CN118036571A

CN118036571A - Composition correction interaction method and device based on artificial intelligence and electronic equipment

Info

Publication number: CN118036571A
Application number: CN202410176413.9A
Authority: CN
Inventors: 王艺深; 王娜
Original assignee: Netease Youdao Information Technology Beijing Co Ltd
Current assignee: Netease Youdao Information Technology Beijing Co Ltd
Priority date: 2024-02-08
Filing date: 2024-02-08
Publication date: 2024-05-14

Abstract

The embodiment of the invention provides a composition correction interaction method and device based on artificial intelligence and electronic equipment, wherein the method comprises the following steps: identifying text information corresponding to the text to be modified; correcting the text information by using a preset large language model to obtain an initial correction result, wherein the initial correction result comprises a first correction result, a second correction result and/or a third correction result; generating a browsing interface for displaying the first correcting result, the second correcting result and/or the third correcting result in a preset mode; responding to the adjustment operation received based on the browsing interface, and adjusting the first correction result, the second correction result and/or the third correction result; and generating a target display page according to the adjusted correction result, and displaying the target display page. Therefore, the adjustment operation of the user aiming at the browsing interface can be received, and the initial correction result is adjusted, so that the target display page finally displayed to the user is more ideal.

Description

Composition correction interaction method and device based on artificial intelligence and electronic equipment

Technical Field

The embodiment of the invention relates to the technical field of intelligent correction, in particular to a composition correction interaction method and device based on artificial intelligence and electronic equipment.

Background

This section is intended to provide a background or context to the embodiments of the invention that are recited in the claims. The description herein is not admitted to be prior art by inclusion in this section.

In the technical field of intelligent correction, especially in the technical field of composition correction, generally, after a user uploads a composition, a system directly generates correction results of the composition based on a large language model and directly displays the correction results to the user, but the correction results generated by the system cannot be manually adjusted, so that the correction results finally displayed to the user are not ideal. Therefore, how to implement interface interaction in the process of composition correction becomes a technical problem to be solved.

Disclosure of Invention

However, for reasons that manual adjustment of the system-generated correction results is not realized at all during the composition correction, the prior art generally directly displays the system-generated correction results to the user.

In the prior art, therefore, it is easy to cause the modification result finally presented to the user to be less than ideal.

Therefore, an improved composition correction method is highly needed to realize the manual adjustment of the correction results generated by the system, so that the correction results finally displayed to the user are more ideal.

In this context, the embodiment of the invention is expected to provide a composition correction interaction method and device based on artificial intelligence and electronic equipment.

In a first aspect of the embodiment of the present invention, there is provided an artificial intelligence based composition correction interaction method, including:

Identifying text information corresponding to the to-be-modified text, wherein the text information is obtained based on the acquired text content of the to-be-modified text or is obtained based on the acquired picture content identification of the to-be-modified text;

Correcting the text information by using a preset large language model to obtain an initial correction result, wherein the initial correction result comprises a first correction result for correcting characters in the text information, a second correction result for evaluating sentences and paragraphs in the text information and/or a third correction result for evaluating the whole text to be corrected;

Generating a browsing interface for displaying the first correcting result, the second correcting result and/or the third correcting result in a preset mode;

responding to the adjustment operation received based on the browsing interface, and adjusting the first correction result, the second correction result and/or the third correction result;

And generating a target display page according to the adjusted correction result, and displaying the target display page.

In a second aspect of the embodiments of the present invention, there is provided an artificial intelligence based composition alteration interaction apparatus, the apparatus comprising:

The identification module is used for identifying text information corresponding to the to-be-modified text, wherein the text information is obtained based on the acquired text content of the to-be-modified text or is obtained based on the acquired picture content identification of the to-be-modified text;

The correction module is used for correcting the text information by utilizing a preset large language model to obtain an initial correction result, wherein the initial correction result comprises a first correction result for correcting characters in the text information, a second correction result for evaluating sentences and paragraphs in the text information and/or a third correction result for evaluating the whole text to be corrected;

the generation module is used for generating a browsing interface for displaying the first correction result, the second correction result and/or the third correction result in a preset mode;

The adjustment module is used for responding to the adjustment operation received based on the browsing interface and adjusting the first correction result, the second correction result and/or the third correction result;

And the display module is used for generating a target display page according to the adjusted correction result and displaying the target display page.

In a third aspect of the embodiments of the present invention, there is provided an electronic device including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory perform communication with each other through the communication bus;

A memory for storing a computer program;

and the processor is used for realizing the artificial intelligence-based composition correction interaction method according to any one of the first aspect when executing the program stored in the memory.

In a fourth aspect of embodiments of the present invention, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the artificial intelligence based composition altering interaction method of any of the first aspects.

In a fifth aspect of embodiments of the present invention, there is provided a computer program product comprising a computer program stored on a non-volatile computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to perform the artificial intelligence based composition altering interaction method of any of the first aspects.

According to the composition correction interaction method, device and electronic equipment based on artificial intelligence, after the text information of the composition to be corrected is corrected by utilizing the preset large language model, the adjustment operation of a user for a browsing interface is received, and the first correction result, the second correction result and/or the third correction result are adjusted in response to the adjustment operation, so that interface interaction in the composition correction process is realized, and a target display page finally displayed to the user is more ideal.

Drawings

The above, as well as additional purposes, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description when read in conjunction with the accompanying drawings. Several embodiments of the present invention are illustrated by way of example, and not by way of limitation, in the figures of the accompanying drawings and in which:

fig. 1 schematically shows a schematic view of an application scenario for composition correction according to an embodiment of the present invention;

FIG. 2 schematically illustrates a flow chart of an artificial intelligence based composition altering interaction method according to an embodiment of the invention;

FIG. 3 schematically illustrates a schematic view of a browsing interface according to an embodiment of the present invention;

FIG. 4 schematically shows a schematic view of a browsing interface showing a first modification result according to an embodiment of the present invention;

FIG. 5 schematically illustrates an interface diagram of a new character modification result according to an embodiment of the present invention;

FIG. 6 schematically illustrates an interface diagram of a delete character adaptation result according to an embodiment of the invention;

FIG. 7 schematically illustrates an interface diagram for paragraph merging according to an embodiment of the present invention;

FIG. 8 schematically illustrates an interface schematic of paragraph splitting according to an embodiment of the invention;

FIG. 9 schematically illustrates a view of a browsing interface showing sentence scores according to an embodiment of the present invention;

FIG. 10 schematically illustrates a view of a browsing interface showing a section comment according to an embodiment of the present invention;

FIG. 11 schematically illustrates a schematic diagram of a target presentation page according to an embodiment of the present invention;

FIG. 12 schematically illustrates an interface diagram of an upload interface according to an embodiment of the present invention;

FIG. 13 schematically illustrates a functional block diagram of an artificial intelligence based composition altering interaction device according to an embodiment of the invention.

Fig. 14 schematically shows a functional block diagram of an electronic device according to an embodiment of the invention.

In the drawings, the same or corresponding reference numerals indicate the same or corresponding parts.

Detailed Description

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 presented merely to enable those skilled in the art to better understand and practice the invention and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

Summary of The Invention

The inventor finds that the existing composition correction mode is generally that after a user uploads the composition content, the system directly generates correction results of the composition content based on a large language model and directly displays the correction results to the user, so that the correction results generated by the system cannot be manually adjusted, and the correction results finally displayed to the user are not ideal.

Based on the method, after the large language model generates the composition correction result, the interface interaction in the composition correction process is realized by receiving the adjustment operation of the user on the browsing interface with the correction result and responding to the adjustment operation to adjust the correction result.

Having described the basic principles of the present invention, various non-limiting embodiments of the invention are described in detail below.

Application scene overview

First, referring to fig. 1, an application scenario of an artificial intelligence-based composition correction interactive method according to an embodiment of the present invention will be described in detail.

Fig. 1 schematically shows an application scenario of an artificial intelligence based composition altering interaction method implemented by the present invention. It should be noted that fig. 1 is only an example of an application scenario where an embodiment of the present invention may be applied to help those skilled in the art understand the technical content of the present invention, and does not mean that the embodiment of the present invention may not be used in other devices, systems, environments, or scenarios.

As shown in fig. 1, in a composition correction application scenario, a system architecture of an artificial intelligence based composition correction interaction method according to an embodiment of the present invention may include a terminal device 101 and a server 102. The terminal device 101 may be provided with a user operation interface to obtain, in real time, text content or picture content of the to-be-modified text uploaded by the user. After acquiring the text content or the picture content of the text to be modified uploaded by the user, the terminal device 101 can identify text information corresponding to the text to be modified according to the text content or the picture content, and then send the text information corresponding to the text to be modified to the server 102; the server 102 corrects the text information according to a preset large language model to obtain an initial correction result, and sends the initial correction result to the terminal equipment 101; and then the terminal equipment 101 generates a browsing interface according to the initial correction result, adjusts the initial correction result in response to the adjustment operation received based on the browsing interface, generates a target display page, and displays the target display page through the user operation interface.

Optionally, after the terminal device 101 identifies the text information corresponding to the text to be modified according to the text content or the picture content of the text to be modified, the terminal device 101 may modify the text information according to a preset large language model to obtain an initial modification result, then generate a browsing interface according to the initial modification result, and adjust the initial modification result in response to an adjustment operation received based on the browsing interface, generate a target display page, and display the target display page through the user operation interface.

Optionally, after the terminal device 101 identifies the text information corresponding to the text to be modified according to the text content or the picture content of the text to be modified, the terminal device 101 and the server 102 may cooperate to modify the text information according to a preset large language model to obtain an initial modification result, then generate a browsing interface according to the initial modification result, and respond to an adjustment operation received based on the browsing interface to adjust the initial modification result, generate a target display page, and display the target display page through the user operation interface.

Terminal device 101 may be a variety of electronic devices including, but not limited to, a smart phone, a tablet computer, a laptop portable computer, a desktop computer, and the like.

The server 102 may interact with the terminal device 101 through a network to receive or send messages or the like. For example, the server 102 may receive text information corresponding to the to-be-modified text sent by the terminal device 101, or may send an initial modification result generated according to a preset large language model to the terminal device 101.

It should be noted that, the steps in the artificial intelligence-based composition correction interaction method according to the embodiment of the present invention may be performed by the terminal device 101 or the server 102, which is not limited in this embodiment of the present invention.

Exemplary method

An artificial intelligence based composition altering interaction method according to an exemplary embodiment of the present invention is described below with reference to fig. 2 in conjunction with the application scenario of fig. 1. It should be noted that the above application scenario is only shown for the convenience of understanding the spirit and principle of the present invention, and the embodiments of the present invention are not limited in any way. Rather, embodiments of the invention may be applied to any scenario where applicable.

Referring first to FIG. 2, a flow chart of an artificial intelligence based composition altering interaction method in accordance with an embodiment of the present invention is schematically shown. As shown in fig. 2, the artificial intelligence based composition altering interaction method may include:

Step 202, identifying text information corresponding to the to-be-modified text, wherein the text information is obtained based on the text content of the obtained to-be-modified text or is obtained based on the picture content identification of the obtained to-be-modified text.

Specifically, the above-mentioned composition to be modified may refer to a composition to be modified uploaded by a user (such as a student, a teacher, etc.), and specifically may be an article of any language type such as english, chinese, etc.

In the step, the user can upload the picture content of the text to be modified (such as related pictures of the text to be modified) through the user operation interface of the electronic equipment, and can also directly input the text content of the text to be modified through the user operation interface of the electronic equipment, so that the electronic equipment can identify the text information corresponding to the text to be modified by utilizing the text content of the text to be modified or the picture content of the text to be modified, and the text information corresponding to the text to be modified is convenient to process subsequently.

It should be noted that the number of the to-be-modified texts may be one or multiple, and the embodiment of the present invention is not limited specifically. When the number of the to-be-modified texts is multiple, a user can upload the to-be-modified texts through the tabs on the user operation interface at one time, and at the moment, the electronic equipment can identify text information corresponding to the to-be-modified texts in batches, so that the to-be-modified texts can be modified conveniently and simultaneously.

And 204, correcting the text information by using a preset large language model to obtain an initial correction result, wherein the initial correction result comprises a first correction result for correcting characters in the text information, a second correction result for evaluating sentences and paragraphs in the text information and/or a third correction result for evaluating the whole of the text to be corrected.

Specifically, the large language model is trained in advance and can be used for correcting text information to obtain an initial correction result. The large language model may be a generative pre-training transducer (GENERATIVE PRE-Trained Transformer, abbreviated as GPT) model, a deep bi-directional pre-training transducer (Bidirectional Encoder Representation from Transformers, abbreviated as BERT) model based on semantic understanding, a Transformers model, a T5 (Text-to-Text Transfer Transformer) model, etc., which is not limited in this embodiment. The initial correction result may include a combination of one or more of a first correction result for correcting characters in the text information of the to-be-corrected text, a second correction result for evaluating sentences and paragraphs in the text information of the to-be-corrected text, and a third correction result for evaluating the to-be-corrected text as a whole. It should be noted that, the first, second and/or third modification results may be generated at one time by using a large language model, that is, text information of the text to be modified is input into the large language model, and the large language model simultaneously generates the first, second and/or third modification results; the method can also be generated for multiple times according to the process steps, namely when the characters in the text information of the text to be corrected are required to be corrected, the text information of the text to be corrected is input into a large language model, a first correction result is generated by the large language model at the same time, when sentences and paragraphs in the text information of the text to be corrected are required to be evaluated, the text information of the text to be corrected is input into the large language model, a second correction result is generated by the large language model at the same time, and when the whole evaluation of the text to be corrected is required to be performed, the text information of the text to be corrected is input into the large language model, and a third correction result is generated by the large language model at the same time.

Step 206, generating a browsing interface for displaying the first, second and/or third modification results in a preset mode.

After the correction result of the corresponding stage is generated according to the selection of the user, the correction result can be typeset and displayed, so that a corresponding browsing interface is obtained. For example, when the user executes the first-stage composition correction flow, a browsing interface for displaying the first correction result in a preset mode may be generated; when the user executes the second-stage composition correction flow, a browsing interface for displaying a second correction result in a preset mode can be generated; when the user executes the composition correction flow of the third stage, a browsing interface for displaying the third correction result in a preset mode can be generated.

Step 208, adjusting the first modification result, the second modification result and/or the third modification result in response to the adjustment operation received based on the browsing interface.

Specifically, the adjustment operation may be any operation such as clicking, sliding, dragging, etc. input by the user on the browsing interface, or may be a voice instruction or a gesture instruction input by the user based on the browsing interface, which is not limited in particular.

In this step, the adjustment operation input by the user may be received by using the browsing interface, and after the adjustment operation input by the user is received, the first modification result, the second modification result and/or the third modification result in the browsing interface are adjusted, so as to implement interface interaction in the modification process. Specifically, one or more of the first, second and third modification results may be adjusted according to the user's needs.

And 210, generating a target display page according to the adjusted correction result, and displaying the target display page.

In this step, after the first, second and/or third modification results are adjusted, a target display page may be generated according to the adjusted modification results, and the target display page may be displayed to the user, so that the user may conveniently view the final modification effect.

In this embodiment, after the text information of the to-be-modified composition is modified by using the preset large language model, an adjustment operation of the user for the browsing interface is received, and the first modification result, the second modification result and/or the third modification result are adjusted in response to the adjustment operation, so that interface interaction in the composition modification process is realized, and a target display page finally displayed to the user is more ideal.

In some embodiments, the browsing interface includes a preview area and an operation area, the preview area is used for displaying an altering preview image, and the altering preview image is generated based on text content of the text to be altered or picture content of the text to be altered; the operation area is used for displaying the correction content details corresponding to the initial correction result and receiving adjustment operation for the correction content details, and the correction content details are generated based on text information.

Specifically, the modification preview image is a preview image with modification marks generated based on text content of the text to be modified or picture content of the text to be modified. The correction mark can be an annotation symbol for modifying the character, such as an annotation symbol corresponding to a new character, a deleted character, a replaced character and the like; or the sentences or paragraphs can be marked by colors, such as green marks for bright sentences, blue marks for sentences to be modified, and the like. The preview area and the operation area are located in different areas of the browsing interface, for example, a left area of the browsing interface may be used as the preview area, and a right area of the browsing interface may be used as the operation area; the right area of the browsing interface may be used as the preview area, and the left area of the browsing interface may be used as the operation area; the upper area of the browsing interface may be used as the preview area, and the lower area of the browsing interface may be used as the operation area; the lower area of the browsing interface may be used as the preview area, and the upper area of the browsing interface may be used as the operation area. As an alternative embodiment, the left area of the browsing interface may be used as the preview area 301, and the right area of the browsing interface may be used as the operation area 302, as shown in fig. 3. In this way, the user can view the modification preview image in the preview area 301 of the left area, view the modification content details in the operation area 302 of the right area, and input the adjustment operation to adjust the modification content details, thereby providing convenience for the user to view the modification result and adjust the modification result.

In some embodiments, generating a browsing interface that displays the first modification result, the second modification result, and/or the third modification result in a preset mode includes:

Typesetting and displaying the first correction result, the second correction result and/or the third correction result according to the operation flow in stages to obtain a browsing interface, wherein the first correction result, the second correction result and the third correction result are displayed in batches according to the stage of the operation flow, and the first correction result is displayed before the second correction result and the third correction result.

The composition and correction operation flow can be performed in multiple stages, for example, the composition and correction operation flow can include a first stage for modifying characters in text information of the composition to be corrected, a second stage for evaluating sentences and paragraphs in the text information of the composition to be corrected, a third stage for evaluating the composition to be corrected as a whole, and the like, so that one or more stages can be selectively performed by a user according to correction requirements of the user, thereby obtaining correction results required by the user, and meanwhile, typesetting and displaying the first correction result, the second correction result and/or the third correction result in stages according to the operation flow, thereby obtaining a browsing interface for displaying the first correction result, the second correction result and/or the third correction result in a preset mode. By the mode, the first correction result, the second correction result and the third correction result can be displayed in batches according to the stage of the operation flow, and the user does not need to click to check one by one, so that the convenience of checking the correction results by the user is further improved; and if the user is not satisfied with the first correction result and the second correction result and the third correction result, the user can manually debug or supplement correction content, so that the convenience of the user in adjusting the correction result is further improved.

The first modification result is displayed before the second modification result and the third modification result in an externalization manner, because the first modification result needs to be generated before the second modification result and the third modification result. That is, when the text to be modified is modified, it is necessary to modify the characters in the text information of the text to be modified, and then evaluate sentences and paragraphs in the text information of the text to be modified or evaluate the text to be modified as a whole. Thus, when the sentence evaluation, the segment evaluation or the general evaluation of the to-be-corrected composition is generated by using the large language model, the input text information is ensured to be accurate, and the accuracy of the sentence evaluation, the segment evaluation or the general evaluation is improved.

In the prior art, different highlighting colors are displayed correspondingly according to error types to prompt a user to check, but as the highlighting colors do not have error type recognition, when the error types are too many, multiple colors are gathered together, a 'full focus' visual defocus feeling can be generated, and the user needs to click to check detailed error types and change suggestions, so that the operation is troublesome. In the embodiment of the invention, the correction operation flow can be split into: the method has the advantages that three flow nodes of character endorsement, sentence endorsement and paragraph endorsement and the whole overall comment are displayed in batches according to the current flow node, and users do not need to click to view one by one.

In some embodiments, under the condition that the first correction result is displayed on the browsing interface, the correction preview is a character correction preview, the correction content details are first correction details for correcting characters in the text information, the character correction preview is displayed with correction symbols corresponding to the corrected characters, the display positions of the correction symbols in the character correction preview are determined according to the positions of the corrected characters in the character correction preview, and different correction symbols correspond to different character correction types;

Responding to the adjustment operation received based on the browsing interface, adjusting the first correction result, the second correction result and/or the third correction result, including:

And responding to a first adjustment operation, and synchronously adjusting the first correction details and the character correction preview, wherein the first adjustment operation refers to an operation input in an operation area corresponding to the first correction details.

Specifically, the first modification details are different from the character modification preview image in that: the first correction details are correction results displayed based on the text information of the identified text to be corrected, and the character correction preview is correction results displayed based on the original picture content or the original text content of the text to be corrected input by the user; the first modification details can be used for checking modification detailed information of each modified character, and the character modification preview image only displays annotation symbols corresponding to the modified characters. The character to be modified may include a word unit to be modified, punctuation marks to be modified, and the like.

As an alternative embodiment, the browsing interface for displaying the first correction result may be as shown in fig. 4, where the character correction preview (i.e. the character correction preview in fig. 4) displays the annotation symbol corresponding to the modified character, such as the newly added character symbol, the deleted character symbol, the replaced character symbol, and so on. The display position of the annotation symbol in the character modification preview is determined according to the position of the modified character in the character modification preview. Specifically, a character recognition technology, such as an optical character recognition (Optical Character Recognition, abbreviated as OCR) technology, may be used to identify the position of the modified character in the batch to be modified and the size of each character (i.e., the area occupied by each character), and then calculate the position and size of the character in the modification preview according to the position of the character in the batch and the size of each character. For example, when deleting the "contracted" word in fig. 4, the position and the size of the "contracted" word in the character correcting preview image can be determined according to the position and the size of the "contracted" word in the composition, and then the deleted character symbol is displayed at the predetermined area of the position; when the "shallow" word in fig. 4 is newly added, the position of the "shallow" word in the character correcting preview can be determined according to the positions and sizes of the "short" word and the "shallow" word in the composition, then the newly added character symbol is displayed at the preset area of the position, and the like, so that all the character modifications in the composition to be corrected can be completed.

And when a user inputs a first adjustment operation in an operation area corresponding to the first correction details, the first correction details and the character correction preview can be synchronously adjusted. For example, characters are newly added, deleted, and the like at the same time as the first modification details and the character modification preview. Therefore, after the artificial intelligence is automatically corrected, a user only needs to check the accuracy of the correction result, the browsing pressure of the whole text is reduced, and in order to ensure the flexibility of use and the accuracy of the correction result, the user is supported to check leakage and repair deficiency, the correction result is manually added, and the error correction result is also supported to be recovered.

In some embodiments, the first adjustment operation includes at least one of a character modification new operation and a character modification deletion operation;

And responding to the first adjustment operation, and synchronously adjusting the first correction details and the character correction preview, wherein the method comprises the following steps of:

Responding to a character correcting new operation, correcting the first correcting details and the character which is not corrected temporarily in the character correcting preview, wherein the new character correcting operation comprises a selection operation of any character in the text information and a selection operation of correcting type; and/or the number of the groups of groups,

And responding to the character correcting and deleting operation, correcting and deleting the first correcting details and the corrected characters in the character correcting preview, wherein the character correcting and deleting operation comprises a selection operation of any corrected character in the text information and a selection operation based on a restoration menu in a first prompt box, and the first prompt box is generated in response to the selection operation of any corrected character in the text information.

Specifically, as shown in fig. 5, when the user needs to add a new character correction result, a certain non-correction character in the text information may be selected first, and then a correction type selection menu may be popped up near the character, and the user may select the correction type by selecting the correction type selection menu, and then pop up an editing window, where the user completes editing. For example, when the user needs to modify the two words in fig. 5, the two words may be selected first, at this time, the modification type selection menu shown in the upper diagram in fig. 5 is displayed on the browsing interface, after the user selects the menu on the right side, the browsing interface pops up the editing window shown in the lower diagram in fig. 5, at this time, the user may replace the two words with any word that wants to be replaced in the editing window, and then click the "complete" button to newly display the corresponding modification result in the first modification details and the character modification preview diagram. By adopting the mode, the first correction details and the character correction results in the character correction preview can be newly added.

As shown in fig. 6, when the user needs to delete the character correcting result, a corrected character in the text information may be selected first, and then a first prompt box may pop up near the character, and the user may view the modification details of the character in the first prompt box, and may select a restoration menu in the first prompt box to restore the correcting result of the character. For example, when the user needs to restore the modification of the "stare" word in fig. 6, the user may first select the "fix" word, and at this time, the browsing interface will display the first prompt box shown in fig. 6, and when the user clicks the "restore" button in the first prompt box, the user may restore the "fix" word to the "stare" word, and delete the corresponding modification result in the first modification details and the character modification preview. By adopting the mode, the deletion of the first correcting details and the character correcting result in the character correcting preview image can be realized.

In this embodiment, the above manner may be adopted to quickly implement the operations of adding and deleting the character correction result, so as to improve the efficiency of composition correction. Of course, in practical application, other ways may be used to implement the addition and deletion of the character correcting result, and the embodiment of the invention is not limited in particular.

In some embodiments, the first adjustment operation further comprises at least one of a paragraph merge operation, a paragraph split operation, and a paragraph order adjustment operation;

In response to paragraph merging operation, merging any two adjacent paragraphs in the text information, wherein the paragraph merging operation is clicking operation on a first icon, the first icon is in a display state when an operation response instance moves to an area where two adjacent paragraphs to be merged are located, and is in a hidden state when the operation response instance moves out of the area where the two adjacent paragraphs to be merged are located; and/or the number of the groups of groups,

Responding to paragraph splitting operation, namely splitting any paragraph in the text information, wherein the paragraph splitting operation is clicking operation on a second icon, the second icon is in a display state when an operation response example moves to an area where the paragraph to be split is located, and is in a hidden state when the operation response example moves out of the area where the paragraph to be split is located; and/or the number of the groups of groups,

And responding to paragraph sequence adjusting operation, and adjusting the ordering position of any paragraph in the text information, wherein the paragraph sequence adjusting operation is a drag operation of a paragraph card corresponding to the paragraph to be adjusted.

Specifically, the operation response example may be a cursor formed by a mouse on the interface, or an input signal formed by a click operation performed by a user on the interface, which is not limited in this embodiment.

As shown in fig. 7, when the user needs to merge paragraphs, the operation response instance may be moved to the region where two adjacent paragraphs to be merged are located, and the region will display the first icon at this time, and then the user may click on the first icon, so that the merging operation of the two adjacent paragraphs can be implemented. For example, when the user needs to merge two text segments in fig. 7, the operation response instance may be moved to the middle of the two text segments, and the browsing interface displays a first icon as shown in fig. 7, and after the user clicks the first icon, the user can merge the upper segment and the lower segment. It should be noted that, the first icon is in a display state when the operation response instance moves to the region where the two adjacent paragraphs to be merged are located, and is in a hidden state when the operation response instance moves out of the region where the two adjacent paragraphs to be merged are located, so that paragraph merging caused by misoperation of a user can be effectively prevented.

As shown in fig. 8, when the user needs to split a paragraph, the operation response instance may be moved to the area where the paragraph to be split is located, and the area displays a second icon at this time, and then the user may click on the second icon, so as to implement the splitting operation of the paragraph. For example, when the user needs to split a text segment in fig. 8, the operation response instance may be moved to the position of the text segment, and at this time, the browsing interface displays a second icon as shown in fig. 8, and after the user clicks the second icon, the user can split the text segment. It should be noted that, the second icon is in a display state when the operation response instance moves to the region where the paragraph to be split is located, and is in a hidden state when the operation response instance moves out of the region where the paragraph to be split is located, so that the paragraph splitting caused by misoperation of a user can be effectively prevented.

Under the condition that a user needs to adjust the paragraph sequence, the paragraph card corresponding to the paragraph to be adjusted can be directly dragged, and then the paragraph card is released after being dragged to the target position, so that the paragraph sequence adjustment can be completed. The paragraph card is understood to be an independent gray area corresponding to each paragraph shown in fig. 7 or 8. It should be noted that, after the user completes paragraph merging, paragraph splitting or paragraph sequence adjustment, the electronic device may reorder all paragraphs of the pending composition.

In this embodiment, the above manner may be adopted to quickly implement operations such as merging, splitting, and sequence adjustment on paragraphs, so as to improve efficiency of composition correction. Of course, in practical application, other manners may be adopted to implement merging, splitting and sequence adjustment of paragraphs, and embodiments of the present invention are not limited in particular.

In some embodiments, when the second correction result is displayed on the browsing interface, the correction preview is a sentence correction preview or a paragraph correction preview, the correction content details are second correction details for evaluating sentences and paragraphs in the text information, the sentence correction preview is used for performing color filling display on the evaluated sentences, different filling colors represent different evaluation types, and the paragraph correction preview is used for performing highlighting display on the evaluated paragraphs;

and in response to a second adjustment operation, synchronously adjusting the second correction details and the sentence correction preview, or in response to the second adjustment operation, synchronously adjusting the second correction details and the paragraph correction preview, wherein the second adjustment operation refers to an operation input in an operation area corresponding to the second correction details.

Specifically, the sentence correction preview may perform color filling display on the sentence to be evaluated, where different filling colors represent different evaluation types, for example, filling excellent sentences with green, filling sentences to be optimized with blue, and so on. Thus, the user can quickly locate the corresponding sentence through the sentence preview and quickly judge the quality of the sentence according to the color. The paragraph correction preview can highlight the evaluated paragraphs, and normally display the non-evaluated paragraphs. Therefore, the user can quickly locate the corresponding paragraph through the paragraph preview, and the user can conveniently compare and check the paragraph comments. The second altering detail is different from the sentence altering preview image or the paragraph altering preview image in that: the second correction details are correction results displayed based on the text information of the identified text to be corrected, and the sentence correction preview image or the paragraph correction preview image is correction results displayed based on the picture content or the text content of the original text to be corrected input by the user; the second correction details can be used for checking the modification detailed information of each rated sentence or paragraph, and the sentence correction preview image or paragraph correction preview image only displays the color information or brightness information and the like corresponding to the rated sentence or paragraph.

And when a user inputs a second adjustment operation in an operation area corresponding to the second modification details, the second modification details and the sentence modification preview image can be synchronously adjusted, or the second modification details and the paragraph modification preview image can be synchronously adjusted.

Therefore, after the artificial intelligence is automatically modified, the user can check the sentence evaluation and the paragraph evaluation in batches without clicking one by one, and the evaluated sentences and paragraphs can be displayed in the preview image in a distinguishing way, so that the user can conveniently and quickly position the sentences and paragraphs.

In some embodiments, the second adjustment operation includes at least one of a regeneration operation of a single sentence comment, a regeneration operation of all sentence comments, and a cancel application operation of the single sentence comment;

And in response to the second adjustment operation, performing synchronous adjustment on the second correction details and the sentence correction preview, including:

any sentence comment in the second correction details and the sentence correction preview is adjusted in response to a regeneration operation of the single sentence comment, wherein the regeneration operation of the single sentence comment is a clicking operation of a first control in an area where the sentence comment to be adjusted is located, the first control is in a display state when an operation response instance moves to the area where the sentence comment to be adjusted is located, and is in a hidden state when the operation response instance moves out of the area where the sentence comment to be adjusted is located; and/or the number of the groups of groups,

Responding to the regenerating operation of all sentence comments, and adjusting the second correcting details and all sentence comments in the sentence correcting preview, wherein the regenerating operation of all sentence comments is the clicking operation of a second control at a first preset position of a browsing interface; and/or the number of the groups of groups,

And canceling the application of any sentence comment in the second correction details and the sentence correction preview in response to a cancellation application operation of the single sentence comment, wherein the cancellation application operation of the single sentence comment is a click operation of a third control in an area where the sentence comment to be cancelled is located.

As an alternative embodiment, the browsing interface for displaying the second modification result may be as shown in fig. 9. In the case that the user needs to regenerate a single sentence comment, the operation response instance may be moved to the region where the sentence comment to be adjusted is located, where the region may display a first control, such as the "inaccurate, GPT re-optimized" link identified in fig. 9. When the user clicks the first control, a new sentence comment of the sentence can be regenerated by using the large language model. By adopting the mode, the repeated execution of the correction flow on a single sentence evaluation can be supported without repeated execution of the correction flow on all sentence evaluation, thereby improving the operation flexibility and reducing the system load. It should be noted that, the first control is in a display state when the operation response instance moves to the region where the sentence evaluation to be adjusted is located, and is in a hidden state when the operation response instance moves out of the region where the sentence evaluation to be adjusted is located, so that the regeneration of a new sentence evaluation caused by misoperation of a user can be effectively prevented.

Under the condition that the user needs to regenerate all sentence scores, the user can directly click on a second control at the first preset position of the browsing interface, such as the uppermost control of regeneration in fig. 9, and at this time, new sentence scores of all sentences can be regenerated by using the large language model. It should be noted that, the first preset position may be set according to actual needs, for example, an uppermost area or a lowermost area of the browsing interface.

When the user needs to cancel the application of a single sentence comment, the user can directly click a third control in the area where the sentence comment to be cancelled is located, for example, a switch of "application" corresponding to each sentence comment in fig. 9, and then the application of the sentence comment can be cancelled. It should be noted that the third controls are in one-to-one correspondence with the sentence evaluations, and each sentence evaluation corresponds to one third control, so that a user can determine whether to apply the sentence evaluations according to the third control corresponding to each sentence evaluation, thereby determining the sentence evaluation finally displayed on the target display page.

By the method, the adjustment of single sentence evaluation or all sentence evaluation can be quickly realized, so that the efficiency of correcting the composition is improved. Of course, in practical application, sentence evaluation adjustment may be implemented in other manners, and the embodiment of the present invention is not limited specifically.

In some embodiments, the second adjustment operation further includes at least one of a single segment review regeneration operation, a regeneration operation of all segment reviews, and a cancel application operation of the single segment review;

and in response to the second adjustment operation, performing synchronous adjustment on the second modification details and the paragraph modification preview image, including:

any section comment in the second correction details and the section correction preview is adjusted in response to a regeneration operation of the single section comment, wherein the regeneration operation of the single section comment is a clicking operation of a fourth control in an area where the section comment to be adjusted is located, the fourth control is in a display state when an operation response example moves to the area where the section comment to be adjusted is located, and is in a hidden state when the operation response example moves out of the area where the section comment to be adjusted is located; and/or the number of the groups of groups,

Responding to the regenerating operation of all section comments, and adjusting the second correcting details and all section comments in the section correcting preview image, wherein the regenerating operation of all section comments is clicking operation of a fifth control at a second preset position of the browsing interface; and/or the number of the groups of groups,

And canceling the application of any section comment in the second correction details and the section correction preview in response to the cancellation application operation of the single section comment, wherein the cancellation application operation of the single section comment is a clicking operation of a sixth control in the area where the section comment to be cancelled is located.

As shown in fig. 10, in the case where the user needs to regenerate a single segment comment, the operation response instance may be moved to the area where the segment comment to be adjusted is located, where the area may display a fourth control, such as the "re-moistened" link identified in fig. 10. When the user clicks the fourth control, a new paragraph comment of the paragraph can be regenerated by using the large language model. It should be noted that, the fourth control is in a display state when the operation response instance moves to the region where the section comment to be adjusted is located, and is in a hidden state when the operation response instance moves out of the region where the section comment to be adjusted is located, so that the regeneration of a new section comment caused by misoperation of a user can be effectively prevented.

Under the condition that the user needs to regenerate all section comments, the user can directly click on a fifth control at a second preset position of the browsing interface, such as the uppermost control of regeneration in fig. 10, and at this time, a large language model can be utilized to regenerate new section comments of all sections. It should be noted that the second preset position may be set according to actual needs, for example, an upper area or a lower area of the browsing interface.

When the user needs to cancel the application of the single section comment, the user can directly click a sixth control in the region where the section comment to be cancelled is located, for example, an application switch corresponding to each frequency band in fig. 10, and then the application of the section comment can be cancelled. It should be noted that the sixth controls are in one-to-one correspondence with the section comments, and each section comment corresponds to one sixth control, so that a user can determine whether to apply the section comments according to the sixth control corresponding to each section comment, thereby determining the section comment finally displayed by the target display page.

By adopting the mode, the adjustment of single section evaluation or all section evaluation can be quickly realized, so that the efficiency of composition correction is improved. It should be noted that, when the sentence comment content slides to the tail end, tab will automatically switch to the paragraph comment and the preview is synchronously switched, as shown in fig. 10, which matches the offline modification sequence, and the operation is natural and smooth.

In some embodiments, under the condition that the third modification result is displayed on the browsing interface, the modification preview image is a sentence modification preview image or a paragraph modification preview image, and the modification content details are third modification details for performing overall evaluation on the text information;

and responding to a third adjustment operation, and adjusting the third modification details, wherein the third adjustment operation refers to an operation input in an operation area corresponding to the third modification details.

Specifically, the third modification details can be understood as contents for evaluating the to-be-modified composition from a plurality of different dimensions, such as contents from a plurality of dimensions including encouragement speech, general evaluation, language expression, structural logic, content depth, article bright spots, to-be-enhanced items, modification suggestions and the like, so that the granularity of information can be refined, and professional and quality evaluation contents can be given.

And when a user inputs a third adjustment operation in an operation area corresponding to the third modification details, the third modification details can be adjusted.

Thus, after the artificial intelligence is automatically modified, the user can view the general comments in batches without clicking one for viewing.

In some embodiments, the third adjustment operation includes at least one of a regeneration operation of all the critique, a cancel application operation of all the critique, and a cancel application operation of a single critique;

And adjusting the third modification details in response to the third adjustment operation, including:

responding to the regenerating operation of all the criticisms, and adjusting all the criticisms in the third modification details, wherein the regenerating operation of all the criticisms is the clicking operation of a seventh control at a third preset position of the browsing interface; and/or the number of the groups of groups,

Responding to the cancel application operation of all the criticisms, and adjusting all the criticisms in the third modification details, wherein the cancel application operation of all the criticisms is the click operation of an eighth control at a fourth preset position of the browsing interface, and the target display page does not contain the criticisms after the criticisms are cancelled; and/or the number of the groups of groups,

And canceling the application of any general comment in the third correction details in response to a cancel application operation of the single general comment, wherein the cancel application operation of the single general comment is a click operation of a ninth control in an area where the general comment is to be canceled, and the target display page does not contain the general comment after the cancel application.

Specifically, as shown in fig. 10, in the case that the user needs to regenerate all the critique, the user may click on the seventh control at the third preset position of the browsing interface, so that a new critique may be regenerated. It should be noted that, the third preset position may be set according to actual needs, for example, an upper area or a lower area of the browsing interface.

And under the condition that the user needs to cancel all the general comments of the application, the user can directly click an eighth control at a fourth preset position of the browsing interface, and then the application of all the general comments can be canceled. It should be noted that, the fourth preset position may be set according to actual needs, for example, an upper area or a lower area of the browsing interface.

And under the condition that the user needs to cancel the application single general comment, the user can directly click a ninth control in the area where the general comment to be cancelled is located, and then the application of the general comment can be cancelled. It should be noted that the ninth controls are in one-to-one correspondence with the general scores, and each general score corresponds to one ninth control, so that a user can determine whether to apply the general scores according to the ninth control corresponding to each general score, thereby determining the general score finally displayed on the target display page.

By the method, adjustment of single overall evaluation or all overall evaluation can be achieved rapidly, and efficiency of composition correction is improved.

In some embodiments, the text content in the modification preview is displayed in alignment with the text content in the modification content details, and when the modification result corresponding to the target paragraph in the modification content details is in the adjusted state, the target paragraph is highlighted in the corresponding region in the modification preview.

Specifically, the text content in the correction preview image is always aligned with the text content in the correction content detail, for example, when the correction preview image shows the first few pieces of content of the to-be-corrected text, the character modification, sentence evaluation or segment evaluation of the first few pieces of content of the to-be-corrected text is also shown in the correction content detail; when the last pieces of content of the to-be-modified text are displayed in the modification preview, character modification, sentence evaluation or segment evaluation of the last pieces of content of the to-be-modified text are also displayed in the modification content details. When the user slides the correction content details up and down, the correction preview image is scrolled along with the correction content details, so that the text content in the correction preview image is consistent with the text content in the correction content details in a browsing area all the time. By the method, when the initial correction result is adjusted, a user can conveniently compare and check the text content in the correction preview image with the text content in the correction content detail. In addition, when the correction result (such as clicking or editing text) corresponding to the target paragraph in the correction content detail is adjusted, the corresponding area of the target paragraph in the correction preview image can be highlighted, so that the non-associated paragraph can be weakened, the user can focus on the current operation, and the visual dispersion is avoided.

In some embodiments, generating the target presentation page according to the adjusted modification result includes:

dividing the adjusted correction preview image according to paragraphs to obtain a plurality of preview fragments, wherein the adjusted correction preview image is marked with annotation symbols corresponding to the adjusted first correction result;

And typesetting the adjusted second correction result and the adjusted third correction result based on the plurality of preview segments to obtain a target display page, wherein the sentence evaluation sum Duan Pingfen applied in the adjusted second correction result is distributed near the corresponding preview segment, and the total evaluation applied in the adjusted third correction result is distributed at the uppermost position and/or the last lower position of the target display page.

Specifically, when the target display page is generated according to the adjusted correction result, the adjusted correction preview image is divided according to paragraphs to obtain a plurality of preview segments, then sentence evaluation and segment evaluation applied in the adjusted second correction result are typeset near the corresponding preview segments, and total evaluation applied in the adjusted third correction result is typeset at the uppermost position and/or the last lower position of the target display page to obtain the target display page, as shown in fig. 11.

It should be noted that the target display page may have various display effects, such as handwriting photo-taking board, printing board, and the like. The handwriting photographing plate is different from the printing body plate in that: the preview segment of the former is a picture corresponding to a section of characters in the photographed picture of the handwriting, and the preview segment of the latter is a section of characters in the character content of the printing body. Considering uncontrollable factors such as large difference of writing effects of different student users, large section of sequence adjustment and the like, when the effect of the handwriting photo plate is unsatisfactory, the alternative printing body version long chart is directly applied without re-correction, so that the repeated labor probability can be reduced.

In the prior art, most of the output objects which do not support downloading are mainly full-text general scores and multidimensional scores on the content side, the granularity of information is coarse, and the comparison relation between labels and original sentences or paragraphs is weak. In the embodiment of the invention, the project can be debugged according to the writing key points and requirements corresponding to the grades of students and the service requirements, so that the artificial intelligence (ARTIFICIAL INTELLIGENCE, which is simply called AI) can have the key points to produce the total evaluation contents of a plurality of dimensions; and after the paragraph sequence and the size range are calculated, the composition original image is cut into a plurality of fragments according to paragraphs, the corresponding annotation content is matched according to the set layout rule, and the comparison relation between the labels and the original images in the output object is effectively enhanced.

In some embodiments, the text information is modified by using a preset large language model to obtain an initial modification result, including:

acquiring feature information corresponding to the to-be-modified text, wherein the feature information comprises learning stage information of an author of the to-be-modified text and the type of the to-be-modified text;

matching target prompt information matched with the characteristic information from a preset prompt information library, wherein a plurality of prompt information corresponding to different characteristic information are stored in the preset prompt information library, and the plurality of prompt information comprise target prompt information;

And inputting the target prompt information and the text information into a preset large language model for correction, and obtaining an initial correction result.

Specifically, the learning stage information of the author of the to-be-modified composition may refer to the student grade, the stage of the course learned by the student, and the like. The types of the composition to be modified can be classified according to the language of the composition, such as Chinese composition, english composition and the like; or classifying according to the subject matter of the composition, such as memorizing people, notes, scenes, etc. The preset prompt information library stores a plurality of different prompt information (namely, prompt) corresponding to different characteristic information, the prompt information is a guiding text provided when the user interacts with the large language model, and the large language model can accurately know the intention and the problem of the user by writing clear and definite prompt, so that more accurate answers are obtained.

In the process of generating an initial correction result by using a preset large language model, feature information corresponding to a text to be corrected can be acquired first, then target prompt information matched with the feature information is matched from a preset prompt information library, and then the target prompt information and text information are input into the preset large language model for correction, so that the initial correction result is obtained.

By the method, the target prompt information can be determined according to the characteristic information corresponding to the text to be modified, and then the initial modification result is obtained by utilizing the target prompt information, so that the initial modification result is more accurate.

In some embodiments, the second modification result and the third modification result each include a plurality of evaluation results with different dimensions, and the dimensions and the number of the evaluation results are obtained by writing target prompt information.

Specifically, the sentence evaluation and the paragraph evaluation in the second modification result and the total evaluation in the third modification result may be plural, and the sentences, paragraphs, and the whole in the to-be-modified text may be evaluated from plural different dimensions. As an alternative implementation mode, sentence evaluation and segment evaluation can produce four dimensions of omnipotent beginning, suggestion, color rendering and bright spot, and total evaluation can produce eight-dimensional content of encouraging speaking, total evaluation, language expression, structural logic, content depth, article bright spot, item to be lifted and modification suggestion. In practical application, target prompt information can be written according to practical requirements, so that the dimensions and the number of sentence evaluation, segment evaluation and general evaluation can be determined. The dimensions and numbers in the above examples are for illustrative purposes only and do not limit the invention.

In this way, the granularity of the information can be refined, and a professional and quality sentence evaluation, segment evaluation and general evaluation correction reference can be given.

In some embodiments, identifying text information corresponding to the text to be modified includes:

Judging whether the acquired text of the to-be-modified text is a handwriting text or not;

Under the condition that the acquired text of the to-be-modified text is a handwriting text, identifying the handwriting text to obtain text information, wherein the font type of the text information is consistent with that of the handwriting text;

And under the condition that the acquired text of the to-be-modified text is a non-handwritten text, identifying the non-handwritten text to obtain text information, wherein the non-handwritten text is a printed text with any font type, and the font type of the text information is consistent with that of the non-handwritten text.

Specifically, after uploading the text to be modified by the user, whether the acquired text of the text to be modified is a handwriting text can be judged first, and if the acquired text of the text to be modified is the handwriting text, the handwriting text is identified, so that text information consistent with the font type of the handwriting text is obtained; and if the acquired text of the to-be-modified text is the non-handwritten text, identifying the non-handwritten text to obtain text information consistent with the font type of the non-handwritten text.

In this way, the electronic device can preferentially identify the handwriting text, and if the handwriting text is non-handwriting text, the electronic device automatically switches to print text identification without manually selecting the font type by the user.

In some embodiments, before identifying the text information corresponding to the text to be modified, the method further comprises:

displaying prompt information in an uploading interface of the to-be-modified text, wherein the prompt information is used for prompting the requirements required by uploading the picture;

and under the condition that the picture content of the to-be-modified text is received, cutting the picture content of the to-be-modified text.

Specifically, the prompt information may be text, picture, etc., and the present invention is not limited specifically.

When the user needs to upload the picture content of the text to be modified, the picture content of the text to be modified can be uploaded according to the prompt information displayed in the uploading interface. As shown in fig. 12, the left area performs example guidance on the uploading standard, and reduces the probability of abnormal situations such as "handwriting blurring, uneven brightness, font missing, left and right column writing" and the like. In addition, after the uploading of the picture content of the to-be-modified composition is successful, the electronic device can also automatically cut the picture content of the to-be-modified composition, for example, automatically cut a picture containing multiple pages of composition content into multiple pictures containing one page of composition content, or cut the redundant blank frame positions on the pictures, and the like. Therefore, more standard pictures can be formed, the picture contents can be conveniently and subsequently identified, text information of the to-be-modified text is obtained, and a corresponding modification preview picture is generated. In addition, after the uploading of the picture content of the to-be-modified composition is successful, the user can drag and change the picture display sequence so as to improve the fault tolerance.

In addition, when the user needs to upload text content of the text to be modified, the uploading interface can be switched through a switching button on the uploading interface, so that a new uploading interface is obtained. On the new upload interface, the user may manually input or copy the text field into the input box to enable uploading of the text content of the text to be modified. Of course, there is also a switch button in the new upload interface in order for the user to switch the upload interface to the original upload interface.

In the prior art, whether a user selects picture uploading or text input uploading, the picture uploading or text input uploading is converted into printed text, so that the text recognition in reading is improved, but the scene is modified off line. In the off-line correction scene, a teacher can directly annotate on student original texts, writing is a mode of judging and scoring the composition, and fonts are the fastest mode of distinguishing the composition of the student. However, the prior art cannot directly annotate the student text, and cannot judge the composition according to the writing quality. In the embodiment of the invention, the final output object still uses the student handwriting, and the annotating symbol or annotating content is reasonably matched to the corresponding position of the original text by combining the annotating habit, the information arrangement mode, the symbol use logic and the like of the offline teacher, so that the offline correcting scene and the correcting effect are maximally fitted. Secondly, in the prior art, OCR recognition depends on writing standard, and once the original image has scenes such as paragraph exchange, multi-character deletion and the like, the accuracy of OCR recognition is low. Most of the bid products are uploaded, namely, corrected, and few of the bid products have manual adjustment processes, but the adjustment efficiency is low because the adjustment has no additional quick operation. In the embodiment of the invention, because the picture is exemplified during the uploading of the composition, the picture cutting function is added, and the picture standardization is guided, the OCR recognition accuracy is higher. Before the user sends the text to the GPT, an original text correction flow is introduced, and quick operations such as paragraph merging and sequence adjustment are given to the user in the flow, so that the manual adjustment efficiency is improved.

Exemplary apparatus

Having introduced the method of the exemplary embodiment of the present invention, next, a description will be given of the topic recommendation device of the exemplary embodiment of the present invention with reference to fig. 4.

FIG. 13 schematically illustrates a functional block diagram of an artificial intelligence based composition altering interaction device according to an embodiment of the invention. As shown in fig. 13, the apparatus 1300 may include:

The identifying module 1302 is configured to identify text information corresponding to the to-be-modified text, where the text information is obtained based on the obtained text content of the to-be-modified text or is obtained based on the obtained picture content identification of the to-be-modified text;

The correction module 1304 is configured to correct the text information by using a preset large language model to obtain an initial correction result, where the initial correction result includes a first correction result for correcting characters in the text information, a second correction result for evaluating sentences and paragraphs in the text information, and/or a third correction result for integrally evaluating a to-be-corrected text;

the generating module 1306 is configured to generate a browsing interface that displays the first modification result, the second modification result, and/or the third modification result in a preset mode;

an adjustment module 1308, configured to adjust the first modification result, the second modification result, and/or the third modification result in response to an adjustment operation received based on the browsing interface;

And the display module 1310 is configured to generate a target display page according to the adjusted correction result, and display the target display page.

In one embodiment of the invention, the browsing interface comprises a preview area and an operation area, wherein the preview area is used for displaying an altering preview image which is generated based on text content of a text to be altered or picture content of the text to be altered; the operation area is used for displaying the correction content details corresponding to the initial correction result and receiving adjustment operation for the correction content details, and the correction content details are generated based on text information.

In one embodiment of the invention, the generating module 1306 includes:

And the typesetting and displaying sub-module is used for typesetting and displaying the first correction result, the second correction result and/or the third correction result in stages according to the operation flow to obtain a browsing interface, wherein the first correction result, the second correction result and the third correction result are displayed in batches according to the stage of the operation flow, and the first correction result is displayed before the second correction result and the third correction result.

In one embodiment of the invention, under the condition that a first correction result is displayed on a browsing interface, the correction preview is a character correction preview, the correction content details are first correction details for correcting characters in text information, the character correction preview is displayed with annotation symbols corresponding to the corrected characters, the display positions of the annotation symbols in the character correction preview are determined according to the positions of the corrected characters in the character correction preview, and different annotation symbols correspond to different character correction types; the adjustment module 1308 includes:

And the first adjustment sub-module is used for responding to a first adjustment operation and synchronously adjusting the first correction details and the character correction preview, wherein the first adjustment operation is an operation input in an operation area corresponding to the first correction details.

In one embodiment of the present invention, the first adjustment operation includes at least one of a character modification new operation and a character modification deletion operation; the first adjustment submodule includes:

The correcting unit is used for responding to the character correcting new operation and correcting the first correcting details and the character which is not corrected temporarily in the character correcting preview image, wherein the new character correcting operation comprises a selecting operation for any character in the text information and a selecting operation for correcting types; and/or the number of the groups of groups,

And the correction deleting unit is used for correcting and deleting the first correction details and corrected characters in the character correction preview image in response to a character correction deleting operation, wherein the character correction deleting operation comprises a selection operation of any corrected character in the text information and a selection operation based on a restoration menu in a first prompt box, and the first prompt box is generated in response to the selection operation of any corrected character in the text information.

In one embodiment of the present invention, the first adjustment operation further includes at least one of a paragraph merging operation, a paragraph splitting operation, and a paragraph order adjustment operation; the first adjustment sub-module further includes:

The merging unit is used for responding to paragraph merging operation, and merging any two adjacent paragraphs in the text information, wherein the paragraph merging operation is clicking operation on a first icon, the first icon is in a display state when an operation response example moves to an area where the two adjacent paragraphs to be merged are located, and is in a hidden state when the operation response example moves out of the area where the two adjacent paragraphs to be merged are located; and/or the number of the groups of groups,

The splitting unit is used for responding to the paragraph splitting operation to split any paragraph in the text information, wherein the paragraph splitting operation is clicking operation on a second icon, the second icon is in a display state when the operation response example moves to the area where the paragraph to be split is located, and is in a hidden state when the operation response example moves out of the area where the paragraph to be split is located; and/or the number of the groups of groups,

The first adjusting unit is used for responding to paragraph sequence adjusting operation, and adjusting the ordering position of any paragraph in the text information, wherein the paragraph sequence adjusting operation is a drag operation of a paragraph card corresponding to the paragraph to be adjusted.

In one embodiment of the invention, under the condition that the second correction result is displayed on the browsing interface, the correction preview image is a sentence correction preview image or a paragraph correction preview image, the correction content details are second correction details for evaluating sentences and paragraphs in the text information, the sentence correction preview image is used for carrying out color filling display on the evaluated sentences, different filling colors represent different evaluation types, and the paragraph correction preview image is used for carrying out highlighting display on the evaluated paragraphs; the adjustment module 1308 further includes:

And the second adjustment sub-module is used for synchronously adjusting the second correction details and the sentence correction preview images in response to a second adjustment operation, or synchronously adjusting the second correction details and the paragraph correction preview images in response to the second adjustment operation, wherein the second adjustment operation refers to an operation input in an operation area corresponding to the second correction details.

In one embodiment of the present invention, the second adjustment operation includes at least one of a regeneration operation of a single sentence comment, a regeneration operation of all sentence comments, and a cancel application operation of a single sentence comment; the second adjustment submodule includes:

The second adjusting unit is used for responding to the regenerating operation of the single sentence evaluation, and adjusting any sentence evaluation in the second correcting detail and the sentence correcting preview, wherein the regenerating operation of the single sentence evaluation is the clicking operation of the first control in the area where the sentence evaluation to be adjusted is located, the first control is in a display state when the operation response example moves to the area where the sentence evaluation to be adjusted is located, and is in a hidden state when the operation response example moves out of the area where the sentence evaluation to be adjusted is located; and/or the number of the groups of groups,

The third adjusting unit is used for responding to the regenerating operation of all sentence scores, and adjusting the second correcting details and all sentence scores in the sentence correcting preview, wherein the regenerating operation of all sentence scores is the clicking operation of the second control at the first preset position of the browsing interface; and/or the number of the groups of groups,

And the first application unit is used for responding to the cancellation application operation of the single sentence evaluation, and canceling the application of any sentence evaluation in the second correction details and the sentence correction preview, wherein the cancellation application operation of the single sentence evaluation is the clicking operation of a third control in the area where the sentence evaluation to be cancelled is located.

In one embodiment of the present invention, the second adjustment operation further includes at least one of a single segment score regenerating operation, and a single segment score canceling operation; the second adjustment sub-module further includes:

the fourth adjusting unit is used for responding to the regenerating operation of the single section comment and adjusting any section comment in the second correcting detail and the paragraph correcting preview image, wherein the regenerating operation of the single section comment is the clicking operation of a fourth control in the area of the section comment to be adjusted, the fourth control is in a display state when the operation response example moves to the area of the section comment to be adjusted, and is in a hidden state when the operation response example moves out of the area of the section comment to be adjusted; and/or the number of the groups of groups,

The fifth adjusting unit is used for responding to the regenerating operation of all section comments and adjusting the second correcting details and all section comments in the paragraph correcting preview, wherein the regenerating operation of all section comments is the clicking operation of a fifth control at a second preset position of the browsing interface; and/or the number of the groups of groups,

And the second application unit is used for responding to the cancel application operation of the single section comment, and canceling the application of any section comment in the second correction details and the section correction preview, wherein the cancel application operation of the single section comment is a click operation of a sixth control in the area where the section comment to be canceled is located.

In one embodiment of the invention, under the condition that the third modification result is displayed on the browsing interface, the modification preview image is a sentence modification preview image or a paragraph modification preview image, and the modification content details are third modification details for performing overall evaluation on the text information; the adjustment module 1308 further includes:

and the third adjustment sub-module is used for responding to a third adjustment operation to adjust the third modification details, wherein the third adjustment operation refers to the operation input in the operation area corresponding to the third modification details.

In one embodiment of the present invention, the third adjustment operation includes at least one of a regenerating operation of all the critique, a cancel application operation of all the critique, and a cancel application operation of a single critique; the third adjustment submodule includes:

The sixth adjusting unit is used for responding to the regenerating operation of all the criticisms, and adjusting all the criticisms in the third modification details, wherein the regenerating operation of all the criticisms is the clicking operation of a seventh control at a third preset position of the browsing interface; and/or the number of the groups of groups,

A seventh adjustment unit, configured to adjust all the total scores in the third modification details in response to an application canceling operation of all the total scores, where the application canceling operation of all the total scores is a click operation of an eighth control at a fourth preset position of the browsing interface, and the target display page does not include the total scores after being cancelled; and/or the number of the groups of groups,

And the third application unit is used for responding to the application canceling operation of the single general comment and canceling the application of any general comment in the third correction details, wherein the application canceling operation of the single general comment is the clicking operation of the ninth control in the area where the general comment is to be canceled, and the target display page does not contain the general comment after the application canceling.

In one embodiment of the invention, the text content in the modification preview is displayed in alignment with the text content in the modification content detail, and when the modification result corresponding to the target paragraph in the modification content detail is in the adjusted state, the corresponding region of the target paragraph in the modification preview is highlighted.

In one embodiment of the invention, the display module 1310 includes:

The marking sub-module is used for dividing the adjusted correction preview image according to paragraphs to obtain a plurality of preview fragments, wherein the adjusted correction preview image is marked with annotation symbols corresponding to the adjusted first correction result;

And the typesetting sub-module is used for typesetting the adjusted second correction result and the adjusted third correction result based on the plurality of preview segments to obtain a target display page, wherein the sentence evaluation and Duan Pingfen applied in the adjusted second correction result are distributed near the corresponding preview segments, and the total evaluation applied in the adjusted third correction result is distributed at the uppermost position and/or the last lower position of the target display page.

In one embodiment of the invention, the adaptation module 1304 includes:

The obtaining submodule is used for obtaining the characteristic information corresponding to the to-be-modified text, wherein the characteristic information comprises learning stage information of an author of the to-be-modified text and the type of the to-be-modified text;

The matching sub-module is used for matching target prompt information matched with the characteristic information from a preset prompt information library, wherein a plurality of prompt information corresponding to different characteristic information are stored in the preset prompt information library, and the plurality of prompt information comprise target prompt information;

and the correction sub-module is used for inputting the target prompt information and the text information into a preset large language model for correction, so as to obtain an initial correction result.

In one embodiment of the present invention, the second modification result and the third modification result each include a plurality of evaluation results with different dimensions, and the dimensions and the number of the evaluation results are obtained by compiling target prompt information.

In one embodiment of the invention, the identification module 1302 includes:

the judging submodule is used for judging whether the acquired text of the to-be-modified text is a handwriting text or not;

The first recognition sub-module is used for recognizing the handwriting text to obtain text information when the acquired text of the to-be-modified text is the handwriting text, and the font type of the text information is consistent with that of the handwriting text;

and the second recognition sub-module is used for recognizing the non-handwritten text to obtain text information when the acquired text of the to-be-modified text is the non-handwritten text, wherein the non-handwritten text is a printed text with any font type, and the font type of the text information is consistent with the font type of the non-handwritten text.

In one embodiment of the invention, the apparatus 1300 further comprises:

The display module is used for displaying prompt information in an uploading interface of the to-be-modified text, wherein the prompt information is used for prompting the requirements for uploading the pictures;

and the clipping module is used for clipping the picture content of the text to be modified under the condition that the picture content of the text to be modified is received.

The apparatus 1300 according to the embodiment of the present invention has been described and explained in detail above in connection with the method, and will not be described again here.

In a third aspect of the embodiments of the present invention, there is provided an electronic device, as shown in fig. 14, comprising a processor 1411, a communication interface 1412, a memory 1413, and a communication bus 1414, wherein the processor 1411, the communication interface 1412, the memory 1413 complete communication with each other through the communication bus 1414,

A memory 1413 for storing a computer program;

In one embodiment of the present application, the processor 1411 is configured to implement the artificial intelligence based composition modifying interaction method provided in any of the foregoing method embodiments when executing the program stored on the memory 1413.

The embodiment of the application also provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the artificial intelligence based composition correction interaction method provided by any one of the method embodiments described above.

Embodiments of the present application also provide a computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to implement the artificial intelligence based composition altering interaction method provided by any of the method embodiments described above.

It should be noted that although several devices or sub-devices of the electronic apparatus are mentioned in the above detailed description, this division is not mandatory only. Indeed, the features and functions of two or more of the devices described above may be embodied in one device, in accordance with embodiments of the present invention. Conversely, the features and functions of one device described above may be further divided into multiple devices to be embodied.

Furthermore, although the operations of the methods of the present invention are depicted in the drawings in a particular order, this is not required or suggested that these operations must be performed in this particular order or that all of the illustrated operations must be performed in order to achieve desirable results. Rather, the steps depicted in the flowcharts may change the order of execution. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step to perform, and/or one step decomposed into multiple steps to perform.

It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable medium that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable medium produce an article of manufacture including instruction means which implement the function/act specified in the flowchart and/or block diagram block or blocks.

The computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other devices to cause a series of operational steps to be performed on the computer, other programmable apparatus or other devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions/acts specified in the flowchart and/or block diagram block or blocks.

Use of the verb "comprise," "include" and its conjugations in this application does not exclude the presence of elements or steps other than those stated in the application. The article "a" or "an" preceding an element does not exclude the presence of a plurality of such elements.

While the spirit and principles of the present invention have been described with reference to several particular embodiments, it is to be understood that the invention is not limited to the disclosed embodiments nor does it imply that features of the various aspects are not useful in combination, nor are they useful in any combination, such as for convenience of description. The invention is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. The scope of the following claims is to be accorded the broadest interpretation so as to encompass all such modifications and equivalent structures and functions.

Claims

1. An artificial intelligence based composition correction interaction method, which is characterized by comprising the following steps:

2. The method of claim 1, wherein the browsing interface includes a preview area and an operation area, the preview area is used for displaying a correction preview image, and the correction preview image is generated based on text content of the to-be-corrected text or picture content of the to-be-corrected text; the operation area is used for displaying the correction content details corresponding to the initial correction result and receiving adjustment operation for the correction content details, and the correction content details are generated based on the text information.

3. The method according to claim 1, wherein the generating a browsing interface that displays the first, second and/or third modification results in a preset mode comprises:

Typesetting and displaying the first correction result, the second correction result and/or the third correction result according to the operation flow in stages to obtain the browsing interface, wherein the first correction result, the second correction result and the third correction result are displayed in batches according to the stage of the operation flow, and the first correction result is displayed in a manner of being exteriorized before the second correction result and the third correction result.

4. The method according to claim 2, wherein, in the case that the first modification result is displayed on the browsing interface, the modification preview is a character modification preview, the modification content details are first modification details for modifying characters in the text information, the character modification preview displays modification symbols corresponding to the modified characters, and display positions of the modification symbols in the character modification preview are determined according to positions of the modified characters in the character modification preview, and different modification types of the modification symbols correspond to different character modification types;

The responding to the adjustment operation received based on the browsing interface adjusts the first correction result, the second correction result and/or the third correction result, including:

5. The method according to claim 2, wherein in the case that the second correction result is displayed on the browsing interface, the correction preview image is a sentence correction preview image or a paragraph correction preview image, the correction content details are second correction details for evaluating sentences and paragraphs in the text information, the sentence correction preview image is used for performing color filling display on the evaluated sentences, different filling colors represent different evaluation types, and the paragraph correction preview image is used for performing highlighting display on the evaluated paragraphs;

And in response to a second adjustment operation, synchronously adjusting the second correction details and the sentence correction preview, or in response to a second adjustment operation, synchronously adjusting the second correction details and the paragraph correction preview, wherein the second adjustment operation refers to an operation input in an operation area corresponding to the second correction details.

6. The method according to claim 2, wherein in a case where the browsing interface displays the third modification result, the modification preview is the sentence modification preview or the paragraph modification preview, and the modification content details are third modification details for performing overall evaluation on the text information;

7. An artificial intelligence based composition altering interaction device, the device comprising:

8. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;

A memory for storing a computer program;

a processor for implementing the artificial intelligence based composition modifying interaction method of any one of claims 1-6 when executing a program stored on a memory.

9. A computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the artificial intelligence based composition altering interaction method of any of claims 1-6.

10. A computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to implement the artificial intelligence based composition altering interaction method of any one of claims 1-6.