Rapid article generation method based on multiple artificial intelligence tools

Abstract

The invention relates to a rapid article generation method based on multiple artificial intelligence tools. The method comprises the steps that the artificial intelligence tools ki expand input content into an article outline according to an article framework; the artificial intelligence tool ki fills the key points of the sub-paragraph Pi and obtains an initial text according to the sentence relation and the expression technique, and after filling is finished, the running time is returned to the processor; and the processor determines the efficiency sequence of the artificial intelligence tool ki according to the running time Ti, and performs one-to-one optimization processing on the paragraphs Sj of the initial text according to the efficiency sequence from high to low. According to the arrangement, the article is filled according to the article framework, and a central thought and a writing technique which can reflect the target article are generated according to the logic of the target article, so that a user can understand deep and difficult-to-understand articles, hidden thoughts and expression techniques; the initial text is processed according to the efficiency sequence through the artificial intelligence tool ki, and the article generation speed and readability are improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence language technology, and in particular to a method for quickly generating articles based on multiple artificial intelligence tools. Background Art

[0002] With the continuous development of big data technology, natural language processing and other artificial intelligence technologies, the exploration and practice of using algorithms to automatically write and generate articles has gradually been launched. Writing mainly examines word order, grammar, content and language logic. The text generation model obtained by training a large amount of high-quality composition data can automatically generate high-quality compositions.

[0003] In the related technology, the following methods are usually adopted when using artificial intelligence tools to write articles: one is to use the relevant information of the article to train the deep neural network model, and then use the trained deep neural network model to generate articles. Since this method completely relies on deep neural networks, the content of the generated articles is uncontrollable and it is easy to have semantic incoherence. The other is to manually set a specific template in a specific field, so as to fill the article information according to the template and improve the content of the article. The generated article is similar to the template only in outline, and the key elements such as the delicate expression of thoughts and feelings that the article should contain and the rich and diverse expression methods are not fully considered. This leads to the fact that when users read such articles, they cannot deeply understand the author's creative intentions and core ideas, and it is even more difficult to learn effective writing skills and strategies from them, so that they cannot have a positive role in promoting their own writing ability. This greatly limits the application value and development prospects of such article generation methods in writing teaching, literary creation assistance and other fields. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a method for rapid article generation based on multiple artificial intelligence tools, the method comprising the following steps:

[0005] The user selects the target article in the sample database;

[0006] Extract the article framework and expression techniques based on the target article to construct the target article code; the target article code includes paragraph S j 、Sub-paragraph P i , sentence relation R n and article difficulty; each paragraph S j Each contains multiple sub-paragraphs P i , each subparagraph P i It includes multiple sentences and there is a sentence relationship R between the multiple sentences n ;

[0007] The artificial intelligence tool iAccording to different sub-paragraphs P i The content input by the user is expanded to form key points according to the main theme, and a plurality of the key points form an article outline;

[0008] Each paragraph S j Different sub-paragraphs P i The key points are input into different artificial intelligence tools i , and different paragraphs S j Subparagraphs P with the same symbol i Artificial intelligence tools with the same symbol k i , the artificial intelligence tool ki will follow the sub-paragraph P i The corresponding sentence relationship and expression method, the sub-paragraph P i Fill in the key points and get the initial text, and return the running time to the processor after filling;

[0009] The processor is based on the running time T i Determine the AI ​​toolk i The effectiveness ranking of the artificial intelligence tool k i In order of effectiveness from high to low, paragraph S of the initial text j One-to-one contextual coherence and optimization.

[0010] Preferably, the shortest value priority sequence sorting mechanism is included, and the processor calculates the artificial intelligence tool k according to the shortest value priority sequence sorting mechanism. i The cumulative average running time ratio is used to determine the AI ​​tool k. i performance ranking.

[0011] Preferably, the shortest value first sequence sorting mechanism comprises the following steps:

[0012] S1. Setting a time range of the processing time T, wherein the time range is divided into short, medium and long;

[0013] S2, each sub-paragraph P i Randomly match AI tools k i , and the artificial intelligence tool k i Returns the processed text content and running time T i ;

[0014] S3. Each AI tool k i Returns the running time T i Compare with the set time range and get the running time T i The time range of

[0015] S4. Calculate the running time T i The time range is short and the number of AI tools k i Returns the running time T i The ratio of the total number of AI tools is determined based on the ratio of the number of AI tools. i performance ranking.

[0016] Preferably, there are two ways to select the target article: first, keywords are obtained based on the article content and article type input by the user, and the target article is searched in the sample database based on the keywords; second, the user selects the target article in the sample database, and optimizes the article content by analogy reasoning between the abstract of the target article and the article content input by the user.

[0017] Preferably, the article types include practical articles and literary creation articles;

[0018] When the article type is a practical article, the article content input by the user will be expanded according to the first expansion method to obtain multiple main words, and the multiple main words are in an OR relationship with each other, and these main words and the article type are in an AND relationship, so as to form keywords;

[0019] When the article type is a literary creation article, the article content input by the user will be expanded according to the second expansion method to obtain multiple main words, and the multiple main words will be in an or relationship with each other, and these main words will be in an and relationship with the article type, so as to form keywords.

[0020] Preferably, the first expansion algorithm is to expand the content of the article input by the user into content similar to the subject of the target article by analogical reasoning;

[0021] The second expansion method is to expand the article content input by the user into related words related to the target article through a big data model.

[0022] Preferably, a text analysis tool is included, and the target article coding is obtained by classifying the contents with similar themes, concepts or functions in the target article into nodes of the same category through the text analysis tool, and analyzing the relationship between multiple nodes to obtain the article framework and expression method.

[0023] Preferably, the sample database classifies articles according to article categories, and the articles in each category are sorted according to the difficulty of the articles.

[0024] Preferably, the artificial intelligence tool k i According to the difficulty of the target article, the input article content is expanded and filled with the corresponding difficulty.

[0025] The beneficial effects of the present invention are as follows: 1. The content input by the user is expanded into the main points of the article according to the main theme of the sub-paragraph Pi through the expansion module, and the main points are filled in according to the sentence relationship and expression method corresponding to the sub-paragraph Pi through different artificial intelligence tools ki, so that the generated article has the same emotional color and main theme as the target article, and the expression method and sentence relationship used in each paragraph are consistent. However, the content input by the user can be changed. We can input things that we can understand as content, and generate an article that can reflect the central idea and writing method of the target article according to the logic of the target article. This is conducive to users' understanding of profound and difficult articles, hidden ideas, expression methods, etc.

[0026] 2. By processing the initial text with the artificial intelligence tool ki in order from high to low efficiency, the speed of article generation can be increased. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying creative labor, other drawings obtained based on these drawings still belong to the scope of the present invention.

[0028] Figure 1 It is a principle block diagram of the present invention;

[0029] Figure 2 A diagram showing the relationship between the article outline of the present invention and the processing of the artificial intelligence tool;

[0030] Figure 3 A relationship diagram for connecting the artificial intelligence tools for the article paragraphs of the present invention;

[0031] Figure 4 This is a diagram of an implementation example of the present invention. DETAILED DESCRIPTION

[0032] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0033] It should be noted that all expressions using "first" and "second" in the embodiments of the present invention are for distinguishing two non-identical entities with the same name or non-identical parameters. It can be seen that "first" and "second" are only for the convenience of expression and should not be understood as limitations on the embodiments of the present invention. The subsequent embodiments will not explain this one by one.

[0034] The directions and positions mentioned in the present invention, such as "upper", "lower", "front", "back", "left", "right", "inside", "outside", "top", "bottom", "side", etc., are only for reference to the directions or positions of the drawings. Therefore, the directions and positions used are for explaining and understanding the present invention, but not for limiting the scope of protection of the present invention.

[0035] like Figures 1 to 4 As shown in the figure, a method for quickly generating articles based on multiple artificial intelligence tools in an embodiment of the present invention comprises the following steps:

[0036] The user selects the target article in the sample database;

[0037] Extract the article framework and expression techniques based on the target article to construct the target article code; the target article code includes paragraph S j 、Sub-paragraph P i , sentence relation R n and article difficulty; each paragraph S j Each contains multiple sub-paragraphs P i , each subparagraph P i It includes multiple sentences and there is a sentence relationship R between the multiple sentences n ;

[0038] The artificial intelligence tool i According to different sub-paragraphs P i The content input by the user is expanded to form key points according to the main theme, and a plurality of the key points form an article outline;

[0039] Each paragraph S j Different sub-paragraphs P i The key points are input into different artificial intelligence tools i , and different paragraphs S j Subparagraphs P with the same symbol i Artificial intelligence tools with the same symbol k i , the artificial intelligence tool ki will follow the sub-paragraph P i The corresponding sentence relationship and expression method, the sub-paragraph P i Fill in the key points and get the initial text, and return the running time to the processor after filling;

[0040] The processor is based on the running time T i Determine the AI ​​toolk i The effectiveness ranking of the artificial intelligence tool k i In order of effectiveness from high to low, paragraph S of the initial text j One-to-one contextual coherence and optimization.

[0041] It can be understood that the content input by the user is expanded into the main points of the article according to the main theme of the sub-paragraph Pi, and the main points are filled in according to the sentence relationships and expression methods corresponding to the sub-paragraph Pi through different artificial intelligence tools ki, so that the generated article has the same emotional color and main theme as the target article, and the expression methods and sentence relationships used in each paragraph are consistent. However, the content input by the user can be changed. We can input things that we can understand as content, and generate an article that can reflect the central idea and writing method of the target article according to the logic of the target article. This is conducive to users' understanding of profound and difficult articles, hidden ideas, expression methods, etc.

[0042] Sentence relations can be expressed as conjunctions. By treating conjunctions as nodes, the sentence relations between paragraphs can be understood through nodes.

[0043] The processor calculates the artificial intelligence tool k according to the shortest value priority sequence sorting mechanism. i The cumulative average running time ratio is used to determine the AI ​​tool k. i performance ranking.

[0044] The shortest value first sequence sorting mechanism comprises the following steps:

[0045] S1. Setting a time range of the processing time T, wherein the time range is divided into short, medium and long;

[0046] S2, each sub-paragraph P i Randomly match AI tools k i , and the artificial intelligence tool k i Returns the processed text content and running time T i ;

[0047] S3. Each AI tool k i Returns the running time T i Compare with the set time range and get the running time T i The time range of

[0048] S4. Calculate the running time T i The time range is short and the number of AI tools k i Returns the running time T i The ratio of the total number of AI tools is determined based on the ratio of the number of AI tools. i performance ranking.

[0049] It can be understood that the present invention can increase the speed of article generation by processing the initial text with the artificial intelligence tool ki in order of efficiency from high to low.

[0050] There are two ways to select the target article: first, obtain keywords based on the article content and article type input by the user, and search for the target article in the sample database based on the keywords; second, the user selects the target article in the sample database, and optimizes the article content by analogy reasoning between the abstract of the target article and the article content input by the user.

[0051] It is understandable that in method one, using the article content and article type input by the user as keywords can easily obtain the target article that the user wants, but in method two, there will be a problem that the target article selected by the user does not match the article content input by the user, so it is necessary to compare the summary of the target article with the article content input by the user to ensure that the article content input by the user is consistent with the target article.

[0052] The types of articles include practical style articles and literary creation articles;

[0053] When the article type is a practical article, the input content is expanded according to the first expansion method to obtain multiple main words, and the multiple main words are in an OR relationship with each other, and the main words and the article type are in an AND relationship, so as to form keywords;

[0054] When the article type is a literary creation article, the input content will be expanded according to the second expansion method to obtain multiple main words, and the multiple main words will be in an or relationship with each other, and these main words will be in an and relationship with the article type, so as to form keywords.

[0055] It is understandable that there are essential differences between practical articles and literary creation articles. Practical articles focus more on whether the matter is real and intuitive, while literary creation articles can be exaggerated and ironic; so when faced with two different types of articles, the method of selecting the corresponding target article based on the content of the article input by the user is also different. Expanding the input content is to expand the scope of the user's search for the target article.

[0056] The first expansion algorithm is to expand the content of the article into content similar to the subject of the target article by analogical reasoning;

[0057] The second expansion method is to expand the article content into synonyms similar to the target article through a big data model.

[0058] It is understandable that when the article type is a practical style article, the article content input by the user is similar to the theme of the target article through analogical reasoning. The similarity here refers to the consistency of emotional color and the type of article object (if the object is the weather, the article content should also be about the weather).

[0059] When the article type is a literary creation article, the article content input by the user can be expanded into similar related words with the target article. The related words refer to words that can be expanded through rhetoric techniques such as personification and metaphor, or words related to the target article, or words related through knowledge spectrum construction, etc.

[0060] The sample database classifies articles according to article categories, and the articles in each category are sorted according to the difficulty of the articles. The difficulty levels are divided into primary school, junior high school, high school, university, etc. Based on the difficulty selected by the user, the filling module will fill in the article outlines corresponding to different difficulties.

[0061] It is understandable that by setting the difficulty of the article, the subsequent use of the expansion module to expand the article content into key points and fill in the article outline based on the key points all need to refer to the article difficulty selected by the user, so that the generated article can be understood by the user.

[0062] It includes a text analysis tool, which classifies the contents with similar themes, concepts or functions in the target article into nodes of the same category, and analyzes the relationship between multiple nodes to obtain the article framework and expression techniques to form the target article code; the target article code includes paragraphs, sub-paragraph themes, sentence relationships, expression techniques and article difficulty.

[0063] It can be understood that through the text analysis tool, each paragraph used in the target article, the main theme of the sub-paragraph, the sentence relationship, the expression method and the difficulty of the article can be identified so that the subsequent expansion module can expand the article content into key points and fill in the article outline based on the key points.

[0064] By coding and classifying the article, sentences with similar themes, concepts or functions are grouped into nodes of the same category, so as to grasp the distribution and relationship of different types of sentences in the paragraph from a macro perspective, and then analyze the paragraph structure. For example, the paragraphs that explain the problem in the article can be grouped into one category, and the paragraphs that analyze the solution can be grouped into another category, so as to determine whether the article is developed according to the structure of "raising a problem-analyzing a problem-solving a problem" or other logical structures.

[0065] The artificial intelligence tool expands the article content into multiple main points according to the target article coding, and the multiple main points form the article outline.

[0066] The artificial intelligence tool fills in the article outline according to the paragraphs, sub-paragraph themes, sentence relationships, expression techniques and article difficulty and obtains the initial text, and returns the filling time to the processor after the filling is completed.

[0067] It is understandable that by using artificial intelligence tools to fill in the article outline and obtain the initial text according to the paragraphs, sub-paragraph themes, sentence relationships, expression techniques and article difficulty, the generated article can be more in line with the target article and help users understand the target article.

[0068] The text analysis tool includes a sentiment pre-training model. Large language models such as BERT (Bidirectional Encoder Representations from Transformers) and GPT (Generative Pretrained Transformer) series can be used to identify sentiment tone. These models are trained on a large amount of text and can understand the semantics and sentiment tendency of the text. For example, BERT can add a special tag (such as [CLS]) at the beginning of a sentence. After being processed by the model, the output vector corresponding to the tag can be used for classification tasks, including sentiment classification. By performing word segmentation on the target article and inputting it into the pre-training model, the model can output the sentiment category of the target article (such as cheerful, sad, etc.).

[0069] The text analysis tools include sentiment analysis libraries such as TextBlob and VADER (Valence Aware Dictionary and sEntiment Reasoner). TextBlob is a simple NLP library that provides sentiment analysis capabilities. It determines the sentiment tone by judging the sentiment polarity (positive or negative) and subjectivity (subjective or objective) of the words in the text. VADER is a tool specifically used for text sentiment analysis such as social media. It has a predefined sentiment vocabulary dictionary that can judge the sentiment intensity and tone based on the appearance of the words in the text and the context. For example, in VADER, the word "excellent" has a higher positive sentiment score, while "terrible" has a higher negative sentiment score.

[0070] Each word in the emotional vocabulary dictionary is marked with an emotional category (such as happiness, sadness, anger, etc.) and an emotional intensity (such as mild, moderate, severe). When expanding key points and filling in text, appropriate words are selected from the vocabulary according to the emotional tone specified by the user. At the same time, a rule engine is established to determine the combination mode and sentence structure between words to meet the specific emotional tone. For example, for the emotional tone of sadness, the rule engine may stipulate that the sentence structure uses long sentences to express deep emotions, and tends to use words such as "sad" and "heartbroken" in the sentence.

[0071] The following takes a long article as an example to illustrate the relationship between the article outline and the artificial intelligence tool ki in the present invention:

[0072] Figure 2 The figure is a processing relationship diagram of the article outline and the artificial intelligence tool of the present invention. Each paragraph is Sj, and there are multiple sub-paragraphs Pi in S1, which are sent to different artificial intelligence tools respectively. The different artificial intelligence tools in the present invention are collectively referred to as AIGPT. AIGPT-i includes ChatGPT, Siri, or Wenxinyiyan, which are source servers with artificial intelligence processing characteristics and services. After the AIGPT-i is processed, each returns the processing time ti (t1<>t2<>…<>ti), and the time returned is different from each other. This is the first stage. The processing process of the first stage is one-to-one, and the AIGPT service source is randomly matched.

[0073] The second stage of AI processing is performed on the Sj chapters, and the corresponding AIGPT-k is matched according to the average cumulative number of the shortest values, as well as the recommended execution order. Figure 3 This is a relationship diagram of the article paragraphs connected to the artificial intelligence tools of the present invention. Each S can correspond to multiple AIGPTs, but each S can only be assigned one AIGPT.

[0074] Figure 4 This is an example of an article. The first stage generates the initial text according to the article outline, and the second stage improves the coherence and optimization of the initial text. The steps are as follows:

[0075] The first paragraph S1 has three sub-paragraphs P1, P2, and P3. The second paragraph S2 has three sub-paragraphs P1, P2, and P3. The third paragraph S3 has two sub-paragraphs P1 and P2.

[0076] S1: P1+P2+P3 can initially be matched to AIGPT-1, AIGPT-2, or AIGPT-3, each generating three running times of t1+t2+t3.

[0077] S2: P1+P2+P3 can initially be matched to AIGPT-1, AIGPT-2, or AIGPT-3, each generating three running times of t1+t2+t3.

[0078] S3: P1+P2 can initially be matched to AIGPT-1, AIGPT-2, or AIGPT-3, each generating two running times such as t1+t2.

[0079] Shortest value first sequence sorting mechanism:

[0080] Step 11. Input chapters S1, S2, S3…Sj, i.e. P1, P2, P3,…Pi contained in each chapter;

[0081] Step 12. The user sets the range of processing time t, short, normal, or long;

[0082] Step 13. P1, P2, P3, ...Pi sequences are randomly matched to AIGPT-1, AIGPT-2, ..., AIGPT-k;

[0083] Step 14. AIGPT-1, AIGPT-2, ..., AIGPT-k return the processed text content and the processing time t1, t2, t3, ...;

[0084] Step 15. Classify the time t1, t2, t3, ... as long, short, or average;

[0085] Step 16. The cumulative number ratio of average running time of the artificial intelligence tools AIGPT-1, AIGPT-2, ... is evaluated;

[0086] Step 17. Optimize the ranking of artificial intelligence tools;

[0087] The AI ​​tool time algorithm includes the following formulas and steps:

[0088]

[0089] Step 21: Input P1, P2, P3, …, Pi;

[0090] Step 22: Output t1, t2, …, ti;

[0091] Step 23: For i=0→n;

[0092] Step 24: Execute formula (1) to obtain the processing time ti of the Pi segment;

[0093] Step 25: Take out the minimum value min{t1, t2, …, ti};

[0094] Step 26: If the minimum values ​​are the same, select the first one;

[0095] Step 27: Execute the shortest value first sequence sorting mechanism to obtain the optimal sorting;

[0096] Step 28: For j = 0 → m;

[0097] Step 29:

[0098] T j =Execute formula (4) to obtain the total execution time of the jth

[0099] Send to the corresponding AIGPT in the order of obtaining the best sorting

[0100] Cumulative running time of each S chapter

[0101] 13};

[0102] After the first stage is completed, each segment can be judged according to the running time, and given one of three options: short, normal, or long. The short, normal, or long time range can be set by the user.

[0103] The second stage is the matching mechanism, which will summarize, logically adjust and format each paragraph, as shown in Table 1. So input the artificial intelligence tool again, select the artificial intelligence tool with the shortest average number of values ​​in the first stage, which represents good performance, and sort it from the shortest to the longest, and input it in order. The paragraphs that exceed the limit will be re-arranged; Table 2 shows the optimal sorting of artificial intelligence tools from large to small, AIGPT-1→AIGPT-3→AIGPT-2; Table 3 is the symbol table of artificial intelligence tool time algorithm.

[0104] Table 1. Best ranking of AI tools

[0105]

[0106] Table 2. Ranking of paragraph matching AI tools

[0107] Serial number Paragraph number Artificial Intelligence Tools 1 S1 AIGPT-1 2 S2 AIGPT-3 3 S3 AIGPT-2 4 S4 AIGPT-1 5 S5 AIGPT-3

[0108] Table 3. Symbol table

[0109]

[0110] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as ROM / RAM, disk, CD-ROM, etc.

[0111] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

[0112] It should be noted that the embodiments of the present invention can be implemented by hardware, software or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. Those skilled in the art will appreciate that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in processor control codes, such as such codes provided on a programmable memory or a data carrier such as an optical or electronic signal carrier.

[0113] In addition, although the operation of the method of the present invention is described in a particular order in the accompanying drawings, this does not require or imply that these operations must be performed in this particular order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps described in the flow chart can change the order of execution. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step for execution, and / or one step can be decomposed into multiple steps for execution. It should also be noted that the features and functions of two or more devices according to the present invention can be embodied in one device. On the contrary, the features and functions of a device described above can be further divided into being embodied by multiple devices.

[0114] Although the invention has been described with reference to several specific embodiments, it should be understood that the invention is not limited to the specific embodiments disclosed. The invention is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A rapid article generation method based on multiple artificial intelligence tools, characterized by: The method comprises the following steps: The user selects the target article in the sample database; Based on the target article, the article framework and expression techniques are extracted to construct the target article code; the target article code includes paragraph S j 、Sub-paragraph P i , sentence relation R n and article difficulty; each paragraph S j Each contains multiple sub-paragraphs P i , each subparagraph P i It includes multiple sentences and there is a sentence relationship R between the multiple sentences n ; Artificial Intelligence Tools i According to different sub-paragraphs P i The content input by the user is expanded to form key points according to the main theme, and a plurality of the key points form an article outline; Each paragraph S j Different sub-paragraphs P i The key points are input into different artificial intelligence tools i , and different paragraphs S j Subparagraphs P with the same symbol i Artificial intelligence tools with the same symbol k i , the AI ​​tool ki will follow the sub-paragraph P i The corresponding sentence relationship and expression method, the sub-paragraph P i Fill in the key points and get the initial text, and return the running time to the processor after filling; The processor is based on the running time T i Determine the AI ​​toolk i The effectiveness ranking of the artificial intelligence tool k i In order of effectiveness from high to low, paragraph S of the initial text j One-to-one contextual coherence and optimization.

2. The method for rapid article generation based on multiple artificial intelligence tools according to claim 1, characterized in that: The processor calculates the artificial intelligence tool k according to the shortest value priority sequence sorting mechanism. i The cumulative average running time ratio is used to determine the AI ​​tool k. i performance ranking.

3. The method for rapid article generation based on multiple artificial intelligence tools according to claim 2, characterized in that: The shortest value first sequence sorting mechanism comprises the following steps: S1. Setting a time range of the processing time T, wherein the time range is divided into short, medium and long; S2, each sub-paragraph P i Randomly match AI tools k i , and the artificial intelligence tool k i Returns the processed text content and running time T i ; S3. Each AI tool k i Returns the running time T i Compare with the set time range and get the running time T i The time range of S4. Calculate the running time T i The time range is short and the number of AI tools k i Returns the running time T i The ratio of the total number of AI tools is determined based on the ratio of the number of AI tools. i performance ranking.

4. The method for rapid article generation based on multiple artificial intelligence tools according to claim 1, characterized in that: There are two ways for the user to select a target article in the sample database: first, obtain keywords based on the article content and article type input by the user, and search for the target article in the sample database based on the keywords; Secondly, the user selects a target article from the sample database, and optimizes the article content by analogical reasoning between the abstract of the target article and the article content input by the user.

5. The method for rapid article generation based on multiple artificial intelligence tools according to claim 4 is characterized in that: The types of articles include practical style articles and literary creation articles; When the article type is a practical article, the article content input by the user will be expanded according to the first expansion method to obtain multiple main words, and the multiple main words are in an OR relationship with each other, and these main words and the article type are in an AND relationship, so as to form keywords; When the article type is a literary creation article, the article content input by the user will be expanded according to the second expansion method to obtain multiple main words, and the multiple main words will be in an or relationship with each other, and these main words will be in an and relationship with the article type, so as to form keywords.

6. The method for rapid article generation based on multiple artificial intelligence tools according to claim 5 is characterized in that: The first expansion method is to expand the content of the article input by the user into content similar to the subject of the target article by analogical reasoning; The second expansion method is to expand the article content input by the user into related words related to the target article through a big data model.

7. The method for rapid article generation based on multiple artificial intelligence tools according to claim 1, characterized in that: It includes a text analysis tool. The target article coding is obtained by classifying the contents with similar themes, concepts or functions in the target article into nodes of the same category through the text analysis tool, and analyzing the relationship between multiple nodes to obtain the article framework and expression method.

8. The method for rapid article generation based on multiple artificial intelligence tools according to claim 1, characterized in that: The sample database classifies articles according to article categories, and the articles in each category are sorted according to the difficulty of the articles.

9. The method for rapid article generation based on multiple artificial intelligence tools according to claim 8, characterized in that: The artificial intelligence tool i According to the difficulty of the target article, the input article content is expanded and filled with the corresponding difficulty.

Images