Composition correction method and device, mobile equipment and storage medium

Abstract

The embodiment of the invention provides a composition correction method and device, mobile equipment and a storage medium. According to the method, firstly, feature analysis is performed on a target composition written by a target object at this time to obtain writing features at this time, then the writing features at this time are combined with historical writing levels and historical writing features of the target object to generate a first reference model essay, and a second reference model essay is generated according to the target composition; the first reference model essay is a model essay higher than the historical writing level of the target object by one level, the second reference model essay is a full-score model essay in the standard meaning, and according to the score of the target composition, determining a model essay with the score higher than that of the target composition and similar to that of the target composition from the first reference model essay and the second reference model essay as the target reference model essay, according to the target reference model essay, the target composition is corrected, and the target reference model essay better fits the writing level of the target object, so that the writing level of the target object can be improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of artificial intelligence technology. More specifically, embodiments of the present invention relate to a composition marking method, apparatus, mobile device and storage medium. Background Art

[0002] This section is intended to provide a background or context to embodiments of the invention that are recited in the claims. No description herein is admitted to be prior art by inclusion in this section.

[0003] Artificial Intelligence (AI) essay correction is a good supplement to current essay learning. On the one hand, teachers cannot correct essays for users in detail and one by one; on the other hand, parents lack both time and knowledge level for professional tutoring. Therefore, the emergence of AI correction fills this gap well.

[0004] The overall process of the current AI essay grading solution is: the user takes a photo and uploads it to submit the essay, and AI generates the grading results (including original text error correction, full-text correction, sentence-by-sentence comments, and display of excellent sample essays).

[0005] In the above-mentioned AI composition grading solution, AI supports grading users' compositions according to grade and different standards. Although the correction annotations have been refined, they have not yet truly adapted to the user's own level. Since the writing levels of different users vary greatly, relatively fixed grading standards cannot accurately help children solve composition problems. Summary of the invention

[0006] Therefore, an improved composition marking method is greatly needed to solve the technical problem in the prior art of how to mark the user's composition in a personalized manner in combination with the user's current writing level.

[0007] In this context, embodiments of the present invention are intended to provide a composition marking method, apparatus, mobile device, and storage medium.

[0008] In a first aspect of the embodiments of the present invention, a composition marking method is provided, the method comprising:

[0009] Performing feature analysis on the target composition of the target object this time to obtain the writing features of the target object this time;

[0010] Generate a first reference model essay according to the writing characteristics of this time, the historical writing level of the target object and the historical writing characteristics of the target object, and generate a second reference model essay according to the target essay; wherein the score of the first reference model essay is less than the score of the second reference model essay;

[0011] Determine a target reference essay having a score close to that of the target essay from the first reference essay and the second reference essay; wherein the score of the target reference essay is greater than the score of the target essay;

[0012] The target composition is corrected based on the target reference model.

[0013] In one embodiment of the present invention, before performing feature analysis on the target composition of the target object to obtain the writing features of the target object, the method further includes:

[0014] Based on the historical essays of the target subjects, the general large model is trained and optimized to obtain a level assessment model for assessing writing level;

[0015] The historical writing level and historical writing characteristics of the target object are obtained based on the level assessment model.

[0016] In another embodiment of the present invention, feature analysis is performed on the target composition of the target object to obtain the writing features of the target object, including:

[0017] Acquire a photographed image of the target composition of the target object;

[0018] Performing image analysis on the captured image to obtain composition text information;

[0019] Based on the composition text information, feature analysis is performed to obtain the current writing features of the target object.

[0020] In yet another embodiment of the present invention, determining a target reference essay having a score close to that of the target composition from the first reference essay and the second reference essay comprises:

[0021] Obtaining the score of the first reference essay and the score of the target essay;

[0022] If the score of the target essay is less than the score of the first reference essay, determining the first reference essay as the target reference essay;

[0023] If the score of the target composition is greater than or equal to the score of the first reference essay, then when the score of the target composition is less than or equal to the score of the second reference essay, the second reference essay is determined to be the target reference essay.

[0024] In yet another embodiment of the present invention, when the score of the target composition is less than the score of the second reference model composition, correcting the target composition based on the target reference model composition includes:

[0025] Based on the target reference model essay, the target essay is corrected and a first batch of corrected contents is generated; wherein the first batch of corrected contents is used for the target object to optimize the target essay;

[0026] When the score of the target composition is equal to the score of the second reference model composition, correcting the target composition based on the target reference model composition includes:

[0027] Based on the target reference model essay, second correction content is generated for the target content in the target essay; wherein the second correction content is used to praise the target content.

[0028] In still another embodiment of the present invention, the writing feature includes at least one of a thought feature, an article structure feature, a writing technique feature, and a language usage feature.

[0029] In yet another embodiment of the present invention, the method further comprises:

[0030] Re-evaluate the level of historical writing based on the score of the target essay;

[0031] The historical writing features are updated according to the current writing features.

[0032] In a second aspect of the embodiments of the present invention, a composition correction device is provided, characterized in that the device comprises:

[0033] A feature analysis module is used to perform feature analysis on the target composition of the target subject to obtain the writing features of the target subject;

[0034] A model essay generating module, for determining a first reference model essay according to the writing characteristics, the historical writing level of the target object and the historical writing characteristics of the target object, and determining a second reference model essay according to the target composition; wherein the score of the first reference model essay is less than the score of the second reference model essay;

[0035] A model essay determination module, configured to determine a target reference model essay having a score close to that of the target essay from the first reference model essay and the second reference model essay; wherein the score of the target reference model essay is greater than the score of the target essay;

[0036] A correction module is used to correct the target composition based on the target reference model essay.

[0037] In a third aspect of the embodiments of the present application, a mobile device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0038] Memory, used to store computer programs;

[0039] The processor is used to implement the steps of the composition marking method described in any embodiment of the first aspect when executing the program stored in the memory.

[0040] In a fourth aspect of the implementation of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the composition marking method as described in any embodiment of the first aspect are implemented.

[0041] According to the composition marking method implemented in the embodiment of the present application, a feature analysis is performed on the target composition of the target object to obtain the writing features of the target object; a first reference model essay is generated according to the writing features of the target object, the historical writing level of the target object and the historical writing features of the target object, and a second reference model essay is generated according to the target composition; wherein the score of the first reference model essay is less than the score of the second reference model essay; a target reference model essay with a score close to that of the target composition is determined from the first reference model essay and the second reference model essay; wherein the score of the target reference model essay is greater than the score of the target composition; and the target composition is marked based on the target reference model essay. The method can firstly perform a feature analysis on the target composition written by the target subject this time to obtain the features of this writing, and then combine the features of this writing with the historical writing level and historical writing features of the target subject to generate a first reference model essay, and generate a second reference model essay based on the target composition. The first reference model essay is a model essay with a higher historical writing level than the target subject, and the second reference model essay is a full-mark model essay in the standard sense. According to the score of the target composition, a model essay with a higher score than the target composition and a similar score is determined from the first reference model essay and the second reference model essay as the target reference model essay, and the target composition is corrected according to the target reference model essay. Since the target reference model essay is more in line with the writing level of the target subject, it helps to improve the writing level of the target subject. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, in which:

[0043] Figure 1 A system architecture diagram of a composition correction method according to an embodiment of the present application is schematically shown;

[0044] Figure 2 The following schematically shows a flow chart of a composition correction method according to an embodiment of the present application;

[0045] Figure 3 The flowchart of the composition correction method according to another embodiment of the present application is schematically shown;

[0046] Figure 4 The following schematically shows a schematic diagram of constructing a level assessment model according to another embodiment of the present application;

[0047] Figure 5 The following schematically shows a schematic diagram of understanding the writing characteristics according to another embodiment of the present application;

[0048] Figure 6 A schematic diagram of constructing a reference model according to another embodiment of the present application is schematically shown;

[0049] Figure 7 A schematic diagram of constructing a reference model according to another embodiment of the present application is schematically shown;

[0050] Figure 8 A schematic diagram of determining a model essay for correction according to another embodiment of the present application is schematically shown;

[0051] Fig. 9 The schematic diagram of essay correction according to another embodiment of the present application is schematically shown;

[0052] Fig.10 The structure diagram of a composition correction device according to an embodiment of the present application is schematically shown;

[0053] Fig.11 The structure diagram of a mobile device according to an embodiment of the present application is schematically shown;

[0054] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0055] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided only to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. On the contrary, these embodiments are provided to make the present disclosure more thorough and complete, and to fully convey the scope of the present disclosure to those skilled in the art.

[0056] According to the implementation manner of the present application, a composition marking method, apparatus, mobile device and storage medium are proposed.

[0057] It should be understood herein that any number of elements in the drawings is for illustration rather than limitation, and any naming is only for distinction rather than having any limiting meaning.

[0058] The principles and spirit of the present application are explained in detail below with reference to several representative implementations of the present application. SUMMARY OF THE INVENTION

[0060] As mentioned earlier, in the above-mentioned AI essay grading solution, AI supports grading users' essays by grade and different standards. Although the correction annotations have been refined, they have not yet truly adapted to the user's own level. Since the writing levels of different users vary greatly, relatively fixed correction standards cannot be personalized and combined with the user's current writing level to correct the user's essay.

[0061] In view of the problems existing in the above-mentioned prior art, the present application provides a composition correction method, device, mobile device and storage medium. The method can first analyze the characteristics of the target composition written by the target object this time to obtain the writing characteristics of this time, and then combine the writing characteristics of this time with the historical writing level and historical writing characteristics of the target object to generate a first reference model essay, and generate a second reference model essay based on the target composition. The first reference model essay is a model essay with a higher historical writing level than the target object, and the second reference model essay is a full-score model essay in the standard sense. According to the score of the target composition, a model essay with a higher score than the target composition and a similar score is determined from the first reference model essay and the second reference model essay as the target reference model essay, and the target composition is corrected according to the target reference model essay. Since the target reference model essay is more in line with the writing level of the target object, it helps to improve the writing level of the target object.

[0062] After introducing the basic principles of the present invention, various non-limiting embodiments of the present invention are described in detail below.

[0063] Application Scenario Overview

[0064] First reference Figure 1 , Figure 1 The system architecture diagram of a composition marking method is shown in FIG. 1 . The system architecture includes at least a data acquisition module 101 and a data processing module 102. The data acquisition module 101 and the data processing module 102 establish a communication connection. Specifically, the system architecture can be an intelligent terminal, such as a tablet computer, a personal laptop computer, a learning machine, an intelligent mobile terminal (such as a mobile phone), etc.

[0065] The method can be applied to the data processing module 102 in the system architecture, wherein the data acquisition module 101 can be an image acquisition module in the system architecture, and of course can also be an image or document recording a composition obtained from other terminals, without limitation.

[0066] Exemplary Methods

[0067] Combine the following Figure 1For application scenarios, refer to Figure 2 The composition correction method according to the exemplary embodiment of the present invention is described. It should be noted that the above application scenarios are only shown to facilitate understanding of the spirit and principle of the present invention, and the embodiments of the present invention are not limited in this respect. On the contrary, the embodiments of the present invention can be applied to any applicable scenario.

[0068] Next, the scheme is described in detail. The embodiment of the present application discloses a composition correction method, such as Figure 2 , methods include:

[0069] Step 201 , performing feature analysis on the target composition of the target object this time to obtain the writing features of the target object this time.

[0070] The target object refers to the writer of the target composition, such as students. The target composition refers to the composition to be corrected using the composition correction method.

[0071] Step 202, generating a first reference model essay based on the current writing characteristics, the historical writing level of the target object and the historical writing characteristics of the target object, and generating a second reference model essay based on the target composition; wherein the score of the first reference model essay is less than the score of the second reference model essay.

[0072] The first reference essay can be an essay with a higher level of historical writing than the target subject, and the second reference essay can be a full-mark essay in the standard sense. For example, the writing level is divided into different levels such as excellent, good, average, and poor. If the historical writing level of the target subject is average, the first reference essay can be an essay with a good grade, and the second reference essay can be an essay with an excellent grade. For another example, taking the full score of writing level as 100 points, if the historical writing level of the target subject is 70 points, the first reference essay can be an essay with 80 points, and the second reference essay can be a full-mark essay with 100 points. It should be noted that the historical writing level of the target subject does not represent the level of the target essay this time. For example, if the historical writing level of the target subject is 70 points, the target essay this time may exceed the level and reach 88 points, or it may not reach the level and only have 60 points, without restriction.

[0073] Step 203, determining a target reference essay having a score close to that of the target composition from the first reference essay and the second reference essay; wherein the score of the target reference essay is greater than the score of the target composition.

[0074] In one embodiment, a target reference essay having a score close to that of a target composition is determined from a first reference essay and a second reference essay, including: obtaining the score of the first reference essay and the score of the target composition; if the score of the target composition is less than the score of the first reference essay, determining the first reference essay as the target reference essay; if the score of the target composition is greater than or equal to the score of the first reference essay, then determining the second reference essay as the target reference essay if the score of the target composition is less than or equal to the score of the second reference essay.

[0075] In this embodiment, the selection criteria of the target reference model essay is to determine a model essay with a higher score than the target composition and a score close to the target composition from the first reference model essay and the second reference model essay. When selecting the model essay, the size relationship between the score of the first reference model essay and the score of the target composition can be first determined. If the score of the target composition is less than the score of the first reference model essay, the first reference model essay can be determined as the target reference model essay. If the score of the target composition is greater than or equal to the score of the first reference model essay, the size relationship between the score of the target composition and the score of the second reference model essay can be further determined. If the score of the target composition is less than or equal to the score of the second reference model essay, the second reference model essay is used as the target reference model essay. It should be noted that since the second reference model essay is a full score model essay in the standard sense, the step of determining the score of the target composition and the score of the second reference model essay can be omitted, that is, when the score of the target composition is greater than or equal to the score of the first reference model essay, the second reference model essay can be directly determined as the target reference model essay. An example is given below. For example, if the historical writing level of the target object mentioned above is 70 points, the first reference essay can be an essay with 80 points, and the second reference essay can be a full-mark essay with 100 points. If the score of the target essay this time reaches 88 points, since the score of the first reference essay is lower than that of the target essay, the second reference essay is determined as the target reference essay. If the score of the target essay this time is 70 points, since both the first reference essay and the second reference essay are higher than the score of the target essay this time, but the score of the first reference essay is closer to the score of the target essay, the first reference essay is determined as the target reference essay.

[0076] Step 204, correcting the target composition based on the target reference model essay.

[0077] In one embodiment, when the score of the target composition is less than the score of the second reference model essay, the target composition is corrected based on the target reference model essay, including: correcting the target composition based on the target reference model essay and generating a first batch of corrected contents; wherein the first batch of corrected contents are used to optimize the target composition for the target object.

[0078] When the score of the target composition is equal to the score of the second reference model essay, the target composition is corrected based on the target reference model essay, including: generating second correction content for the target content in the target composition based on the target reference model essay; wherein the second correction content is used to praise the target content.

[0079] In this embodiment, when correcting, if the score of the target composition is less than the score of the second reference model, the target reference model can be used as a benchmark to correct the target composition and generate the first batch of corrections, such as polishing the general sentences in the target composition. The target object can understand the shortcomings and room for improvement in this target composition based on the first batch of corrections, which is convenient for improving the writing level of the target object. In the case where the score of the target composition is equal to the score of the second reference model, that is, the target composition is also a full-mark composition under the same standard, then the second batch of corrections can be generated for the beautiful paragraphs and sentences in the target composition. The second batch of corrections can be content that encourages and praises the corresponding paragraphs or sentences, so that the target object can obtain positive feedback after writing a full-mark composition, and motivate the target object to further consolidate and improve the current writing level.

[0080] The method can firstly perform a feature analysis on the target composition written by the target subject this time to obtain the features of this writing, and then combine the features of this writing with the historical writing level and historical writing features of the target subject to generate a first reference model essay, and generate a second reference model essay based on the target composition. The first reference model essay is a model essay with a higher historical writing level than the target subject, and the second reference model essay is a full-mark model essay in the standard sense. According to the score of the target composition, a model essay with a higher score than the target composition and a similar score is determined from the first reference model essay and the second reference model essay as the target reference model essay, and the target composition is corrected according to the target reference model essay. Since the target reference model essay is more in line with the writing level of the target subject, it helps to improve the writing level of the target subject.

[0081] In one embodiment, before performing feature analysis on the target composition of the target object and obtaining the writing features of the target object, the method also includes: training and optimizing a general large model based on the historical composition of the target object to obtain a level assessment model for assessing writing level; and obtaining the historical writing level and historical writing features of the target object based on the level assessment model.

[0082] In this embodiment, firstly, the general large model is trained and optimized based on the historical composition of the target object, so that the general large model is more suitable for evaluating the composition written by the target object. The model after training and optimization can be called a level evaluation model or a user composition level evaluation model, and the historical writing level and historical writing features of the target object can be extracted from the historical composition of the target object based on the level evaluation model. Of course, the current writing level and the current writing features can also be extracted from the target composition of the target object based on the level evaluation model. Writing features (including historical writing features and current writing features) can be divided into multiple dimensions, such as ideas features, article structure features, writing techniques features and language usage features. Among them, article structure features refer to what structure users generally like to use in writing, such as total-division-total, etc. Writing technique features refer to what descriptive techniques users are generally good at using in writing, such as metaphors, personification, etc. Language usage features refer to what adjectives users generally like to use in writing, and what are the high-frequency words. Idea features are the ideas expansion of which aspects of the composition topic, the main line of writing, etc., without restriction.

[0083] In one embodiment, the method further includes: re-evaluating the level of historical writing based on the score of the target composition; and updating the historical writing features according to the current writing features.

[0084] In this embodiment, after obtaining the score of the target composition, the historical writing level of the target object can be re-evaluated based on the level assessment model, and the historical writing features in the level assessment model can be updated according to the current writing features of the target composition. In other words, the level assessment model can be updated in a timely manner with each uploaded composition, and the user composition collection and prompt words (prompt) can be input into the level assessment model to store the user's latest writing level grade, writing features, etc.

[0085] In one embodiment, a feature analysis is performed on the target composition of the target object to obtain the writing features of the target object, including: obtaining a captured image of the target composition of the target object; performing image analysis on the captured image to obtain text information of the composition; and performing feature analysis based on the text information of the composition to obtain the writing features of the target object.

[0086] In this embodiment, the target composition uploaded by the target object is generally a captured image, and the composition text information can be obtained through image analysis. For example, optical character recognition technology (OCR) can be used to analyze the image to obtain the composition text information, and feature analysis can be performed based on the composition text information to obtain the writing features of the target object.

[0087] In the above embodiments, based on the analysis of the user's personal writing ability, AI is used to evaluate the user's historical writing level, and then combined with the level of the essay submitted this time, the most suitable personalized correction and tutoring suggestions for the user are given, thereby improving the effect of essay correction.

[0088] In a specific embodiment, the composition correction method is as follows: Figure 3 ,include:

[0089] Step S1, constructing a level assessment model.

[0090] This process is a pre-process, which is a process that is independent of the user's operation. That is, this process does not occur after the user submits the composition for correction, but is an independent process that is carried out in advance. The schematic diagram of the level assessment model is as follows: Figure 4 The platform will store and analyze each corrected essay uploaded by the user, and eventually come up with a level assessment model (i.e., user essay level assessment model). The model content can include at least two aspects: the user's writing level and writing characteristics. The user's writing level can be divided into three levels: excellent, good, and poor; writing characteristics can be divided into three dimensions: article structure characteristics, writing technique characteristics, and language usage characteristics. Article structure characteristics refer to what structure users generally like to use in writing, such as general-specific-general, etc. Writing technique characteristics refer to what descriptive techniques users are generally good at using in writing, such as metaphors, personification, etc. Language usage characteristics refer to which adjectives users generally like to use in writing and which words are used frequently.

[0091] This process is updated regularly, and is updated with a piece of composition uploaded by the user as input. The user composition collection is input into the "User Composition Level Evaluation Model" together with the prompt (prompt word), and the user's latest writing level grade, writing characteristics, etc. are obtained and stored. The "User Composition Level Evaluation Model" can evaluate the user's writing level according to the writing level scoring principle provided by the teaching and research (provided in the form of prompts), and summarize the user's writing characteristics according to its own cognition. The "user writing level grade" and "user writing characteristics" output here will participate in the construction of reference model essays in subsequent processes. The user composition level evaluation model can be a general large model, which can include user composition feature summary model, reference article writing model, article scoring model and composition correction model, etc. according to its specific functions, without restriction.

[0092] Step S2: Understanding the characteristics of the user's writing.

[0093] When users take a photo and upload their own composition, it goes through a series of analyses (OCR images are analyzed into text) and then input into the "User Composition Feature Summary Model" in the form of text. Through prompt debugging, this model will analyze the user's composition this time. The user's thought characteristics, article structure characteristics, writing technique characteristics, language usage characteristics, etc. are analyzed respectively. Thought characteristics refer to the core context of the user's writing this time, for example, what ideas are expressed by describing what events, experiences, and scenery, etc. The schematic diagram of understanding the characteristics of this writing is as follows Figure 5 Through this process, the core purpose is to let AI understand how the user wrote the essay this time, so as to facilitate the construction of subsequent reference essays, that is, how to construct a better essay based on the user’s current writing ideas.

[0094] Step S3, reference sample article construction.

[0095] The construction of reference essays is to prepare for the subsequent process of essay correction. The reference essays are divided into two parts, namely Reference Essay 1 (i.e. the first reference essay) and Reference Essay 2 (i.e. the second reference essay). Reference Essay 1 refers to the target essay given by combining the writing characteristics of this user, the writing level of historical users, and the writing characteristics of historical users. This essay is not the best essay with 100 points, but it is a essay that is closer to the actual situation of the user.

[0096] In the generation process of reference essay 1, the reference essay construction diagram is as follows Figure 6 The user's current writing features output in step S2, the user's writing level output in step S1, and the user's writing features are combined with the prompt and input into the "reference article writing model". The goal is to output a model article that is one level higher than the user's previous writing level.

[0097] In the generation process of Reference Sample 2, the reference sample construction diagram is as follows Figure 7 Reference Essay 2 refers to the essay that the model considers to be the best and most standard essay based on the essay written by the user this time. It can be understood that Reference Essay 2 is a standard essay, that is, a 100-point essay.

[0098] Step S4, determining the sample paper to be corrected.

[0099] Determine the diagram of the revised essay as follows Figure 8, after the generation of reference essays 1 and 2, it is necessary to determine the only corrected essay. This is also the key step in realizing personalized essay correction in this essay correction method. Through analysis and evaluation, it is determined whether the user uses excellent essays close to his or her level as the benchmark for correction, or uses a 100% perfect standard essay as the benchmark for correction. First, the user's essay this time, together with reference essay 1, is input into the "article scoring model" in combination with the prompt (prompt word). The model scores according to the essay scoring standard corresponding to the user's grade (the model can be input through the prompt). Compare the user's writing score this time with the score of the reference essay simulated by AI that is adapted to the user's historical level. If the user's writing score this time is higher, it means that the user's writing level this time is higher than the historical target level, and reference essay 2 is selected as the correction essay. If the user's writing score this time is lower, it means that the user's writing level has not reached the historical target level, and reference essay 1 is selected as the correction essay.

[0100] Step S5, making corrections.

[0101] Compared with the conventional AI essay correction process, the process of inputting essays, matching correction standards and then generating correction examples is done. This solution does the opposite. In the aforementioned process, the target reference example is first confirmed, and the user's current writing is corrected based on the reference example to generate correction content. Not only is the correction content more suitable for the user's actual situation, but it can also help users learn better from the target example. Fig. 9 The specific process is to input the composition uploaded by the user into the "composition correction model". The model generates correction content based on the learning and understanding of the reference essay.

[0102] For example, a user uploads a filmed composition. The user's composition goes through the OCR process and is input into the "User Composition Feature Summary Model" in text format to output the features of this writing, including: ideas, article structure, writing techniques, and language usage. The above results are transferred to the "Reference Model Generation Model", which outputs Reference Model 1 and Reference Model 2 based on the user's historical writing level and historical writing features (article structure, writing techniques, and language usage). Reference Model 1 and the user's composition are input into the "Article Scoring Model" to compare the scores of Reference Model 1 and the user's composition, and further confirm the reference model used for this correction. The user's composition is transferred to the "Composition Correction Model", and the model corrects the user's composition based on its understanding of the reference model and generates correction content.

[0103] Compared with traditional composition correction, AI composition correction has the characteristics of effectiveness and low cost. Compared with manual composition correction, AI correction can be initiated by users anytime and anywhere without being restricted by time and place. At the same time, compared with the high cost of manual comments, the cost of AI correction is significantly lower. In the above-mentioned embodiments of the present application, the advantages of personalization and self-adaptation are added to AI composition correction, so that the correction results are more high-quality and more authoritative, and the correction opinions put forward can fit the user's current composition level and meet the user's actual situation. At the same time, with the dynamic changes in the user's writing level, the correction results can also be adapted accordingly, which helps to improve the user's writing level.

[0104] Exemplary Devices

[0105] After introducing the method of the exemplary embodiment of the present invention, next, refer to Fig.10 Based on the same inventive concept as the above-mentioned composition correction method embodiment, the present application embodiment also provides a composition correction device, including:

[0106] The feature analysis module 1001 is used to perform feature analysis on the target composition of the target object this time to obtain the writing features of the target object this time;

[0107] The model essay generating module 1002 is used to determine a first reference model essay according to the writing characteristics, the historical writing level of the target object and the historical writing characteristics of the target object, and to determine a second reference model essay according to the target composition; wherein the score of the first reference model essay is less than the score of the second reference model essay;

[0108] The model essay determination module 1003 is used to determine a target reference model essay with a score close to that of the target essay from the first reference model essay and the second reference model essay; wherein the score of the target reference model essay is greater than the score of the target essay;

[0109] The correction module 1004 is used to correct the target composition based on the target reference model essay.

[0110] The device can first perform feature analysis on the target composition written by the target object this time to obtain the features of this writing, and then combine the features of this writing with the historical writing level and historical writing features of the target object to generate a first reference model essay, and generate a second reference model essay based on the target composition. The first reference model essay is a model essay with a higher historical writing level than the target object, and the second reference model essay is a full-mark model essay in the standard sense. According to the score of the target composition, a model essay with a higher score than the target composition and a similar score is determined from the first reference model essay and the second reference model essay as the target reference model essay, and the target composition is corrected according to the target reference model essay. Since the target reference model essay is more in line with the writing level of the target object, it helps to improve the writing level of the target object.

[0111] like Fig.11 As shown, the embodiment of the present application provides a mobile device, which can be a smart mobile terminal such as a tablet computer, a learning machine, a mobile phone, or other smart terminals, without limitation, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114.

[0112] Memory 113, used for storing computer programs;

[0113] In one embodiment, the processor 111 is used to execute the program stored in the memory 113 to implement the composition correction method provided by any of the above method embodiments, including:

[0114] Performing feature analysis on the target composition of the target object this time to obtain the writing features of the target object this time;

[0115] Generate a first reference model essay according to the writing characteristics of this time, the historical writing level of the target object and the historical writing characteristics of the target object, and generate a second reference model essay according to the target essay; wherein the score of the first reference model essay is less than the score of the second reference model essay;

[0116] Determine a target reference essay having a score close to that of the target essay from the first reference essay and the second reference essay; wherein the score of the target reference essay is greater than the score of the target essay;

[0117] The target composition is corrected based on the target reference model.

[0118] The communication bus mentioned in the above terminal can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0119] The communication interface is used for communication between the above terminal and other devices.

[0120] The memory may include a random access memory (RAM) or a non-volatile memory, such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0121] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0122] It should be noted that although several devices or sub-devices are mentioned in the above detailed description, this division is not mandatory. In fact, according to an embodiment of the present invention, the features and functions of two or more devices described above can be embodied in one device. Conversely, the features and functions of one device described above can be further divided into multiple devices to be embodied.

[0123] In addition, although the operations of the method of the present invention are described in a specific order in the accompanying drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. On the contrary, the steps depicted in the flow chart can be performed in a different order. Additionally or alternatively, some steps can be omitted, multiple steps can be combined into one step, and / or one step can be decomposed into multiple steps.

[0124] It should be understood that each box in the flowchart and / or block diagram and the combination of boxes in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine, and these computer program instructions are executed by a computer or other programmable data processing device to produce a device that implements the functions / operations specified in the boxes in the flowchart and / or block diagram.

[0125] These computer program instructions may also be stored in a computer-readable medium that enables a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable medium produce a product that includes an instruction device that implements the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0126] Computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operational steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby enabling the instructions executed on the computer or other programmable device to provide a process for implementing the functions / operations specified in the blocks in the flowchart and / or block diagram.

[0127] The use of the verbs "comprise", "include" and their conjugations mentioned in the application documents does not exclude the presence of elements or steps other than those recorded in the application documents. The article "a" or "an" before an element does not exclude the presence of a plurality of such elements.

[0128] Although the spirit and principle of the present invention have been described with reference to several specific embodiments, it should be understood that the present invention is not limited to the disclosed specific embodiments, and the division of various aspects does not mean that the features in these aspects cannot be combined to benefit, and this division is only for the convenience of expression. The present invention is intended to cover various modifications and equivalent arrangements included in the spirit and scope of the attached claims. The scope of the attached claims conforms to the broadest interpretation, thereby including all such modifications and equivalent structures and functions.

Claims

1. A composition correction method, characterized in that: The method comprises: Performing feature analysis on the target composition of the target object this time to obtain the writing features of the target object this time; Generate a first reference model essay according to the writing characteristics of this time, the historical writing level of the target object and the historical writing characteristics of the target object, and generate a second reference model essay according to the target essay; wherein the score of the first reference model essay is less than the score of the second reference model essay; Determine a target reference essay having a score close to that of the target essay from the first reference essay and the second reference essay; wherein the score of the target reference essay is greater than the score of the target essay; The target composition is corrected based on the target reference model.

2. The method according to claim 1, characterized in that Before analyzing the characteristics of the target composition of the target object and obtaining the writing characteristics of the target object, the method further includes: Based on the historical essays of the target subjects, the general large model is trained and optimized to obtain a level assessment model for assessing writing level; The historical writing level and historical writing characteristics of the target object are obtained based on the level assessment model.

3. The method according to claim 1, characterized in that: The target composition of the target object is analyzed to obtain the writing characteristics of the target object, including: Acquire a photographed image of the target composition of the target object; Performing image analysis on the captured image to obtain composition text information; Based on the composition text information, feature analysis is performed to obtain the current writing features of the target object.

4. The method according to claim 1, characterized in that: Determining a target reference essay having a score close to that of the target essay from the first reference essay and the second reference essay comprises: Obtaining the score of the first reference essay and the score of the target essay; If the score of the target essay is less than the score of the first reference essay, determining the first reference essay as the target reference essay; If the score of the target composition is greater than or equal to the score of the first reference essay, then when the score of the target composition is less than or equal to the score of the second reference essay, the second reference essay is determined to be the target reference essay.

5. The method according to claim 4, characterized in that When the score of the target composition is less than the score of the second reference model composition, correcting the target composition based on the target reference model composition includes: Based on the target reference model essay, the target essay is corrected and a first batch of corrected contents is generated; wherein the first batch of corrected contents is used for the target object to optimize the target essay; When the score of the target composition is equal to the score of the second reference model composition, correcting the target composition based on the target reference model composition includes: Based on the target reference model essay, second correction content is generated for the target content in the target essay; wherein the second correction content is used to praise the target content.

6. The method according to any one of claims 1 to 5, characterized in that: The writing characteristics include at least one of the following: thought characteristics, article structure characteristics, writing technique characteristics and language usage characteristics.

7. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: Re-evaluate the level of historical writing based on the score of the target essay; The historical writing features are updated according to the current writing features.

8. A composition correction device, characterized in that: The device comprises: A feature analysis module is used to perform feature analysis on the target composition of the target subject to obtain the writing features of the target subject; A model essay generating module, for determining a first reference model essay according to the writing characteristics, the historical writing level of the target object and the historical writing characteristics of the target object, and determining a second reference model essay according to the target composition; wherein the score of the first reference model essay is less than the score of the second reference model essay; A model essay determination module, configured to determine a target reference model essay having a score close to that of the target essay from the first reference model essay and the second reference model essay; wherein the score of the target reference model essay is greater than the score of the target essay; A correction module is used to correct the target composition based on the target reference model essay.

9. A mobile device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor is used to implement the steps of the composition correction method described in any one of claims 1 to 7 when executing the program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the composition marking method according to any one of claims 1 to 7 are implemented.

Images