Interaction-based form intelligent optimization method and device, electronic equipment and storage medium

Abstract

The application relates to the technical field of interactive form intelligent optimization, and discloses an interactive form intelligent optimization method and device, an electronic device and a storage medium, wherein the method comprises the following steps: displaying two pages respectively carrying original form content and form content filled by a large language model, superimposing a difference confirmation panel on a modified form page, and centrally presenting field names, original values and changed values of all modified fields in the form of a changed item list; in response to a first operation of a user on a changed item in the difference confirmation panel, corresponding modified form content is highlighted in the modified form page; and in response to a completion instruction, the difference between the modified form content and the original form content and the filled form content is displayed on the modified form page. The application has the beneficial technical effect that the generated target form content is accurate and reliable, and the controllability of interaction and the reliability of results in a man-machine cooperation scene are improved.

Description

Technical Field

[0001] This application relates to the field of interactive form intelligent optimization technology, and in particular to an interactive form intelligent optimization method, apparatus, electronic device and storage medium. Background Technology

[0002] In the fields of artificial intelligence and human-computer interaction, intelligent form filling systems based on large language models have been widely used in scenarios requiring structured information collection. However, in the process of large language model-assisted filling, existing intelligent form systems directly output the filling results. If the content to be filled is large, it requires a significant amount of work and time, leading to low operational efficiency for users during the form confirmation process after the large language model has filled the form. Summary of the Invention

[0003] Therefore, it is necessary to propose an interactive form intelligent optimization method, device, electronic device and storage medium to address the existing interactive form intelligent optimization problem.

[0004] An interactive form intelligent optimization method, the method comprising: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

[0005] Furthermore, before the steps of displaying the filled form content (after field filling is completed by the large language model) on the modified form page and displaying the original form content on the original form page, the method further includes: When the original form page is fully loaded, cache the field identifiers and current values ​​of all input controls on the original form page as a snapshot of the original state; After the large language model completes the field prediction and assignment operations, the assigned form content is rendered to the modified form page, and the current state of the modified form page is compared with the original state snapshot to generate the change item in the differential confirmation panel.

[0006] Furthermore, each of the change entries in the differential confirmation panel also includes a change confidence indicator, which is presented next to the field name of the corresponding change entry in the form of a color grade or a percentage value; the color grade or percentage value is determined based on the confidence score assigned to the corresponding fill operation by the large language model.

[0007] Further, the step of displaying the corresponding modification form content on the modification form page in response to a first operation on any change item in the difference confirmation panel includes: In response to a first operation on any change entry in the differential confirmation panel, the target field corresponding to the target change entry and its current change value are highlighted on the modification form page in a manner different from the display of other fields.

[0008] Furthermore, after the step of displaying the corresponding modification form content on the modification form page in response to the first operation on any change item in the difference confirmation panel, the method further includes: In response to the confirmation trigger operation of the target change item in the differential confirmation panel, the status of the target change item is marked as confirmed, and a lock icon is displayed in the corresponding target field in the modification form page; In response to the cancellation trigger operation of the target change entry in the differential confirmation panel, the status of the target change entry is marked as cancelled, and the current value of the corresponding target field in the modification form page is restored to the original value.

[0009] Further, the step of displaying the differences between the modified form content and the original form content and the filled form content, respectively, on the modified form page in response to the completion instruction of the difference confirmation panel, includes: In response to a trigger operation on the completion control in the difference confirmation panel, a difference summary information bar is generated and displayed in the top or side area of ​​the modified form page; wherein the difference summary information bar includes at least: a first difference statistics item, used to indicate the total number of changed fields in the modified form content relative to the original form content; and a second difference statistics item, used to indicate the total number of unchanged fields in the modified form content relative to the original form content.

[0010] Furthermore, in response to the completion instruction on the difference confirmation panel, after displaying the differences between the modified form content and the original form content and the filled form content on the modified form page, the method further includes: A required field completion guide panel is overlaid on the modified form page; wherein, the required field completion guide panel displays the field names of at least one missing required field in the current modified form content in order of business priority; In response to receiving completion information for the currently displayed missing required fields, the system parses the completion information and writes the parsed completion values ​​into the corresponding required fields in the modified form content, thereby obtaining the modified form content.

[0011] An interactive form intelligent optimization device, the device comprising: The first display module is used to display the original form page and the modified form page, including the target form; The second display module is used to display the filled form content after the fields are filled by the large language model on the modified form page, and to display the original form content on the original form page; The first response module is used to display a difference confirmation panel overlaid on the modified form page in response to the completion of the large language model filling; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; The second response module is used to respond to the first operation on any change item in the difference confirmation panel and display the corresponding modification form content on the modification form page. The third response module is used to respond to the completion instruction of the difference confirmation panel and display the differences between the modified form content and the original form content and the filled form content on the modified form page. The generation module is used to generate the target form corresponding to the modified form content in response to the new form generation instruction.

[0012] An electronic device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

[0013] A computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the following steps: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

[0014] The beneficial effects of this application are as follows: By displaying two pages, one containing the original form content and the other containing the form content populated by the large language model, a difference confirmation panel is overlaid on the modified form page. All modified fields, their original values, and changed values ​​are presented in a centralized list of changed items. This allows users to quickly understand all the modifications to the large language model within a single interface area, avoiding the need to search for changed fields one by one in dense forms. Furthermore, in response to the user's first action on a changed item in the difference confirmation panel, the corresponding modified form content is highlighted on the modified form page. Upon completion, the page displays the differences between the modified form content and the original and populated form content, providing users with a summary overview of the differences before final confirmation. This provides sufficient verification for the final generation process, ensuring the accuracy and reliability of the final generated target form content and improving the controllability and credibility of the results in human-computer collaboration scenarios. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] in: Figure 1 This is a diagram illustrating the application environment of an interactive form intelligent optimization method in one embodiment. Figure 2 This is a flowchart of an interactive form intelligent optimization method in one embodiment; Figure 3 This is a structural block diagram of an interactive form intelligent optimization device in one embodiment; Figure 4 This is a structural block diagram of an electronic device in one embodiment. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] Figure 1 This is a diagram illustrating an interactive form-based intelligent optimization application environment in one embodiment. (Refer to...) Figure 1This interactive form intelligent optimization method is applied to an interactive form intelligent optimization system. The system includes a terminal 110 and a server 120. The terminal 110 and server 120 are connected via a network. The terminal 110 can be a desktop terminal or a mobile terminal; a mobile terminal can be at least one of a mobile phone, tablet, or laptop. The server 120 can be a standalone server or a server cluster consisting of multiple servers. The terminal 110 displays the original form page and the modified form page, including the target form, while the server 120 generates the target form corresponding to the modified form content.

[0019] like Figure 2 As shown, in one embodiment, an interactive form intelligent optimization method is provided. This method can be applied to both terminals and servers; this embodiment uses terminal application as an example. The interactive form intelligent optimization method specifically includes the following steps: S1: Displays the original form page and the modified form page, including the target form; S2: On the modified form page, display the filled form content after the fields are filled by the large language model, and on the original form page, display the original form content; S3: In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; S4: In response to the first operation on any change item in the difference confirmation panel, display the corresponding modification form content on the modification form page; S5: In response to the completion instruction of the difference confirmation panel, on the modified form page, display the differences between the modified form content and the original form content and the filled form content respectively; S6: In response to the new form generation instruction, generate the target form corresponding to the modified form content.

[0020] As described in step S1 above, the original form page and the modified form page, including the target form, are displayed. Through the terminal or server display interface, two different forms—the original form page and the modified form page—are presented to the user simultaneously. The original form page contains the initial form content without intervention from the large language model, including raw data manually entered by the user or pre-filled by the system. The modified form page, on the other hand, is the target carrier for subsequent field filling and intelligent optimization by the large language model. Specifically, it refers to a parallel display interface with the same structure as the original form page, used to carry the large language model's filling results and the user's modification operations. In other words, the original form page is the form page that the user initially opens or loads, containing raw form content that has not been processed by the large language model; the modified form page is a separate parallel display area set up to carry the large language model's filling results. The two pages are displayed using a left-right column, a top-bottom column, or a tabbed switching method.

[0021] In one specific embodiment, the terminal device (such as a personal computer or tablet computer) displays a left-right column layout: the left area is the original form page, and the right area is the modified form page. The user first opens a customer information registration form in a browser or application. This form contains more than ten fields, including name, license plate number, identification code, vehicle model, purchase date, contact number, and email address. In the original form page, some of these fields may be empty, while others may be filled with initial content manually entered by the user. The modified form page is a blank copy with the exact same structure as the original form page but with content yet to be filled. The two pages are displayed simultaneously, facilitating a field-by-field comparison between the content generated by the large language model and the original content. In another embodiment, a top-bottom column layout can also be used, with the original form page displayed in the upper half and the modified form page in the lower half, to accommodate different screen sizes and user operating habits.

[0022] As described in step S2 above, the modified form page displays the filled form content after the fields are filled by the large language model, and the original form page displays the original form content. The large language model is invoked to intelligently fill in each form field on the modified form page, and the filling result is presented on the modified form page in real time; at the same time, the original form page retains its original content.

[0023] As described in step S3 above, in response to the large language model completing the population, a difference confirmation panel is overlaid on the modified form page. This difference confirmation panel is an independent UI component overlaid on the modified form page, used to centrally display all changes made by the large language model to the form fields. Each entry includes the field name, original value, and changed value. The modified form page displays at least one change entry in a list, each entry including the field name, original value, and changed value of the changed field. After detecting that the large language model has returned the population result and completed rendering, a difference confirmation panel is automatically overlaid on top of the modified form page. This panel displays all changes made by the large language model to the form fields in a list format—that is, which field values ​​changed, what the original values ​​were, and what the changed values ​​were.

[0024] In one specific embodiment, after the large language model completes the filling of the customer information registration form, it automatically displays a semi-transparent difference confirmation panel overlaid on the right side of the form modification page. This panel lists all the changed field entries in tabular form, for example: Each change entry occupies one line. Field names are highlighted in bold, original values ​​are shown in gray to indicate what has been replaced, and changed values ​​are shown in blue or green to indicate what has been newly entered. Change statistics are displayed at the top of the panel, such as a total of 5 fields showing changes. Users can quickly understand the specific modifications made to the form by the large language model without having to compare the left and right pages one by one.

[0025] As described in step S4 above, in response to the first operation on any change item in the differential confirmation panel, the corresponding modified form content is displayed on the modified form page. The first operation refers to the interactive action performed by the user on the change item to trigger field-based location, including but not limited to single-click, double-click, hover, or touch-click. When the user performs the first operation (e.g., single-click, double-click, hover, or touch-click) on a change item in the differential confirmation panel, the specific field corresponding to that change item and its modified content are highlighted or otherwise visually emphasized on the modified form page. This panel item-page field linkage location mechanism allows users to quickly jump from the abstract change list to the specific form field location, achieving a seamless transition from overview to detailed examination. Each change item stores the DOM element identifier or field path of the corresponding field. When the user triggers the first operation, the target field on the modified form page is located through this identifier and highlighted or scrolled for location.

[0026] In one specific embodiment, the difference confirmation panel lists change entries for the identification code field (original value pending → changed to the corresponding identification code). When the user clicks on an entry, the modification form page immediately scrolls to the location of the identification code field, and a blue highlighted border appears around the field, fading to a light color after 2 seconds. Simultaneously, the input box for that field gains focus, allowing the user to directly edit it. In another embodiment, when the user hovers the mouse over the purchase date change entry, a floating preview appears at the corresponding purchase date field location on the modification form page, displaying the complete context information of that field. When the user clicks on the entry, the page directly jumps to that field and enters edit mode. In the mobile embodiment, when the user touches and clicks on the contact number entry in the difference confirmation panel, the modification form page automatically scrolls to the contact number field, and a virtual keyboard pops up in the field's edit box, allowing the user to directly input changes. The differential confirmation panel is a draggable panel. When the linked highlight is triggered, if the highlighted field is within the coverage area of ​​the differential confirmation panel, the differential confirmation panel will be automatically moved to a position that does not obscure the highlighted field. Alternatively, the linked highlight will be synchronously displayed in the differential confirmation panel as a flashing item.

[0027] As described in step S5 above, in response to the completion command on the differential confirmation panel, the modified form page displays the differences between the modified form content and the original form content, as well as the filled form content. After the user performs confirmation or cancellation operations on all change items in the differential confirmation panel one by one, the completion control switches from an unavailable state to an available state; when all change items have been processed, the user triggers the completion control to issue a completion command. The summary display step, triggered after the user completes the review and confirmation operations on all change items in the differential confirmation panel, displays three sets of comparison information visually on the modified form page when the user issues a completion command (e.g., clicking the confirmation completion button, clicking the submit button, or performing a specific gesture operation): the difference between the modified form content and the original form content, and the difference between the modified form content and the content initially filled in by the Large Language Model (LLM). This three-fold comparison display allows the user to review the entire changes to the form one last time before final submission, ensuring that all modifications meet expectations. For example, different color codes can be used to distinguish three types of content: green indicates the changed parts of the modified content relative to the original content, red indicates the adjusted parts of the modified content relative to the filled content, and blue indicates the parts that are consistent across all three.

[0028] In one specific embodiment, after reviewing all change items one by one in the difference confirmation panel, the user clicks the "Confirm Complete" button at the bottom of the panel. The difference comparison rendering mode is then activated on the modified form page: for each field in the modified form, it is compared with the corresponding field in the original form and the corresponding field in the initial populated content of the large language model. For example, if the modified value of the identifier field is 12345, the original value is pending, and the populated value is 12345, the system displays a green checkmark next to the identifier field, indicating that the modified value has overwritten the original value and matches the populated value. Similarly, if the modified value of the car purchase date field is 2026-07-15, the original value is 2026-06-01, and the populated value is 2026-07-15, a green checkmark is displayed. The modified value of the contact number field is 139xxxx5678, the original value is 138xxxx1234, and the filled value is 139xxxx1234. The system displays a yellow warning icon next to this field, indicating that the user manually modified the value of this field, making it different from the filled value, prompting the user to confirm whether this is their intention. A summary bar is displayed at the top of the entire modification form page: a total of 6 fields have changed, of which 5 are consistent with the filled value suggestion, and 1 was manually adjusted by the user.

[0029] As described in step S6 above, in response to the new form generation instruction, a target form corresponding to the modified form content is generated. Once the user confirms all changes to the modified form content are correct, a new form generation instruction is issued (e.g., clicking the "Generate New Form" button, clicking the "Save" button, or performing a submit operation), solidifying the final content on the modified form page and generating an independent target form. This target form can be a data object stored in a structured data format, such as eXtensible Markup Language (XML), or a formal document output in a visual document format, such as Portable Document Format (PDF), or a data record submitted to a backend database or business system. This step marks the completion of the entire process: original form → large language model population → user review and confirmation → target form generation, achieving a complete transformation from the initial state to the optimized final state.

[0030] In one specific implementation, after confirming that all form modifications are correct, the user clicks the "Generate New Form" button at the bottom of the modified form page. First, the modified form content undergoes data validation, checking that all required fields are filled and that the data format is correct (e.g., email address, phone number, etc.). After successful validation, the system encapsulates the modified form content into a data object containing the following information: form template ID, field names and values, generation timestamp, operator information, and version number (initially V1.0). This data object is stored in the form instance table of the backend database, generating a unique record. Simultaneously, a report generation service is invoked to render the data object into a formatted PDF file. The filename includes the generation date and form type (e.g., Customer Information Registration Form_20260706.pdf), and a download link is provided. After receiving a notification that the target form has been successfully generated, the user can download the PDF file for printing or archiving, or continue to view or edit the target form within the system.

[0031] In one embodiment, before step S2, which involves displaying the filled form content (after field filling is performed by the large language model) on the modified form page and displaying the original form content on the original form page, the method further includes: S101: When the original form page is fully loaded, cache the field identifiers and current values ​​of all input controls on the original form page as a snapshot of the original state; S102: After the large language model completes the field prediction and assignment operation, the assigned form content is rendered to the modified form page, and the current state of the modified form page is compared with the original state snapshot to generate the change item in the differential confirmation panel.

[0032] As described in step S101 above, when the original form page is fully loaded, the field identifiers and current values ​​of all input controls on the original form page are cached as an original state snapshot. The original state snapshot refers to the data set formed by caching the field identifiers and current values ​​of all input controls when the original form page is fully loaded. All input controls on the original form page (including but not limited to text boxes, dropdown selection boxes, radio buttons, checkboxes, date pickers, number input boxes, rich text editors, etc.) have completed the initial rendering of the DOM (Document Object Model), and their current values ​​(which may be empty, default values, or values ​​manually entered by the user) have been determined. At this moment, all input controls on the original form page are traversed, and the field identifiers (such as field names, control IDs, data binding paths, or unique identifiers) and their corresponding current values ​​are extracted one by one. These key-value pairs are then organized into a structured data object—the original state snapshot. By adopting the idea of ​​snapshot isolation, the original data and subsequent changed data are decoupled in the time dimension, ensuring the accuracy and repeatability of change detection.

[0033] As described in step S102 above, after the large language model completes the field prediction and assignment operations, the assigned form content is rendered to the modified form page, and the current state of the modified form page is compared with the original state snapshot to generate the change item in the differential confirmation panel. The execution flow of this step can be divided into three closely connected sub-stages: First, the large language model is called to perform semantic prediction and numerical assignment on each field of the target form. This process involves multiple rounds of prompt word engineering, external knowledge base retrieval, or contextual reasoning. Second, the filling results returned by the large language model (usually a set of key-value pairs of field names and suggested values) are rendered to the corresponding input controls on the modified form page to complete the state update of the modified form page. Then, based on the original state snapshot cached in S101, each input control in the current state of the modified form page is traversed field by field, its current value is extracted, and compared one by one with the original value of the corresponding field in the original state snapshot. The comparison logic follows preset difference determination rules. For example, if the current value is not equal to the original value (including both type and numerical comparisons), the field is determined to be a changed field; if the current value is the same as the original value, it is determined to be an unchanged field. The system organizes all field information (field name, original value, changed value) determined to be changed fields into a structured list of change entries, which serves as the data source for the difference confirmation panel in the subsequent S3 step.

[0034] In one embodiment, each of the change entries in the differential confirmation panel further includes a change confidence indicator, which is presented next to the field name of the corresponding change entry in the form of a color grade or a percentage value; the color grade or percentage value is determined based on the confidence score assigned to the corresponding fill operation by the large language model.

[0035] The change confidence indicator is dynamically associated with the confidence score output by the large language model when generating the field's fill value. The confidence score is a probability measure or uncertainty estimate assigned by the large language model to a specific prediction result during the decoding process, typically reflecting the model's degree of certainty regarding the fill result. This score is presented next to the field name of the corresponding change item in the form of a color-coded level (e.g., red, orange, yellow, and green, corresponding to confidence scores below 60%, 60%-75%, 75%-90%, and above 90%, respectively) or a percentage value (e.g., 92%). Essentially, this transforms the implicit uncertainty measure within the large language model into an explicit visual signal that users can intuitively perceive. By introducing the confidence indicator, users can prioritize focusing their limited attention on low-confidence change items while quickly confirming high-confidence items, thus optimizing the allocation of review resources.

[0036] Confidence Score: In specific technical implementations, when generating sequences, large language models typically output a probability distribution for each predicted token. For the generation of the entire field value, the overall confidence score can be calculated in various ways, such as: taking the average log probability of all generated tokens for that field, taking the minimum token probability of the generated sequence (representing the weakest link in confidence), or using a built-in confidence calibration mechanism (such as probability after temperature scaling). When sending a fill request to the large language model, the system should explicitly request the return of the confidence score for each field in the prompt word or Application Programming Interface (API) parameters, or calculate the confidence score itself by parsing the logits or probability values ​​output by the model after receiving the raw response. In a specific embodiment, the confidence score is calculated as follows: For a sequence {t1, t2, ..., tn} containing n tokens generated by the large language model for field f... n}, record each token i The generation probability p i Then field-level confidence If C_f ≥ 0.85, it is judged as high confidence (green mark); if 0.70 ≤ C_f < 0.85, it is judged as medium confidence (yellow mark); if C_f < 0.70, it is judged as low confidence (red mark).

[0037] In one embodiment, step S4, which involves displaying the corresponding modified form content on the modified form page in response to a first operation on any change item in the differential confirmation panel, includes: S401: In response to a first operation on any change item in the differential confirmation panel, the target field corresponding to the target change item and its current change value are highlighted in the modification form page in a manner different from the display of other fields.

[0038] As described in step S401 above, in response to a first operation on any change item in the differential confirmation panel, the target field corresponding to the target change item and its current change value are highlighted on the modification form page in a display manner distinct from other fields. This is the specific visual response made on the modification form page when a user performs a first operation (e.g., single click, double click, hover, or touch tap) on a change item in the differential confirmation panel. The core features of this response are highlighting and distinguishing it from other fields. Highlighting refers to changing the visual attributes of the target field (such as background color, border color, font weight, adding animation effects, displaying indicator marks, etc.) to make it distinguishable among all fields on the page, thereby guiding the user's visual attention to quickly locate the field. Distinguishing it from other fields emphasizes that this highlighting is a differentiated treatment; that is, the visual style of the target field forms a clear contrast with the default styles of all other non-target fields on the modification form page, allowing the user to easily identify the specific form field associated with the current operation. Furthermore, this step specifically emphasizes highlighting not only the target field (i.e., the overall input area or label area of ​​the field) but also the current changed value of the field (i.e., the specific content of the field). This dual highlighting design allows users to locate the field macroscopically and focus on its specific value microscopically, achieving simultaneous location and value review.

[0039] In one embodiment, after step S4, which involves displaying the corresponding modification form content on the modification form page in response to a first operation on any change entry in the difference confirmation panel, the method further includes: S501: In response to the confirmation trigger operation of the target change item in the differential confirmation panel, the status of the target change item is marked as confirmed, and a lock icon is displayed in the corresponding target field in the modification form page; S502: In response to the cancellation trigger operation of the target change item in the differential confirmation panel, the status of the target change item is marked as cancelled, and the current value of the corresponding target field in the modification form page is restored to the original value.

[0040] As described in step S501 above, in response to the confirmation trigger operation of the target change item in the differential confirmation panel, the status of the target change item is marked as confirmed, and a lock icon is displayed in the corresponding target field on the modification form page. The lock icon is a visual mark displayed in the corresponding field after the change is confirmed, indicating that the field has been manually confirmed and its content is locked and cannot be edited. When a user reviews a change entry in the differential confirmation panel (e.g., has viewed the modified content of the field through the S401 highlight linkage) and determines that the changed value filled by the large language model is accurate and does not need to be modified, the user performs a confirmation trigger operation (e.g., clicks the checkmark button on the right side of the entry, double-clicks the entry, or clicks the confirmation soft key). Upon receiving this operation instruction, two parallel technical actions are executed: First, at the data level, the status field of the change entry is updated from the default "pending" to "confirmed". This status marker is persistently stored in the change entry status array on the front end and is eventually uploaded along with the form submission data as part of the user's review record. Second, at the interface level, a lock icon (such as a padlock icon, a gray background overlay, or a "confirmed" text label) is overlaid on the target field position corresponding to the target change entry on the modification form page. This lock icon has a dual technical meaning: externally (to the user), it indicates that the field content has been manually confirmed and does not require further verification; internally (to the system), it usually means that the input control is set to read-only or disabled to prevent users from accidentally modifying the confirmed field content in subsequent operations, thereby ensuring the integrity and consistency of the data.

[0041] As described in step S502 above, in response to the cancellation trigger operation on the target change item in the difference confirmation panel, the status of the target change item is marked as cancelled, and the current value of the corresponding target field in the modification form page is restored to the original value. When a user reviews a change item and believes that the changed value filled by the large language model is inaccurate, unreasonable, or does not conform to business rules (e.g., the model mistakenly fills in an incorrect number for the contact phone number, or misjudges the identification code as a non-existent identification code), the user performs a cancellation trigger operation (e.g., clicking the right side of the item). Upon receiving the operation (e.g., via the cancel button, swiping the item left, or clicking the reject softkey), two core technical actions are performed: First, at the data level, the status of the change item is updated from the default pending to reverted, explicitly marking the change as unacceptable. Second, at the interface level, the current value of the target field corresponding to the target change item in the modified form page is forcibly restored from the changed value filled by the large language model to the original value in the original state snapshot cached in step S101. This restoration operation involves not only visual numerical refilling but also precise resetting of the input control state (such as text content, selected items in dropdown options, selected states of checkboxes, etc.) and triggering necessary form validation updates (e.g., after restoring an empty value, a required field validation prompt may need to be displayed).

[0042] In one embodiment, step S5, which displays the differences between the modified form content and the original form content and the filled form content respectively, on the modified form page in response to the completion instruction of the difference confirmation panel, includes: S511: In response to the triggering operation of the completion control in the difference confirmation panel, a difference summary information bar is generated and displayed in the top or side area of ​​the modified form page; wherein the difference summary information bar includes at least: a first difference statistics item, used to indicate the total number of changed fields of the modified form content relative to the original form content; and a second difference statistics item, used to indicate the total number of unchanged fields of the modified form content relative to the original form content.

[0043] As described in step S511 above, in response to the triggering operation of the completion control in the difference confirmation panel, a difference summary information bar is generated and displayed in the top or side area of ​​the modified form page. The difference summary information bar refers to the information aggregation area displayed in the top or side of the modified form page after the user triggers the completion command, used to display the difference statistics between the modified content and the original content and the filled content. The difference summary information bar includes at least: a first difference statistics item, used to indicate the total number of changed fields of the modified form content relative to the original form content; and a second difference statistics item, used to indicate the total number of unchanged fields of the modified form content relative to the original form content. When the user completes the item-by-item review decision of all change items in the difference confirmation panel (i.e., all are marked as confirmed or withdrawn), the first summary feedback is triggered by clicking the completion control (such as the confirm completion submission for review or generate summary button). This feedback does not directly display a detailed difference list field by field, but first generates a difference summary information bar in the top or side area of ​​the modified form page, a lightweight and highly readable information aggregation component. The core content of this summary section includes at least two contrasting statistical items: a first difference statistic (total number of changed fields) and a second difference statistic (total number of unchanged fields). These two items together outline the overall change map of the modified form content relative to the original form content. Users can quickly see how many fields have been modified and how many remain unchanged in the highest priority position on the page (top title area or side navigation area) without having to browse through each field individually. This summary statistical information provides a crucial final review window before the user submits the form. Its technical effect is that it helps users quickly determine whether the scope of modifications meets expectations through macro-level data indicators. For example, if a user expects to modify only 3 fields, but the summary shows that 15 fields have changed, the user can immediately realize that the large language model filling may have exceeded the expected range, thus timely interrupting the submission process, returning to check, and avoiding over-modification problems caused by batch automated filling.

[0044] In one embodiment, after step S5, which displays the differences between the modified form content and the original form content and the filled form content respectively, on the modified form page in response to the completion instruction of the difference confirmation panel, the method further includes: S601: Display a required field completion guide panel overlaid on the modified form page; wherein, the required field completion guide panel displays the field names of at least one missing required field in the current modified form content in order of business priority; S602: In response to receiving the completion information for the currently displayed missing required fields, the completion information is parsed, and the parsed completion value is written into the corresponding required field in the modified form content to obtain the modified form content.

[0045] As described in step S601 above, a required field completion guide panel is overlaid on the modified form page. This panel displays the names of at least one missing required field in the current modified form content, arranged according to business priority. The required field completion guide panel is a guiding UI component overlaid on the modified form page after the difference confirmation is completed, used to prompt the user to complete missing required fields according to priority. After the user triggers the completion command (i.e., clicks the "Complete Summary" button in S511) and views the difference summary information bar, the next step of data verification and guidance process is automatically triggered. At this time, all fields in the modified form content are scanned. Based on the preset required field metadata rules in the form template (such as the field-level `required:true` attribute or backend configured verification rules), required fields that are currently empty or have invalid placeholder values ​​are identified. Subsequently, a required field completion guide panel is overlaid on top of the modified form page. Its core feature is displaying missing required fields according to business priority. Business priority refers to sorting missing fields according to preset business rules. For example, core identification fields (such as contract number and applicant's ID number) take precedence over auxiliary information fields (such as emergency contact phone number), or they are arranged according to the business filling order of the form (from top to bottom).

[0046] As described in step S602 above, in response to receiving completion information for the currently displayed missing required fields, the completion information is parsed, and the parsed completion value is written into the corresponding required field in the modified form content to obtain the modified form content. When the user provides completion information for one or more missing required fields in the required field completion guide panel (e.g., entering a value in a text box, selecting from a drop-down option, selecting a date using a date picker, etc.) and triggers submission (e.g., clicking the submit completion button), the following core technical actions are performed: First, the user-input completion information is parsed—this parsing operation is far more than simple data reading, but includes data type conversion (e.g., string to number), format standardization (e.g., date format is unified to YYYY-MM-DD), data cleaning (e.g., removing leading and trailing spaces, filtering illegal characters), and optional semantic validation (e.g., regular expression matching for email format, length validation for phone numbers, etc.). After parsing and validation, the completed value is accurately written into the corresponding required field in the modified form content data model through field identifier (fieldId) mapping, thereby generating an updated modified form content. By securely merging user-completed input into the form data object in a structured and cleaned manner, the correctness of data type and the standardization of format are ensured, avoiding subsequent data storage or business processing anomalies caused by inconsistent user input formats.

[0047] Reference Figure 3 This application provides an interactive form intelligent optimization device, the device comprising: The first display module 902 is used to display the original form page and the modified form page, which include the target form; The second display module 904 is used to display the filled form content after the field filling is completed by the large language model on the modified form page, and to display the original form content on the original form page; The first response module 906 is used to display a difference confirmation panel overlaid on the modified form page in response to the completion of the large language model filling; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; The second response module 908 is used to respond to a first operation on any change item in the difference confirmation panel and display the corresponding modification form content in the modification form page. The third response module 910 is used to respond to the completion instruction of the difference confirmation panel and display the differences between the modified form content and the original form content and the filled form content on the modified form page. The generation module 912 is used to generate the target form corresponding to the modified form content in response to the new form generation instruction.

[0048] In one embodiment, the interactive form intelligent optimization device further includes: The caching module is used to cache the field identifiers and current values ​​of all input controls on the original form page as a snapshot of the original state when the original form page is fully loaded. The rendering module is used to render the assigned form content to the modified form page after the large language model completes the field prediction and assignment operations, and compare the current state of the modified form page with the original state snapshot to generate the change item in the differential confirmation panel.

[0049] In one embodiment, each of the change entries in the differential confirmation panel further includes a change confidence indicator, which is presented next to the field name of the corresponding change entry in the form of a color grade or a percentage value; the color grade or percentage value is determined based on the confidence score assigned to the corresponding fill operation by the large language model.

[0050] In one embodiment, the second response module 908 includes: The differentiation submodule is used to respond to a first operation on any change item in the difference confirmation panel, and to highlight the target field corresponding to the target change item and its current change value in the modification form page in a way that is different from the display of other fields.

[0051] In one embodiment, the interactive form intelligent optimization device further includes: The marking module is used to mark the status of the target change item as confirmed in response to the confirmation trigger operation of the target change item in the differential confirmation panel, and display a lock icon in the corresponding target field in the modification form page; The recovery module is used to respond to the cancellation trigger operation of the target change item in the differential confirmation panel, mark the status of the target change item as cancelled, and restore the current value of the corresponding target field in the modification form page to the original value.

[0052] In one embodiment, the third response module 910 includes: The system includes a submodule for generating and displaying a difference summary information bar in the top or side area of ​​the modified form page in response to a trigger operation of the completion control in the difference confirmation panel; wherein the difference summary information bar includes at least: a first difference statistics item for indicating the total number of changed fields in the modified form content relative to the original form content; and a second difference statistics item for indicating the total number of unchanged fields in the modified form content relative to the original form content.

[0053] In one embodiment, the interactive form intelligent optimization device further includes: The completion module is used to overlay a required field completion guide panel on the modified form page; wherein, the required field completion guide panel displays the field names of at least one missing required field in the current modified form content in order of business priority. The parsing module is used to respond to the received completion information for the currently displayed missing required fields, parse the completion information, and write the parsed completion value into the corresponding required field in the modified form content to obtain the modified form content.

[0054] Figure 4 An internal structural diagram of an electronic device in one embodiment is shown. This electronic device can specifically be a terminal or a server, and more specifically, a computer device. Figure 4 As shown, the electronic device includes a processor, a memory, and a network interface connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, this computer program enables the processor to implement an interactive form intelligent optimization method. The internal memory may also store a computer program, which, when executed by the processor, enables the processor to implement an interactive form intelligent optimization method. Those skilled in the art will understand that... Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0055] In one embodiment, an electronic device is provided, including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the following steps: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

[0056] By displaying two separate pages—one containing the original form content and the other containing the form content populated by the large language model—a difference confirmation panel is overlaid on the modified form page. All modified fields, their original values, and changed values ​​are presented in a centralized list of change items. This allows users to quickly understand all modifications to the large language model within a single interface area, avoiding the need to search for changed fields one by one in a dense form. Then, in response to the user's first action on a change item in the difference confirmation panel, the corresponding modified form content is highlighted on the modified form page. Upon completion, the page displays the differences between the modified form content and the original and populated form content, providing users with a summary overview of the differences before final confirmation. This provides sufficient verification for the final generation process, ensuring the accuracy and reliability of the final generated target form content and improving the controllability and credibility of the results in human-computer collaboration scenarios.

[0057] In one embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, causes the processor to perform the following steps: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

[0058] By displaying two separate pages—one containing the original form content and the other containing the form content populated by the large language model—a difference confirmation panel is overlaid on the modified form page. All modified fields, their original values, and changed values ​​are presented in a centralized list of change items. This allows users to quickly understand all modifications to the large language model within a single interface area, avoiding the need to search for changed fields one by one in a dense form. Then, in response to the user's first action on a change item in the difference confirmation panel, the corresponding modified form content is highlighted on the modified form page. Upon completion, the page displays the differences between the modified form content and the original and populated form content, providing users with a summary overview of the differences before final confirmation. This provides sufficient verification for the final generation process, ensuring the accuracy and reliability of the final generated target form content and improving the controllability and credibility of the results in human-computer collaboration scenarios.

[0059] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0060] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0061] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. An interactive form intelligent optimization method, characterized in that, The method includes: Displays the original form page and the modified form page, including the target form; On the modified form page, the filled form content after the fields are filled by the large language model is displayed, and on the original form page, the original form content is displayed; In response to the completion of the large language model filling, a difference confirmation panel is overlaid on the modified form page; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; In response to a first operation on any change item in the difference confirmation panel, the corresponding modification form content is displayed on the modification form page; In response to the completion command of the difference confirmation panel, the modified form page displays the differences between the modified form content and the original form content and the filled form content, respectively; In response to the command to generate a new form, a target form corresponding to the modified form content is generated.

2. The interactive form intelligent optimization method according to claim 1, characterized in that, Before the steps of displaying the filled form content (after field filling is completed by the large language model) on the modified form page and displaying the original form content on the original form page, the method further includes: When the original form page is fully loaded, cache the field identifiers and current values ​​of all input controls on the original form page as a snapshot of the original state; After the large language model completes the field prediction and assignment operations, the assigned form content is rendered to the modified form page, and the current state of the modified form page is compared with the original state snapshot to generate the change item in the differential confirmation panel.

3. The interactive form intelligent optimization method according to claim 1, characterized in that, Each of the change entries in the differential confirmation panel also includes a change confidence indicator, which is presented next to the field name of the corresponding change entry in the form of a color grade or a percentage value; the color grade or percentage value is determined based on the confidence score assigned to the corresponding fill operation by the large language model.

4. The interactive form intelligent optimization method according to claim 1, characterized in that, The step of displaying the corresponding modified form content on the modified form page in response to a first operation on any change item in the differential confirmation panel includes: In response to a first operation on any change entry in the differential confirmation panel, the target field corresponding to the target change entry and its current change value are highlighted on the modification form page in a manner different from the display of other fields.

5. The interactive form intelligent optimization method according to claim 1, characterized in that, Following the step of displaying the corresponding modified form content on the modified form page in response to a first operation on any change item in the differential confirmation panel, the method further includes: In response to the confirmation trigger operation of the target change item in the differential confirmation panel, the status of the target change item is marked as confirmed, and a lock icon is displayed in the corresponding target field in the modification form page; In response to the cancellation trigger operation of the target change entry in the differential confirmation panel, the status of the target change entry is marked as cancelled, and the current value of the corresponding target field in the modification form page is restored to the original value.

6. The interactive form intelligent optimization method according to claim 1, characterized in that, The step of displaying, on the modified form page, the steps of displaying the differences between the modified form content and the original form content, as well as the filled form content, in response to the completion instruction on the difference confirmation panel includes: In response to a trigger operation on the completion control in the difference confirmation panel, a difference summary information bar is generated and displayed in the top or side area of ​​the modified form page; wherein the difference summary information bar includes at least: a first difference statistics item, used to indicate the total number of changed fields in the modified form content relative to the original form content; and a second difference statistics item, used to indicate the total number of unchanged fields in the modified form content relative to the original form content.

7. The interactive form intelligent optimization method according to claim 1, characterized in that, In response to the completion instruction on the difference confirmation panel, after displaying the steps on the modified form page based on the differences between the modified form content and the original form content and the filled form content, the method further includes: A required field completion guide panel is overlaid on the modified form page; wherein, the required field completion guide panel displays the field names of at least one missing required field in the current modified form content in order of business priority; In response to receiving completion information for the currently displayed missing required fields, the system parses the completion information and writes the parsed completion values ​​into the corresponding required fields in the modified form content, thereby obtaining the modified form content.

8. An interactive form intelligent optimization device, characterized in that, The device includes: The first display module is used to display the original form page and the modified form page, including the target form; The second display module is used to display the filled form content after the fields are filled by the large language model on the modified form page, and to display the original form content on the original form page; The first response module is used to display a difference confirmation panel overlaid on the modified form page in response to the completion of the large language model filling; wherein, the difference confirmation panel lists at least one change item, and each change item includes the field name, original value and changed value of the changed field; The second response module is used to respond to the first operation on any change item in the difference confirmation panel and display the corresponding modification form content on the modification form page. The third response module is used to respond to the completion instruction of the difference confirmation panel and display the differences between the modified form content and the original form content and the filled form content on the modified form page. The generation module is used to generate the target form corresponding to the modified form content in response to the new form generation instruction.

9. A computer-readable storage medium, characterized in that, The system contains a computer program that, when executed by a processor, causes the processor to perform the steps of the interactive form intelligent optimization method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, The device includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the interactive form intelligent optimization method as described in any one of claims 1 to 7.

Images