameters and the training period of the preset large language model so as to improve the user experience of the preset large language model.

Fig. 5 is a schematic diagram of a fifth embodiment of the present disclosure. As shown in fig. 5, an automatic article generating apparatus 500 according to a sixth embodiment of the present disclosure includes:

an obtaining unit 501, configured to obtain a sketch outline and a sketch material of an article to be generated; the method comprises the steps of writing outline to represent outline of an article to be generated, and writing material to represent material of the article to be generated; determining a first prompt word according to the sketch outline, wherein the first prompt word characterizes an evaluation result of the sketch outline;

the adjusting unit 502 is configured to adjust the sketch outline according to the first prompt word, so as to obtain an adjusted sketch outline;

and the generating unit 503 is configured to generate an article to be generated according to the adjusted sketch outline and the adjusted sketch material.

In some embodiments, the acquisition unit 501 includes: a first determining subunit (not shown in the figure) configured to determine, according to a preset first prompting word template and a sketch outline, a first prompting word corresponding to the sketch outline; the preset first prompting word template comprises evaluation conditions of each sketch outline.

In some embodiments, the adjustment unit 502 includes: an evaluation subunit (not shown in the figure) for inputting the first prompt word into a preset large language model and outputting the sketch outline evaluation value and the improvement information; wherein the improvement information characterizes a modification suggestion for the authoring outline; and the adjusting subunit (not shown in the figure) is used for adjusting the sketch outline according to the improvement information if the evaluation value of the sketch outline is smaller than the preset evaluation threshold value, so as to obtain the adjusted sketch outline.

In some embodiments, the adjustment subunit (not shown in the figures) includes: an adjustment module (not shown in the figure) for adjusting the sketch outline based on the modification content indicated by the modification instruction in response to the modification instruction input by the user based on the improvement information, to obtain an adjusted sketch outline; wherein the modification instruction indicates modification content for the authoring outline.

In some embodiments, the acquisition unit 501 includes: a first obtaining subunit (not shown in the figure) for obtaining a writing topic and a reference article of an article to be generated; the writing theme characterizes the theme of the article to be generated; wherein, the reference article characterizes a reference outline of the article to be generated; a second acquisition subunit (not shown in the figure) for acquiring the authoring material according to the authoring subject; a second determining subunit (not shown in the figure) for determining the sketch outline according to the reference article, the sketch theme and the sketch material.

In some embodiments, the second acquisition subunit (not shown in the figures) comprises: a first determining module (not shown in the figure) for determining a material source of an article to be generated; wherein the source of material characterizes the source of the initial authoring material; a first obtaining module (not shown in the figure) for obtaining initial authoring material from a material source according to an authoring subject; the method comprises the steps of firstly, writing materials to represent materials directly obtained from a material source; a second determining module (not shown in the figure) for determining the type of the initial authoring material; the type of the initial writing material characterizes the text structural characteristics of the initial writing material; and processing the initial writing material according to the type of the initial writing material to obtain the writing material.

In some embodiments, wherein the types of initial authoring material include: structural stringent and/or structural non-stringent materials, a second determining module (not shown) comprising: the first processing submodule (not shown in the figure) is used for extracting fragments of the initial writing material to obtain the writing material if the type of the initial writing material is a material with strict structure; and the second processing sub-module (not shown in the figure) is used for carrying out induction summarization processing on the initial writing materials according to a preset large language model to obtain the writing materials if the type of the initial writing materials is the non-rigorous structure materials.

In some embodiments, the second determining subunit (not shown in the figures) comprises: an extraction module (not shown in the figure) for extracting a reference outline from the reference article; a third determining module (not shown in the figure) for determining a second prompting word corresponding to the sketching theme according to a preset second prompting word template, a reference outline, the sketching theme and the sketching material; the preset second prompting word template comprises a generation condition of each sketch outline; the second prompting word represents a generation result of the writing outline; the first generation module (not shown in the figure) is used for inputting the second prompt word into a preset large language model and outputting the sketch outline.

In some embodiments, the generating unit 503 includes: a third determining subunit (not shown in the figure) configured to determine a third prompting word corresponding to the adjusted sketch outline according to a preset third prompting word template, the adjusted sketch outline and the sketch material; the preset third prompting word template comprises generating conditions of each article to be generated; the third prompt word characterizes the generation result of the article to be generated; a first generating subunit (not shown in the figure) is configured to input the third prompt word into a preset large language model, and output an article to be generated.

In some embodiments, further comprising: a color unit comprising: a third obtaining subunit (not shown in the figure) for obtaining the type of the rendering process and obtaining the writing segment to be rendered from the article to be generated; a fourth determining subunit (not shown in the figure) configured to determine a fourth prompting word corresponding to the type of the rendering process according to a preset fourth prompting word template, the type of the rendering process, and the segment to be rendered; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word represents a color rendering processing result of the writing fragment; a second generating subunit (not shown in the figure) for inputting the fourth prompting word into a preset large language model and outputting the segment after the color rendering processing; and a third generating subunit (not shown in the figure) for generating the article to be generated after the rendering processing according to the article to be generated and the segment after the rendering processing.

In some embodiments, further comprising: an insertion unit comprising: a fourth generating subunit (not shown in the figure) for generating a list of images of articles to be selected according to the sketch and/or the articles to be generated; the article image list to be selected characterizes article pictures and/or article videos selected by a user; an inserting subunit (not shown in the figure) for, in response to a selection instruction input by a user based on the article image list to be selected, inserting an article image indicated by the selection instruction into the article to be generated, and generating the article to be generated including the article image; wherein the selection instruction indicates the article image selected by the user.

In some embodiments, a fourth generation subunit (not shown in the figures) comprises: the second acquisition module (not shown in the figure) is used for acquiring an image set of the article to be selected according to the sketch outline and/or the article to be generated; a third obtaining module (not shown in the figure) for obtaining the reference frequency of each article image to be selected in the article image set to be selected; and the second generation module (not shown in the figure) is used for sequencing the article images to be selected according to the reference frequency to generate a list of article images to be selected.

The automatic generation device of the article provided in fig. 5 may perform the steps related to the server in the above corresponding method embodiments, and the implementation principle and technical effects are similar, which are not described herein again.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

According to an embodiment of the present disclosure, there is also provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform the solution provided by any one of the above embodiments.

According to an embodiment of the present disclosure, the present disclosure also provides a computer program product comprising: a computer program stored in a readable storage medium, from which at least one processor of an electronic device can read, the at least one processor executing the computer program causing the electronic device to perform the solution provided by any one of the embodiments described above.

Fig. 6 is a schematic block diagram of an example electronic device 600 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the disclosure described and/or claimed herein.

As shown in fig. 6, the apparatus 600 includes a computing unit 601 that can perform various appropriate actions and processes according to a computer program stored in a Read Only Memory (ROM) 602 or a computer program loaded from a storage unit 608 into a Random Access Memory (RAM) 603. In the RAM603, various programs and data required for the operation of the device 600 may also be stored. The computing unit 601, ROM 602, and RAM603 are connected to each other by a bus 604. An input/output (I/O) interface 605 is also connected to bus 604.

Various components in the device 600 are connected to the I/O interface 605, including: an input unit 606 such as a keyboard, mouse, etc.; an output unit 607 such as various types of displays, speakers, and the like; a storage unit 608, such as a magnetic disk, optical disk, or the like; and a communication unit 609 such as a network card, modem, wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information/data with other devices via a computer network, such as the internet, and/or various telecommunication networks.

The computing unit 601 may be a variety of general and/or special purpose processing components having processing and computing capabilities. Some examples of computing unit 601 include, but are not limited to, a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), various specialized Artificial Intelligence (AI) computing chips, various computing units running machine learning model algorithms, a Digital Signal Processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the respective methods and processes described above, such as an automatic generation method of an article. For example, in some embodiments, the method of automatically generating articles may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program may be loaded and/or installed onto the device 600 via the ROM 602 and/or the communication unit 609. When a computer program is loaded into the RAM603 and executed by the computing unit 601, one or more steps of the above-described automatic article generating method may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the automatic generation method of articles by any other suitable means (e.g., by means of firmware).

Various implementations of the systems and techniques described here above may be implemented in digital electronic circuitry, integrated circuit systems, field Programmable Gate Arrays (FPGAs), application Specific Integrated Circuits (ASICs), application Specific Standard Products (ASSPs), systems On Chip (SOCs), complex Programmable Logic Devices (CPLDs), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include: implemented in one or more computer programs, the one or more computer programs may be executed and/or interpreted on a programmable system including at least one programmable processor, which may be a special purpose or general-purpose programmable processor, that may receive data and instructions from, and transmit data and instructions to, a storage system, at least one input device, and at least one output device.

Program code for carrying out methods of the present disclosure may be written in any combination of one or more programming languages. These program code may be provided to a processor or controller of a general purpose computer, special purpose computer, or other programmable data processing apparatus such that the program code, when executed by the processor or controller, causes the functions/operations specified in the flowchart and/or block diagram to be implemented. The program code may execute entirely on the machine, partly on the machine, as a stand-alone software package, partly on the machine and partly on a remote machine or entirely on the remote machine or server.

In the context of this disclosure, a machine-readable medium may be a tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of a machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

To provide for interaction with a user, the systems and techniques described here can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and pointing device (e.g., a mouse or trackball) by which a user can provide input to the computer. Other kinds of devices may also be used to provide for interaction with a user; for example, feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form, including acoustic input, speech input, or tactile input.

The systems and techniques described here can be implemented in a computing system that includes a background component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such background, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local Area Networks (LANs), wide Area Networks (WANs), and the internet.

The computer system may include a client and a server. The client and server are typically remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. The server can be a cloud server, also called a cloud computing server or a cloud host, and is a host product in a cloud computing service system, so that the defects of high management difficulty and weak service expansibility in the traditional physical hosts and VPS service ("Virtual Private Server" or simply "VPS") are overcome. The server may also be a server of a distributed system or a server that incorporates a blockchain.

It should be appreciated that various forms of the flows shown above may be used to reorder, add, or delete steps. For example, the steps recited in the present disclosure may be performed in parallel or sequentially or in a different order, provided that the desired results of the technical solutions of the present disclosure are achieved, and are not limited herein.

The above detailed description should not be taken as limiting the scope of the present disclosure. It will be apparent to those skilled in the art that various modifications, combinations, sub-combinations and alternatives are possible, depending on design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure are intended to be included within the scope of the present disclosure.

Claims

1. An automatic article generation method, comprising:

2. The method of claim 1, wherein determining a first hint word from the composition outline comprises:

determining a first prompting word corresponding to the sketch outline according to a preset first prompting word template and the sketch outline; the preset first prompting word template comprises evaluation conditions of each writing outline.

3. The method according to claim 1 or 2, wherein adjusting the composition outline according to the first prompt word, to obtain an adjusted composition outline, comprises:

inputting the first prompt word into a preset large language model, and outputting a sketch outline evaluation value and improvement information; wherein the improvement information characterizes a modification suggestion for the authoring outline;

and if the evaluation value of the writing outline is smaller than the preset evaluation threshold value, adjusting the writing outline according to the improvement information to obtain an adjusted writing outline.

4. The method of claim 3, wherein if it is determined that the sketch outline evaluation value is smaller than a preset evaluation threshold, adjusting the sketch outline according to the improvement information to obtain an adjusted sketch outline, including:

Responding to a modification instruction input by a user based on the improvement information, and adjusting the sketch outline based on modification content indicated by the modification instruction to obtain an adjusted sketch outline;

wherein the modification instruction indicates modification content of the sketch outline.

5. The method of any of claims 1-4, wherein obtaining the authoring synopsis and authoring material of the article to be generated comprises:

acquiring a writing theme and a reference article of an article to be generated; wherein the authoring subject characterizes a subject of an article to be generated; the reference article characterizes a reference outline of the article to be generated;

acquiring the writing materials according to the writing subjects;

and determining the sketch outline according to the reference article, the sketch theme and the sketch material.

6. The method of claim 5, wherein obtaining the authoring material in accordance with the authoring subject comprises:

determining a material source of an article to be generated; wherein the source of material characterizes the source of the initial authoring material;

acquiring initial writing materials from the material sources according to the writing subjects; the initial authoring material characterizes the material directly acquired from the material source;

Determining the type of the initial writing material; the type of the initial writing material characterizes the text structural characteristics of the initial writing material; and processing the initial writing material according to the type of the initial writing material to obtain the writing material.

7. The method of claim 6, wherein the type of initial authoring material comprises: and carrying out material processing on the initial writing material according to the type of the initial writing material to obtain the writing material, wherein the material processing comprises the following steps:

if the type of the initial writing material is the material with strict structure, carrying out fragment extraction processing on the initial writing material to obtain the writing material;

and if the type of the initial writing material is the non-rigorous structural material, carrying out induction and summarization treatment on the initial writing material according to a preset large language model to obtain the writing material.

8. The method of claim 5, wherein determining the authoring synopsis based on the reference article, the authoring subject and the authoring material comprises:

extracting the reference outline from the reference article;

Determining a second prompting word corresponding to the sketching theme according to a preset second prompting word template, the reference outline, the sketching theme and the sketching material; the preset second prompting word template comprises a generation condition of each sketch outline; the second prompting word represents a generation result of the writing outline;

and inputting the second prompting word into a preset large language model, and outputting the sketch outline.

9. The method of any of claims 1-8, wherein generating the article to be generated from the adjusted authoring synopsis and the authoring material comprises:

determining a third prompting word corresponding to the adjusted sketch outline according to a preset third prompting word template, the adjusted sketch outline and the sketch material; the preset third prompting word template comprises generating conditions of each article to be generated; the third prompt word characterizes a generation result of the article to be generated;

and inputting the third prompt word into the preset large language model, and outputting the article to be generated.

10. The method of any of claims 1-9, further comprising:

Acquiring the type of the color rendering treatment, and acquiring a writing segment to be subjected to the color rendering treatment from the article to be generated;

determining a fourth prompting word corresponding to the type of the color rendering according to a preset fourth prompting word template, the type of the color rendering and the segment to be color rendering; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word represents a color rendering processing result of the writing segment;

inputting the fourth prompting word into a preset large language model, and outputting a fragment after the color rendering treatment;

and generating the article to be generated after the color rendering treatment according to the article to be generated and the segment after the color rendering treatment.

11. The method of any of claims 1-10, further comprising:

generating a to-be-selected article image list according to the sketch outline and/or the to-be-generated articles; the article image list to be selected characterizes an article picture and/or an article video for the user to select;

responding to a selection instruction input by a user based on the article image list to be selected, inserting an article image indicated by the selection instruction into the article to be generated, and generating the article to be generated comprising the article image; wherein the selection instruction indicates an article image selected by the user.

12. The method of claim 11, wherein the generating a list of candidate article images from the authoring synopsis and/or the candidate articles further comprises:

acquiring an image set of the article to be selected according to the sketch outline and/or the article to be generated;

acquiring the reference frequency of each article image to be selected in the article image set to be selected;

and sorting the article images to be selected according to the reference frequency, and generating the article image list to be selected.

13. An automatic article generation device, comprising:

14. The apparatus of claim 13, wherein the acquisition unit comprises:

The first determining subunit is used for determining a first prompting word corresponding to the sketch outline according to a preset first prompting word template and the sketch outline; the preset first prompting word template comprises evaluation conditions of each writing outline.

15. The apparatus according to claim 13 or 14, wherein the adjustment unit comprises:

the evaluation subunit is used for inputting the first prompt word into a preset large language model and outputting a sketch outline evaluation value and improvement information; wherein the improvement information characterizes a modification suggestion for the authoring outline;

and the adjustment subunit is used for adjusting the writing outline according to the improvement information if the writing outline evaluation value is smaller than the preset evaluation threshold value, so as to obtain the adjusted writing outline.

16. The apparatus of claim 15, wherein the adjustment subunit comprises:

the adjusting module is used for responding to the modification instruction input by the user based on the improvement information, adjusting the sketch outline based on the modification content indicated by the modification instruction, and obtaining an adjusted sketch outline; wherein the modification instruction indicates modification content of the sketch outline.

17. The apparatus according to any one of claims 13-16, wherein the acquisition unit comprises:

the first acquisition subunit is used for acquiring the writing theme and the reference article of the article to be generated; wherein the authoring subject characterizes a subject of an article to be generated; the reference article characterizes a reference outline of the article to be generated;

the second acquisition subunit is used for acquiring the authoring material according to the authoring subject;

and the second determining subunit is used for determining the writing outline according to the reference article, the writing theme and the writing material.

18. The apparatus of claim 17, wherein the second acquisition subunit comprises:

the first determining module is used for determining a material source of an article to be generated; wherein the source of material characterizes the source of the initial authoring material;

the first acquisition module is used for acquiring initial sketching materials from the material sources according to the sketching subjects; the initial authoring material characterizes the material directly acquired from the material source;

the second determining module is used for determining the type of the initial writing material; the type of the initial writing material characterizes the text structural characteristics of the initial writing material; and processing the initial writing material according to the type of the initial writing material to obtain the writing material.

19. The apparatus of claim 18, wherein the type of initial authoring material comprises: a structural stringent material and/or a structural non-stringent material, the second determining module comprising:

the first processing submodule is used for extracting fragments of the initial writing material to obtain the writing material if the type of the initial writing material is the material with strict structure;

and the second processing sub-module is used for carrying out induction summarization processing on the initial writing materials according to a preset large language model to obtain the writing materials if the type of the initial writing materials is the non-rigorous structural materials.

20. The apparatus of claim 17, wherein the second determination subunit comprises:

the extraction module is used for extracting the reference outline from the reference article;

the third determining module is used for determining a second prompt word corresponding to the sketching theme according to a preset second prompt word template, the reference outline, the sketching theme and the sketching material; the preset second prompting word template comprises a generation condition of each sketch outline; the second prompting word represents a generation result of the writing outline;

The first generation module is used for inputting the second prompt word into a preset large language model and outputting the sketch outline.

21. The apparatus according to any one of claims 13-20, wherein the generating unit comprises:

the third determining subunit is used for determining a third prompting word corresponding to the adjusted writing outline according to a preset third prompting word template, the adjusted writing outline and the writing material; the preset third prompting word template comprises generating conditions of each article to be generated; the third prompt word characterizes a generation result of the article to be generated;

the first generation subunit is used for inputting the third prompt word into the preset large language model and outputting the article to be generated.

22. The apparatus of any of claims 13-21, further comprising: a color unit comprising:

the third acquisition subunit is used for acquiring the type of the color rendering processing and acquiring the writing segment to be subjected to the color rendering processing from the article to be generated;

a fourth determining subunit, configured to determine a fourth prompting word corresponding to the type of the rendering process according to a preset fourth prompting word template, the type of the rendering process, and the segment to be rendered; the preset fourth prompting word template comprises a color rendering processing condition of each writing segment; the fourth prompting word represents a color rendering processing result of the writing segment;

The second generation subunit is used for inputting the fourth prompt word into a preset large language model and outputting a segment after the color rendering treatment;

and the third generation subunit is used for generating the article to be generated after the color rendering processing according to the article to be generated and the segment after the color rendering processing.

23. The apparatus of any of claims 13-22, further comprising: an insertion unit comprising:

the fourth generation subunit is used for generating a to-be-selected article image list according to the sketch outline and/or the to-be-generated article; the article image list to be selected characterizes an article picture and/or an article video for the user to select;

the inserting subunit is used for responding to a selection instruction input by a user based on the article image list to be selected, inserting the article image indicated by the selection instruction into the article to be generated and generating the article to be generated comprising the article image; wherein the selection instruction indicates an article image selected by the user.

24. The apparatus of claim 23, wherein the fourth generation subunit comprises:

the second acquisition module is used for acquiring an image set of the article to be selected according to the sketch outline and/or the article to be generated;

The third acquisition module is used for acquiring the reference frequency of each article image to be selected in the article image set to be selected;

and the second generation module is used for sequencing the article images to be selected according to the reference frequency to generate the article image list to be selected.

25. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-12.

26. A non-transitory computer readable storage medium storing computer instructions for causing the computer to perform the method of any one of claims 1-12.

27. A computer program product comprising a computer program which, when executed by a processor, implements the steps of the method of any of claims 1-12.