An interface optimization method, device, equipment, medium and program product

Abstract

Embodiments of the present application provide an interface optimization method and device, equipment, medium and program product; the method comprises: obtaining first interaction information between a target user and a to-be-processed interface having a first component; the to-be-processed interface is generated based on an interface generation instruction input by the target user; based on the first interaction information, a first evaluation parameter corresponding to the first component and preference information of the target user for the first component are determined; the first evaluation parameter represents the degree of negation of the target user to the first component; based on the first evaluation parameter, a target storage area corresponding to the preference information is determined from a plurality of storage areas, and the preference information is stored in the target storage area; the storage time of data in different storage areas is different; if the interface generation instruction is received again within the storage time of the preference information, the preference information is obtained from the target storage area, and the first component in the to-be-processed interface is adjusted based on the preference information to obtain a target interface.

Description

Technical Field

[0001] This application relates to the field of computer technology, specifically to an interface optimization method, apparatus, device, medium, and program product. Background Technology

[0002] With the development of technology, generative artificial intelligence has been widely applied in daily life. Currently, generative AI is typically used to generate and render personalized interactive interfaces in real time based on user-input voice or touch commands. However, related technologies generally focus on how to accurately generate the corresponding interactive interface when the user first inputs a command, but lack a mechanism to automatically adjust the generated interactive interface based on user preferences. This means that if a user is not interested in a component of the interactive interface, they can only manually remove it from the interface each time to optimize the generated interactive interface, resulting in poor interface optimization efficiency and thus affecting the user experience. Summary of the Invention

[0003] This application provides an interface optimization method, apparatus, device, medium, and program product, which solves the problem of poor efficiency in interface optimization in related technologies, thereby improving the user experience.

[0004] In a first aspect, embodiments of this application provide an interface optimization method, the method comprising: acquiring first interaction information between a target user and a pending interface having a first component; wherein the pending interface is generated based on an interface generation instruction input by the target user; determining, based on the first interaction information, a first evaluation parameter corresponding to the first component and the target user's preference information for the first component; wherein the first evaluation parameter characterizes the degree of rejection of the first component by the target user; determining, based on the first evaluation parameter, a target storage area corresponding to the preference information from multiple storage areas, and storing the preference information in the target storage area; wherein the storage time of data in different storage areas is different; if the interface generation instruction is received again within the storage time of the preference information, acquiring the preference information from the target storage area, and adjusting the first component in the pending interface based on the preference information to obtain a target interface.

[0005] It is understood that in the interface optimization method provided in this application embodiment, the target user's preference information for the first component in the interface to be processed can be automatically stored. Then, when the interface generation instruction is received again, the first component in the interface to be processed can be automatically adjusted according to the preference information, without requiring the user to manually repeat the operation, thereby improving the efficiency of interface optimization. Furthermore, by determining the degree of the target user's disapproval of the first component, the target storage area corresponding to the preference information can be determined. This allows the system to distinguish whether the preference is short-term, medium-term, or long-term based on the storage duration of the target storage area, thereby determining whether to temporarily adjust the interface to be processed or to make a permanent adjustment. This avoids the problem of the interface being permanently changed due to user error, thereby improving the reliability of interface optimization.

[0006] In some embodiments, determining the first evaluation parameter corresponding to the first component based on the first interaction information includes: determining the first complexity of the first operation corresponding to the first interaction information and the type of the first operation based on the first interaction information; wherein the type characterizes the strength of the target user's intention to correct the first component; determining the first number of times the interface generation instruction is received within a historical time period; if the first number is greater than a first threshold, determining the historical interface correction time closest to the current time, and determining a first time interval based on the historical interface correction time and the current time; wherein the historical interface correction time is a historical time when the interface to be processed was corrected; if the first number is equal to the first threshold, determining the first time interval as a target value; and determining the first evaluation parameter based on the first complexity, the type, the first number, the first time interval, and the first coefficient.

[0007] It is understandable that by combining the first complexity of the first operation, the type of the first operation, the number of times the same behavior is repeated in the same scenario, and the time interval since the last interface correction, the degree of rejection of the first component in the interface to be processed by the target user can be accurately calculated, so as to achieve accurate quantification of the target user's operation behavior, thereby effectively distinguishing between misoperation and real preference, and providing a reliable quantitative basis for the hierarchical storage of subsequent preference information and automatic adjustment of the interface.

[0008] In some embodiments, determining the target storage region corresponding to the preference information from multiple storage regions based on the first evaluation parameter includes: if the first evaluation parameter is less than a second threshold, determining the target storage region as a first storage region among the multiple storage regions; if the first evaluation parameter is greater than or equal to the second threshold and less than a third threshold, determining the target storage region as a second storage region among the multiple storage regions; if the first evaluation parameter is greater than or equal to the third threshold, determining the target storage region from other storage regions based on the first number of times; wherein, the other storage regions are regions among the multiple storage regions other than the first storage region and the second storage region.

[0009] It is understood that in the interface optimization method provided in this application embodiment, preference information is automatically stored in a matching memory area according to the degree of rejection of the first component by the target user. In this way, when the degree of rejection of the first component is high, the component preference is permanently effective; when the degree of rejection is moderate, the component preference is temporarily effective; and when the degree of rejection is low, the component preference is temporarily stored. This allows for the adoption of differentiated adjustment strategies to optimize the interface based on the intensity of the target user's rejection of the first component, thereby improving the accuracy of interface optimization while avoiding permanent changes to the interface due to user misoperation.

[0010] In some embodiments, determining the target storage region from other storage regions based on the first number of times includes: if the first number of times is equal to the first threshold, or if the first number of times is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, determining the target storage region as a third storage region among the other storage regions; if the first number of times is greater than the first threshold and the number is multiple, determining multiple target historical evaluation parameters from multiple historical evaluation parameters based on the generation time of the historical evaluation parameters; if any target historical evaluation parameter is less than the third threshold, determining the target storage region as the third storage region; if each target historical evaluation parameter is greater than or equal to the third threshold, determining the target storage region as a fourth storage region among the other storage regions.

[0011] It is understood that in the interface optimization method provided in this application embodiment, only truly stable and continuous user preferences can be permanently recorded through dual verification of the historical number of times the interface generation command was received and the number of historical evaluation parameters. In this way, it avoids misjudging single occasional behavior as long-term preference and ensures that the user's true preferences can continue to be effective.

[0012] In some embodiments, this application provides an interface optimization method that further includes: acquiring second interaction information between multiple target users and the interface to be processed within a current first period, and determining a second evaluation parameter for the first component by the multiple target users within the current first period based on the second interaction information; wherein the second evaluation parameter characterizes the degree of collective rejection of the first component by the multiple target users; determining a third evaluation parameter for the first component by the multiple target users within a target adjacent period; wherein the target adjacent period is a first period adjacent to the current first period; if the sub-component library corresponding to the interface to be processed is determined to meet the update conditions based on the second evaluation parameter and the third evaluation parameter, determining a component to be replaced from the sub-component library and determining a candidate component from the target component library; if the candidate component meets the target conditions, replacing the component to be replaced with the candidate component.

[0013] It is understood that in the interface optimization method provided in this application embodiment, based on the degree of collective dissatisfaction of multiple target users with the first component within consecutive adjacent periods, it is automatically determined whether there is a persistent collective dissatisfaction with a certain first component. When it is confirmed that the collective dissatisfaction behavior is a long-term behavior rather than a single misoperation, the component replacement process in the sub-component library corresponding to the interface to be processed is automatically triggered. The component replacement is only performed when it is confirmed that the target component to be replaced meets the target conditions. In this way, it can be ensured that the replaced component is better than the original component in real use scenarios, thereby improving the user experience.

[0014] In some embodiments, determining the second evaluation parameters of multiple target users for the first component within the current first period based on the second interaction information includes: obtaining the first moment when the second interaction information is generated, and determining a second time interval based on the first moment and the current moment; determining the second complexity of the second operation corresponding to the second interaction information and the availability parameter of the first component based on the second interaction information; wherein the availability parameter characterizes the availability of the first component; determining the second evaluation parameter based on the second complexity, the availability parameter, the second time interval, and the second coefficient; and determining that the sub-component library meets the update condition if both the second evaluation parameter and the third evaluation parameter are greater than a third threshold.

[0015] It is understood that in the interface optimization method provided in this application embodiment, by combining the availability parameters of a first component under different dimensions, the complexity of user operation components, and the number of days of data coverage, the failure degree of the first component in the current scenario is comprehensively evaluated, thereby ensuring that only components that consistently perform poorly will be replaced, thereby improving the accuracy and reliability of component optimization.

[0016] In some embodiments, determining candidate components from the target component library includes: determining a plurality of candidate components from the target component library based on the functional information of the component to be replaced; determining the historical preference level of the plurality of target users for each candidate component, and determining the degree of adaptation between the plurality of target users and each candidate component; determining the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed; determining a target score corresponding to each candidate component based on the historical preference level, the degree of matching, the degree of adaptation, and a third coefficient; and determining the candidate component from the plurality of candidate components based on the multiple target scores.

[0017] It is understood that in the interface optimization method provided in this application embodiment, by combining the weighted scores of three dimensions—historical preference degree, scene matching degree, and user adaptation degree—scientific sorting and precise selection of components to be selected are achieved. First, ensure that the components are highly compatible with the current scene. Second, refer to the historical performance of the components. Finally, match the style preferences of the user group. In this way, the candidate components that are most suitable for the user group in the current scene are selected, thereby improving the accuracy of component replacement.

[0018] In some embodiments, the method further includes: acquiring third interaction information between multiple users and the interface to be processed within a second period; constructing multiple sample prompts based on the interface generation instruction, the third interaction information, and the satisfaction status of each user with the post-operation interface; wherein the post-operation interface is obtained after the user adjusts the first component during the interaction with the interface to be processed; determining the confidence level corresponding to the multiple sample prompts; wherein the confidence level characterizes the degree of consistency among the multiple sample prompts; if the confidence level is greater than or equal to a fourth threshold, adjusting the interface generation model based on the multiple sample prompts to obtain an adjusted model.

[0019] It is understood that in the interface optimization method provided in this application embodiment, the interface generation model is incrementally fine-tuned by constructing structured sample prompt information containing "interface generation instructions, user correction behavior and user satisfaction status". This enables the model to learn how to generate an interface that better meets user expectations from a large number of users' real correction behaviors, thereby improving the model's ability to continuously self-evolve based on group feedback. This allows the model to accurately generate an interface that meets user expectations and improve the user experience.

[0020] In some embodiments, determining the confidence level corresponding to the plurality of sample prompts includes: determining the data volume of the plurality of sample prompts and the similarity between the plurality of sample prompts; determining the scene coverage corresponding to the target interface content in the interface to be processed; wherein, the target interface content is multiple pieces of content in the interface to be processed that are not of interest to the user; the scene coverage characterizes the breadth of the scene covered by the target interface content; obtaining the second moment when the third interaction information is generated, and determining a third time interval based on the second moment and the current moment; and determining the confidence level based on the data volume, the similarity, the scene coverage, and the third time interval.

[0021] It is understood that in the interface optimization method provided in this application embodiment, by comprehensively and quantitatively evaluating the four dimensions of sample prompt information, namely data volume, similarity, scene coverage and time interval, the confidence level of group consensus among multiple sample prompt information can be determined. Then, when the sample quality meets the standard, incremental fine-tuning of the interface generation model is triggered, thereby avoiding the model being incorrectly optimized due to insufficient samples, disagreement, single scene or outdated data, and thus improving the accuracy of model adjustment.

[0022] In some embodiments, adjusting the interface generation model based on the plurality of sample prompts to obtain an adjusted model includes: processing the plurality of sample prompts to obtain deviation information; determining a target parameter to be adjusted from a plurality of parameters of the interface generation model based on the deviation information; and adjusting the target parameter of the interface generation model based on the deviation information to obtain the adjusted model.

[0023] It is understood that in the interface optimization method provided in the embodiments of this application, by comparing and analyzing the user's correction behavior with the model's generated results, the specific defects of the model in terms of layout, components, and interaction logic are accurately located, and the target parameters in the model corresponding to the defects are adjusted in a targeted manner. In this way, model optimization can be achieved by only fine-tuning local parameters, without adjusting all model parameters. This significantly reduces the consumption of computing resources and training time while ensuring the optimization effect, and achieves efficient and accurate model optimization.

[0024] Secondly, embodiments of this application provide an interface optimization apparatus, comprising: an acquisition unit, configured to acquire first interaction information between a target user and a pending interface having a first component; wherein the pending interface is generated based on an interface generation instruction input by the target user; a first determination unit, configured to determine, based on the first interaction information, a first evaluation parameter corresponding to the first component and the target user's preference information for the first component; wherein the first evaluation parameter characterizes the degree of rejection of the first component by the target user; a second determination unit, configured to determine, based on the first evaluation parameter, a target storage area corresponding to the preference information from multiple storage areas, and store the preference information in the target storage area; wherein the storage time of data in different storage areas is different; and an adjustment unit, configured to, if the interface generation instruction is received again within the storage time of the preference information, acquire the preference information from the target storage area, and adjust the first component in the pending interface based on the preference information to obtain a target interface.

[0025] Thirdly, embodiments of this application provide an interface optimization device, the device comprising: a processor, a memory, and a communication bus; the communication bus is used to establish a communication connection between the processor and the memory; the processor is used to execute an interface optimization program in the memory to implement the steps of the interface optimization method described in the first aspect.

[0026] Fourthly, embodiments of this application provide a computer-readable storage medium storing one or more programs that can be executed by one or more processors to implement the steps of the interface optimization method described in the first aspect.

[0027] Fifthly, embodiments of this application provide a computer program product, which includes a computer program that, when executed by a processor, implements the interface optimization method described in the first aspect. Attached Figure Description

[0028] Figure 1 A schematic diagram of the implementation process of an interface optimization method provided in this application embodiment. Figure 1 ; Figure 2 A schematic diagram of the implementation process of an interface optimization method provided in this application embodiment. Figure 2 ; Figure 3 This is a schematic diagram of the system architecture corresponding to an interface optimization method provided in an embodiment of this application; Figure 4 A schematic diagram illustrating the implementation process of a component optimization method provided in this application embodiment; Figure 5A schematic diagram illustrating the implementation process of a model optimization method provided in this application embodiment; Figure 6 A schematic diagram of the implementation process of an interface optimization method provided in this application embodiment. Figure 3 ; Figure 7 This is a schematic diagram of the structure of an interface optimization device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an interface optimization device provided in an embodiment of this application. Detailed Implementation

[0029] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described below in conjunction with the accompanying drawings. The embodiments described below are only some embodiments of this application, not all embodiments. Therefore, the described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] In the following description, references to “some embodiments” or “other embodiments” describe a subset of all possible embodiments. However, it is understood that “some embodiments” or “other embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0032] In the following description, the terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first" and "second" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0033] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0034] It should be noted that the following embodiments can all be applied to the field of in-vehicle intelligent cockpits.

[0035] This application provides an interface optimization method, which can be applied to an interface optimization device. (Refer to...) Figure 1 As shown, the method may include steps 101 to 104: Step 101: Obtain the first interaction information between the target user and the interface to be processed with the first component.

[0036] The interface to be processed is generated based on the interface generation instructions input by the target user.

[0037] In this embodiment, the interface to be processed is the interface displayed on the vehicle's central control screen at the current moment; the interface generation instruction is used to generate the interface to be processed, and is input by the target user through voice input or manual input to the vehicle control system; the first component refers to the components included in the interface to be processed. There may be one or more first components.

[0038] For example, if the interface generation command is "help me find suitable camping spots nearby", then the interface to be processed will be a campsite recommendation interface, and the first component can include a campsite list component, a weather component, an estimated trip time component, and a campsite map view component, etc.

[0039] It should be noted that the target user can refer to the current user of the vehicle control system.

[0040] In the embodiments of this application, such as Figure 2 As shown, the first interactive information is the feedback information of the target user to the first component, which may include operation events on the interface to be processed, voice commands input by the target user, facial expressions / eye movement information of the target user, and touch trajectories of the target user on the interface to be processed.

[0041] Specifically, such as Figure 2 As shown, the first interaction information includes explicit feedback information, implicit feedback information, and dialogue context information. Explicit feedback information is information actively expressed by the user, which may include direct actions on the interface to be processed (e.g., closing the weather component via voice / manual means, manually dragging the trip time component, etc.); implicit feedback information is implicit actions by the user (e.g., lingering on the campsite list component for an extended period); dialogue context information includes differences between the interface to be processed and the interface after the target user's actions, and the relationships between multiple rounds of dialogue with the target user.

[0042] In this embodiment of the application, the architecture of the vehicle control system can be as follows: Figure 3 As shown, it may include a generative interactive system, a feedback collection and analysis module, and an optimization execution and evaluation module. It should be noted that the first interactive information is obtained through the feedback collection and analysis module.

[0043] Step 102: Based on the first interaction information, determine the first evaluation parameters corresponding to the first component and the target user's preference information for the first component.

[0044] The first evaluation parameter represents the degree of rejection the target user has of the first component.

[0045] In the embodiments of this application, the first evaluation parameter also represents the degree to which the target user expects to adjust the first component. The larger the value of the first evaluation parameter, the higher the degree of rejection of the first component by the target user (i.e., the less interested the target user is in the first component), and the more the target user wants to modify the first component. Conversely, the smaller the value of the first evaluation parameter, the lower the degree of rejection of the first component by the target user (i.e., the more interested the target user is in the first component), and the less the target user wants to modify the first component.

[0046] In one feasible approach, modifying the first component includes resizing the first component, removing the first component, etc.

[0047] It should be noted that if there are multiple first components, each first component corresponds to a first evaluation parameter.

[0048] In this embodiment, the preference information includes information about first components that the target user is interested in and information about first components that they are not interested in. For example, the preference information might be that the target user is interested in the campsite list component and the estimated travel time, but not in the weather component.

[0049] In the embodiments of this application, such as Figure 2 As shown, the collected first interaction information can be preprocessed using a data preprocessing layer. Then, the first evaluation parameters and preference information are determined based on the preprocessed information. It should be noted that, as... Figure 2 As shown, the preprocessing operations include data cleaning, noise reduction and filtering, format standardization, timestamp alignment, and target user ID association.

[0050] In this embodiment of the application, the data preprocessing layer is located at Figure 3 The feedback acquisition and analysis module in the system architecture shown.

[0051] In one feasible approach, the target user's preference information for the first component can be obtained by directly analyzing the preprocessed information.

[0052] Step 103: Determine the target storage area corresponding to the preference information from multiple storage areas based on the first evaluation parameter, and store the preference information in the target storage area.

[0053] The storage time varies in different storage areas.

[0054] In this embodiment, storage validity period can refer to the validity period of the data, such as 1 day, 3 days, or indefinitely. Furthermore, after the storage validity period expires, the preference information will be automatically deleted from the target storage area.

[0055] It should be noted that the larger the value of the first evaluation parameter, the longer the corresponding preference information is stored. This is because a larger first evaluation parameter indicates that the target user is more likely to want to modify the corresponding first component, which will inevitably be stored in a storage area with a longer storage period. In this way, as long as the interface generation instruction is received again within the storage period, the first component in the interface to be processed can be adjusted according to the stored preference information.

[0056] Specifically, the target storage region corresponding to the preference information can be determined from multiple storage regions based on the comparison results between the first evaluation parameter and the target threshold.

[0057] In one feasible approach, multiple storage areas include short-term memory, medium-term memory, temporary memory, and long-term memory. Data in short-term memory has the shortest storage lifespan, followed by medium-term memory, while data in long-term memory has the longest storage lifespan.

[0058] It should be noted that the process of determining the first evaluation parameter and then determining the target storage area corresponding to the preference information based on the first evaluation parameter is achieved through... Figure 2 The feedback analysis unit in the code is implemented, and the feedback analysis unit also belongs to... Figure 3 It is part of the feedback collection and analysis module. In other words, the feedback collection and analysis module can not only collect the feedback data of the target user on the first component, but also analyze the feedback data to determine the degree of the target user's negative attitude towards the first component (i.e., the first evaluation parameter), and determine the target storage area corresponding to the user's preference information based on the first evaluation parameter.

[0059] Step 104: If an interface generation instruction is received again within the storage period of the preference information, retrieve the preference information from the target storage area, and adjust the first component in the interface to be processed based on the preference information to obtain the target interface.

[0060] In this embodiment, the storage validity period of the preference information can be determined first based on the identifier of the target storage area where the preference information is located and the first mapping relationship. Then, it is monitored in real time whether a user has sent another interface generation command. If no interface generation command is received within the storage validity period, the preference information will be automatically deleted. Even if an interface generation command is received again later, the first component in the interface to be processed will not be adjusted. That is, the interface displayed on the central control screen is the same as the original interface to be processed. The first mapping relationship is the mapping relationship between the identifier of the storage area and the storage validity period of the data.

[0061] It should be noted that since the vehicle control system has already generated the interface to be processed according to the interface generation instruction, after receiving the interface generation instruction again, the vehicle control system will not regenerate the corresponding interface, but will only update the content displayed by the components in the interface, such as updating only the content of the weather component, the content of the estimated travel time component, etc.

[0062] Correspondingly, if an interface generation instruction is received again within the storage period of the preference information, it means that the preference information is still valid. In this case, the preference information needs to be retrieved from the target storage area, and the component to be adjusted and the adjustment operation for the component to be adjusted are determined from the first component in the interface to be processed based on the preference information. The component to be adjusted is then adjusted based on the adjustment operation to obtain the target interface.

[0063] In the embodiments of this application, the adjustment operations can be pre-set. For components that the user is interested in, the adjustment operations can include enlarging the display size of the component, moving the component to the first screen area, expanding the detailed content of the component, and raising the display level of the component, etc.; for components that the user is not interested in, the adjustment operations can include shrinking the display size of the component, moving the component to a non-first screen area, collapsing the detailed content of the component, lowering the display level of the component, and removing the component, etc.

[0064] In one feasible approach, one of the corresponding adjustment operations can be randomly selected as the final adjustment operation to be performed based on preference information.

[0065] In another possible approach, the actions performed by the target user on the component to be adjusted can be determined based on the first interaction information, and these actions can be used as the final adjustment actions. For example, if the component to be adjusted is a weather component, and the target user previously removed the weather component, then simply removing the weather component from the interface to be processed will yield the final target interface.

[0066] It should be noted that the operation of adjusting the interface to be processed based on preference information is performed by... Figure 3 The optimization execution and evaluation module performs the adjustments. Furthermore, in addition to adjusting the user interface, the optimization execution and evaluation model can also update components in the corresponding sub-component library and adjust the interface generation strategy (i.e., the interface generation model).

[0067] It is understood that in the interface optimization method provided in this application embodiment, the first step is to obtain the first interaction information between the target user and the component having the first component. The interface to be processed is generated based on the interface generation command input by the target user. Based on the first interaction information, the first evaluation parameter corresponding to the first component and the target user's preference information for the first component are determined. The first evaluation parameter represents the degree of the target user's negative attitude towards the first component. Then, based on the first evaluation parameter, the target storage area corresponding to the preference information is determined from multiple storage areas, and the preference information is stored in the target storage area. The storage time of data in different storage areas is different. If an interface generation command is received again within the storage time of the preference information, the preference information is obtained from the target storage area, and the first component in the interface to be processed is adjusted based on the preference information to obtain the target interface. In this way, the target user's preference information for the first component in the interface to be processed can be automatically stored, and then when the interface generation command is received again, the first component in the interface to be processed can be automatically adjusted according to the preference information without the user having to repeat manual operation, thereby improving the optimization efficiency of the interface. Furthermore, by assessing the degree of rejection of the first component by the target user, the target storage area corresponding to the preference information can be determined. This allows the system to distinguish whether the preference is short-term, medium-term, or long-term based on the storage duration of the target storage area. Consequently, it determines whether to temporarily adjust the interface to be processed or to make permanent adjustments to the interface to be processed, thus avoiding the problem of the interface being permanently changed due to user misoperation and improving the reliability of interface optimization.

[0068] In some embodiments, determining the first evaluation parameter corresponding to the first component based on the first interaction information further includes the following steps 11 to 15: Step 11: Based on the first interaction information, determine the first complexity of the first operation corresponding to the first interaction information, and the type of the first operation.

[0069] Among them, the type represents the strength of the target user's intention to modify the first component.

[0070] In the embodiments of this application, the first operation may refer to the operation performed by the target user on the first component in the interface to be processed during the interaction with the interface to be processed; the first complexity refers to the complexity of the first operation; the type of the first operation represents the degree to which the target user wants to modify the first component.

[0071] Specifically, the first interaction information can be preprocessed, and the number of operation steps of the first operation of the target user for each first component can be determined based on the preprocessed information. The maximum number of operation steps can be determined from multiple operation steps. Then, each operation step is divided by the maximum number of operation steps, that is, each operation step is normalized to obtain the first complexity of the first operation of the target user for each first component.

[0072] For example, if the target user first clicks to expand the weather widget and then removes it 1 second later, then the target user's first action on the weather widget would be 2 steps.

[0073] It should be noted that each first operation corresponds to a first complexity. In one possible implementation, if the target user has not performed an operation on a certain first component, then the first operation is ignored, its operation steps are 0, and its complexity is naturally 0.

[0074] In this application embodiment, the first operation includes four types: the first type is voice negation (i.e., the target user directly removes / deletes the first component via voice), the second type is manual removal / deletion (i.e., the target user removes / deletes the first component manually), the third type is quick close (i.e., the target user expands the first component and then quickly closes it), and the fourth type is ignore (i.e., the target user does not actually perform any operation on the first component).

[0075] Specifically, the preprocessed information can be analyzed to determine the type of operation performed by the target user for the first operation of each first component.

[0076] Step 12: Determine the first number of times the interface generation command was received within the historical time period.

[0077] In this embodiment of the application, it can be determined whether an interface generation instruction has been received in the historical period. If it has been received, the initial number of times the interface generation instruction has been repeatedly received in the historical period can be directly obtained, and the initial number can be normalized to obtain the first count. If no interface generation instruction has been received, the first count is determined to be 0.

[0078] It should be noted that the first count is the number of times the scene corresponding to the interface generation instruction (e.g., the campsite recommendation interface generation instruction) is repeated (e.g., the destination recommendation scene).

[0079] The historical period can be a pre-set time period.

[0080] In the embodiments of this application, step 203 or step 204 can be executed after step 202.

[0081] Step 13: If the first count is greater than the first threshold, determine the historical interface correction time closest to the current time, and determine the first time interval based on the historical interface correction time and the current time.

[0082] Among them, the historical interface correction moment is the historical moment when the interface to be processed was corrected.

[0083] In this embodiment of the application, the first threshold is 0. That is, if the first number is greater than 0 (indicating that an interface generation instruction was received within the historical period), the historical time when the target user corrected the interface to be processed within the historical period can be obtained. If there is only one historical time, then the historical time is the historical interface correction time, and the first time interval between the historical interface correction time and the current time can be directly determined. If there are multiple historical times, then the historical time closest to the current time is determined from the multiple historical times, which is the historical interface correction time, and then the first time interval between the historical interface correction time and the current time is determined.

[0084] In one feasible approach, the time interval can be measured in days.

[0085] For example, if there is only one historical moment, and the historical moment is 14:00 on May 9th, and the current moment is 15:30 on May 12th, then the first time interval between the two is 3 days.

[0086] Step 14: If the first count equals the first threshold, determine the first time interval as the target value.

[0087] In this embodiment, the target value is 0. Specifically, if the first number is equal to 0, it means that no interface generation instruction was received within the historical time period. In this case, the first time interval can be directly determined to be 0.

[0088] Step 15: Determine the first evaluation parameter based on the first complexity, type, first number, first time interval, and first coefficient.

[0089] In this embodiment, the first coefficient includes a first sub-coefficient, a second sub-coefficient, a third sub-coefficient, and a fourth sub-coefficient, and the sum of the first sub-coefficient, the second sub-coefficient, the third sub-coefficient, and the fourth sub-coefficient is 1. Specifically, the weight value of the type of the first operation can be determined firstly based on the type of the first operation and the second mapping relationship. Then, the weight value can be multiplied by the first sub-coefficient to obtain a first value, the first complexity can be multiplied by the second sub-coefficient to obtain a second value, and the first value can be multiplied by the third sub-coefficient to obtain a third value. Then, an exponential decay operation can be performed on the first time interval, and the obtained operation value can be multiplied by the fourth sub-coefficient to obtain a fourth value. Finally, the first value, the second value, the third value, and the fourth value can be summed to obtain the first evaluation parameter.

[0090] The second mapping relationship is the correspondence between the type of the first operation and multiple weight values. In one possible implementation, the weight value of the first type is 1, the weight value of the second type is 0.8, the weight value of the third type is 0.5, and the weight value of the fourth type is 0.2.

[0091] In one feasible approach, the calculation formula for the first evaluation parameter can be shown in the following formula (1).

[0092] Formula (1) Wherein, CSI represents the first evaluation parameter; This represents the first sub-coefficient (usually 0.5); F represents the weight value of the type of the first operation; The second sub-coefficient is represented (usually taken as 0.2); S represents the first complexity. The third sub-coefficient (usually 0.2) represents the first number (i.e., the number of times the scene is repeated). t1 represents the fourth sub-coefficient (usually taken as 0.1); t1 represents the first time interval; k represents the attenuation coefficient; This indicates exponential decay.

[0093] For example, if F is 0.5, S is 0.2, C is 0 (i.e., the number of scene repetitions is 0), and T is 1, then CSI = 0.5 × 0.5 + 0.2 × 0.2 + 0.2 × 0 + 0.1 × 1 = 0.25 + 0.04 + 0 + 0.1 = 0.39.

[0094] It should be noted that regardless of whether the target user interacts with each first component during the interaction with the interface to be processed, each first component corresponds to a first evaluation parameter. In other words, as long as an interface generation instruction has been received, the corresponding interface will be generated, and each component in the interface will correspond to a first evaluation parameter.

[0095] It is understood that in the interface optimization method provided in the embodiments of this application, by combining the first complexity of the first operation, the type of the first operation, the number of times the same behavior is repeated in the same scenario, and the time interval since the last interface correction, the degree of rejection of the first component in the interface to be processed by the target user is accurately calculated, so as to achieve accurate quantification of the target user's operation behavior, thereby effectively distinguishing between misoperation and real preference, and providing a reliable quantitative basis for the subsequent hierarchical storage of preference information and automatic adjustment of the interface.

[0096] In some embodiments, determining the target storage region corresponding to the preference information from multiple storage regions based on the first evaluation parameter further includes the following steps 21 to 23: Step 21: If the first evaluation parameter is less than the second threshold, the target storage area is determined to be the first storage area among multiple storage areas.

[0097] In this embodiment, the second threshold is an empirically set value, typically 0.3. If the first evaluation parameter is less than the second threshold, it indicates that the target user has a low degree of negative attitude towards a certain first component, meaning that the target user is still quite interested in the first component. Correspondingly, the degree to which the target user wants to modify the first component is also relatively low. In this case, it may be due to a mistake by the target user. In this case, the corresponding preference information for the first component can be stored in the storage area with the shortest storage time. That is, the first storage area with the shortest storage time is directly determined from multiple storage areas, i.e., the short-term memory area is the target storage area corresponding to the preference information.

[0098] In one feasible approach, the data in the first storage area is stored for 1 day, or 24 hours.

[0099] Step 22: If the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, the target storage area is determined to be the second storage area among multiple storage areas.

[0100] In this embodiment, the third threshold is also an empirically set value, typically 0.7. If the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, i.e., 0.3 ≤ first evaluation parameter CSI ≤ 0.7, it indicates that the target user's negative attitude towards a certain first component is moderate. This means that the target user is not very interested in the first component, but has not yet reached the level of aversion. In this case, the corresponding preference information for the first component can be stored in a storage area with a moderate storage time. That is, the second storage area with a moderate storage time is directly determined from multiple storage areas, i.e., the medium-term memory area is the target storage area corresponding to the preference information. In this way, the preference information can be temporarily retained in the medium-term memory area for subsequent short-term interface adjustments, without being quickly forgotten due to a short time limit, or occupying unnecessary storage resources due to a long time limit.

[0101] In one feasible approach, the data in the second storage area is stored for 7 days.

[0102] It should be noted that if the second storage area is determined to be the target storage area, and an interface generation instruction is received again within 7 days, the first component in the interface to be processed needs to be adjusted according to the preference information to obtain the target interface. However, if an interface generation instruction is received again after 7 days, the first component in the interface to be processed does not need to be adjusted because the preference information has expired (i.e., it has been deleted). In this case, the interface finally displayed on the central control screen is the same as the interface to be processed.

[0103] Step 23: If the first evaluation parameter is greater than or equal to the third threshold, determine the target storage area from other storage areas based on the first number.

[0104] Other storage areas are those areas other than the first and second storage areas among the multiple storage areas.

[0105] In this embodiment, if the first evaluation parameter is greater than or equal to the third threshold, i.e., the first evaluation parameter CSI ≥ 0.7, it indicates that the target user has a high degree of disapproval of a certain first component, that is, the target user is completely uninterested in the first component, and correspondingly, the degree to which the target user wants to modify the first component is very high. At this time, it can be determined whether the target user has received an interface generation instruction in the past period of time based on the first count, and then it can be determined based on the determination result that the target user's degree of disapproval of the first component is continuous rather than temporary, and thus the target storage area corresponding to the preference information can be determined from other storage areas based on the result.

[0106] It should be noted that the storage time of data in other storage areas is longer than that of data in the second storage area.

[0107] It is understood that in the interface optimization method provided in this application embodiment, if the first evaluation parameter is less than the second threshold, the target storage area is determined to be the first storage area among multiple storage areas; if the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, the target storage area is determined to be the second storage area among multiple storage areas; if the first evaluation parameter is greater than or equal to the third threshold, the target storage area is determined from other storage areas based on the first evaluation parameter and the first threshold. Thus, according to the degree of rejection of the first component by the target user, preference information is automatically stored in matching memory areas in a hierarchical manner. This allows for permanent effect of component preferences when the degree of rejection of the first component is high, temporary effect of component preferences when the degree of rejection is moderate, and temporary storage of component preferences when the degree of rejection is low. This enables the implementation of differentiated adjustment strategies to optimize the interface based on the intensity of the target user's rejection of the first component, thereby improving the accuracy of interface optimization while preventing users' accidental operations from permanently changing the interface.

[0108] In some embodiments, determining the target storage region from other storage regions based on the first data further includes the following steps 31 to 34: Step 31: If the first count is equal to the first threshold, or if the first count is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, determine the target storage area as the third storage area among other storage areas.

[0109] In this application embodiment, if the first number is equal to the first threshold (i.e., 0), it means that the target user has not received the interface generation instruction in the historical period. In this case, the target user's negative behavior towards the first component is likely to be a single, occasional behavior. At this time, it is necessary to store it in the temporary memory area (i.e., the third storage area).

[0110] In this embodiment, the historical evaluation parameter represents the degree of negative feedback from the target user towards the first component within a historical time period. If the first number is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, it indicates that the target user has only received the interface generation instruction once within the historical time period. In this case, the first evaluation parameter is the target user's second evaluation of the first component. Even if the target user previously held a negative attitude towards the first component, this does not necessarily mean that the negative behavior is continuous rather than occasional. Therefore, regardless of how high the current degree of negative feedback the target user has towards the first component, the preference information should be stored in a temporary memory area first to observe whether the negative behavior will continue to occur in the future, so as to avoid misjudging the target user's single accidental operation as the user's true preference and permanently changing the interface.

[0111] In one feasible approach, the data in the third storage area is stored for 15 days.

[0112] Step 32: If the first number is greater than the first threshold and there are multiple such numbers, determine multiple target historical evaluation parameters from the multiple historical evaluation parameters based on the generation time of the historical evaluation parameters.

[0113] In this embodiment, if the first count is greater than 0 and there are multiple historical evaluation parameters, it indicates that the target interface receiving instruction has been received multiple times (at least twice) within the historical time period. In this case, the generation time of each historical evaluation parameter can be obtained, and the two historical evaluation parameters closest to the current time can be selected from the multiple historical evaluation parameters based on the generation time. These two historical evaluation parameters are then determined as the target historical evaluation parameters. That is, there are two target historical evaluation parameters.

[0114] In the embodiments of this application, step 33 or step 34 can be executed after step 32.

[0115] Step 33: If any target's historical evaluation parameter is less than the third threshold, determine the target storage area as the third storage area.

[0116] In this embodiment of the application, if either of the two target historical evaluation parameters is less than the third threshold, it indicates that the target user's negative attitude towards a certain first component is not that high. In other words, the target user's negative behavior towards the first component is not continuous. The target user may only be uninterested in the first component for a period of time, and may become interested in the first component again after a period of time. That is, the target user's negative behavior towards the first component is an occasional event. In this case, the corresponding preference information can be stored in the temporary storage area (i.e., the third storage area) to adjust the first component in the interface to be processed in the short term.

[0117] It should be noted that if the target user's evaluation parameters for the first component are all less than the third threshold for three consecutive times, the preference information stored in the third storage area needs to be deleted and the preference information stored in the fourth storage area.

[0118] Step 34: If the historical evaluation parameters of each target are greater than or equal to the third threshold, the target storage area is determined to be the fourth storage area among the other storage areas.

[0119] In this embodiment, if both target historical evaluation parameters are greater than or equal to the third threshold, it indicates that the target user's negative behavior toward a certain first component is continuous. This means that the target user is truly and completely uninterested in the first component, and it is not affected by time. In other words, the target user's negative behavior toward the first component is not an occasional occurrence. In this case, it is necessary to determine the fourth storage area (long-term memory area) as the target storage area and store the corresponding preference information in the long-term memory area. In this way, the permanent change of the interface to be processed can be achieved to improve the user experience.

[0120] It should be noted that the data in the fourth storage area is stored permanently. Furthermore, after storing preference information in the permanent memory area, the preference information stored in the temporary memory area must be deleted.

[0121] It is understood that in the interface optimization method provided in this application embodiment, if the first count is equal to the first threshold, or if the first count is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, the target storage area is determined to be the third storage area among other storage areas; if the first count is greater than the first threshold and the number is multiple, multiple target historical evaluation parameters are determined from the multiple historical evaluation parameters based on the generation time of the historical evaluation parameters; if any target historical evaluation parameter is less than the third threshold, the target storage area is determined to be the third storage area; if each target historical evaluation parameter is greater than or equal to the third threshold, the target storage area is determined to be the fourth storage area among other storage areas. In this way, through the dual verification of the historical reception count of the interface generation command and the number of historical evaluation parameters, it can be ensured that only truly stable and continuous user preferences can be permanently recorded. This avoids misjudging single occasional behaviors as long-term preferences and ensures that the user's true preferences can continue to take effect.

[0122] In some embodiments, in addition to adaptive modifications to the interface based on the specific user's personal preferences (i.e. Figure 3 (In addition to the individual layer optimization), it can also permanently replace components in the sub-component library corresponding to the interface being processed, thereby achieving... Figure 3 Component layer optimization in, specifically, such as Figure 4 As shown, the specific process of component layer optimization may include the following steps 201 to 204: Step 201: Obtain the second interaction information between multiple target users and the interface to be processed within the current first period, and determine the second evaluation parameters of multiple target users for the first component within the current first period based on the second interaction information.

[0123] The second evaluation parameter represents the degree of collective rejection of the first component by multiple target users.

[0124] In this embodiment, the first period is a pre-set period. In one possible implementation, the first period is 7 days.

[0125] In this embodiment of the application, the second interactive information and the first interactive information are obtained in the same way, that is, both are obtained through... Figure 3 The feedback collection and analysis module in the middle obtains the second interaction information, and the second interaction information also includes explicit feedback information, implicit feedback information and dialogue context related information.

[0126] It should be noted that the second evaluation parameter is the degree of disapproval of the first component by multiple target users. A high degree of disapproval indicates that multiple target users are not interested in the first component. In other words, the second evaluation parameter is a group indicator that targets a group, while the first evaluation parameter is an individual indicator that targets a single user.

[0127] It should be noted that the second interactive information can also be obtained through... Figure 3 The feedback collection and analysis module collects the data. And it also needs to go through... Figure 3 The data preprocessing layer performs preprocessing operations.

[0128] Step 202: Determine the third evaluation parameters for the first component for multiple target users within adjacent target periods.

[0129] The target adjacent period is the first period adjacent to the current first period.

[0130] In this embodiment, the third evaluation parameter is also a group indicator; the target adjacent period is the previous first period before the current first period (i.e., the first adjacent period), or the next first period after the current first period (i.e., the second adjacent period). Specifically, the first number of times the interface generation command has been received within the historical period can be determined, and the target adjacent period can be determined as the first adjacent period or the second adjacent period based on the first number, and then the third evaluation parameter can be calculated based on the determination result.

[0131] Step 203: If the sub-component library corresponding to the interface to be processed is determined to meet the update conditions based on the second evaluation parameter and the third evaluation parameter, the component to be replaced is determined from the sub-component library, and the candidate component is determined from the target component library.

[0132] In this embodiment of the application, the sub-component library includes all components required to generate the interface to be processed; the target component library includes all components required to generate all interfaces, that is: the sub-component library is a part of the target component library, the target component library is the whole, and the sub-component library is a part of it; the sub-component library satisfying the update condition means that the first component in the sub-component library satisfies the update condition.

[0133] It should be noted that the sub-component library corresponding to the interface to be processed is the same as the sub-component library corresponding to the target interface. In other words, when performing personal layer optimization, the adjustment of the interface to be processed directly adjusts the components in the interface to be processed, but the corresponding sub-component library remains unchanged. However, component layer optimization adjusts from the root of the component. After the adjustment, if an interface generation instruction is received again, the components contained in the generated interface will be different from the original.

[0134] Specifically, you can determine whether the sub-component library meets the update conditions through the following step A1.

[0135] Step A1: If both the second evaluation parameter and the third evaluation parameter are greater than the third threshold, it is determined that the sub-component library meets the update conditions.

[0136] In this embodiment, the third threshold is 0.7. Specifically, if both the second and third evaluation parameters of a certain first component are greater than 0.7, it indicates that the user group, including multiple target users, has a high degree of rejection towards the first component in two adjacent first periods. That is, the user group is not interested in the first component in two adjacent first periods. This means that the user group's rejection of the first component is not an occasional behavior, but a long-term continuous behavior. At this time, it can be determined that the sub-component library meets the update conditions, that is, the corresponding components in the sub-component library need to be updated so that the interface generated based on the components in the sub-component library is more in line with user preferences, thereby improving the user experience.

[0137] It should be noted that as long as the second and third evaluation parameters of any first component in the sub-component library are both greater than the third threshold, the sub-component library meets the update condition, but it is not necessary for the second and third evaluation parameters of all first components to be greater than the third threshold.

[0138] In this embodiment of the application, if it is determined that the sub-component library meets the update conditions, then all first components whose second evaluation parameter and third evaluation parameter are both greater than the third threshold can be selected from multiple first components in the sub-component library, and all of the above first components can be determined as components to be replaced.

[0139] In this embodiment, the candidate component is used to replace the component to be replaced. It should be noted that the candidate component and the component to be replaced have the same function.

[0140] It should be noted that the process of determining the third evaluation parameter and whether the sub-component library meets the update conditions is both done through... Figure 2 The feedback analysis unit is implemented in the feedback acquisition and analysis module.

[0141] Step 204: If the candidate component meets the target conditions, replace the component to be replaced with the candidate component.

[0142] In the embodiments of this application, one component to be replaced corresponds to one candidate component, that is, there are as many candidate components as there are components to be replaced.

[0143] In this embodiment, after identifying the candidate component, 10% of the target users are randomly selected as test users from multiple target users. Then, a new test interface is generated based on the candidate component and other components in the sub-component library and pushed to the aforementioned test users to conduct a 3-day gray-scale A / B test. After the test is completed, the t-test method is used to analyze the test data of the candidate component during the test period (i.e., 3 days) to obtain the difference value (i.e., p-value). Then, the interaction information between multiple test users and the test interface during the test period can be obtained, and the reference evaluation parameters of multiple test users for the candidate component during the test period can be determined based on the interaction information. Further, if the p-value is greater than or equal to 0.95 and the reference evaluation parameter is less than or equal to the third threshold, it is determined that the candidate component meets the target conditions, that is, the effect of the candidate component has reached the expectation; if the p-value is less than 0.95 or the reference evaluation parameter is greater than the third threshold, it is determined that the candidate component does not meet the target conditions, that is, the effect of the candidate component has not reached the expectation.

[0144] Among them, other components are the first components in the sub-component library that are not the components to be replaced.

[0145] In this embodiment of the application, after determining that the candidate component meets the target condition, that is, the effect of the candidate component reaches the expected effect, the component to be replaced in the sub-component library can be replaced with the candidate component. Then, if the interface generation instruction for the same scenario is received again, a new interface including the candidate component and other components will be automatically generated.

[0146] It should be noted that after replacing the component to be replaced, the component can be marked to indicate that it is not suitable for the corresponding scenario. For example, the marking information could be "Destination recommended, but not recommended for the scenario".

[0147] In this embodiment of the application, after determining that the sub-component library meets the update conditions, it can then send... Figure 3 The optimization execution and evaluation module sends component optimization instructions to trigger... Figure 2 The component library optimization process in the system involves determining whether candidate components meet the target conditions through... Figure 3 This is implemented through the optimization execution and evaluation module.

[0148] In this embodiment of the application, after replacing the component to be replaced with the candidate component, if an interface generation instruction corresponding to the interface to be processed is received again, then a new interface needs to be generated again based on the instruction and the components in the updated sub-component library corresponding to the interface to be processed.

[0149] It is understood that in the interface optimization method provided in this application embodiment, the second interaction information between multiple target users and the interface to be processed within the current first period is first obtained, and based on the second interaction information, the second evaluation parameters of multiple target users for the first component within the current first period are determined; then, the third evaluation parameters of multiple target users for the first component within the target adjacent period are determined; then, based on the second evaluation parameters and the third evaluation parameters, it is determined that the sub-component library corresponding to the interface to be processed meets the update conditions, the component to be replaced is determined from the sub-component library, and the candidate component is determined from the target component library; if the candidate component meets the target conditions, the component to be replaced is replaced with the candidate component. In this way, based on the degree of collective rejection of the first component by multiple target users within consecutive adjacent periods, it is automatically determined whether there is a continuous collective dissatisfaction with a certain first component, and when it is confirmed that the collective dissatisfaction behavior is a long-term behavior rather than a single misoperation, the component replacement process in the sub-component library corresponding to the interface to be processed is automatically triggered, and the actual effect of the candidate component is verified through gray-scale A / B testing to ensure that the replaced component is indeed better than the original component in real use scenarios, thus realizing an automatic optimization mechanism of group-driven, periodic verification, and gray-scale confirmation at the component level.

[0150] In some embodiments, determining the second evaluation parameters of multiple target users for the first component within the current first period based on the second interaction information further includes the following steps 41 to 43: Step 41: Obtain the first moment when the second interaction information is generated, and determine the second time interval based on the first moment and the current moment.

[0151] In this embodiment of the application, the first moment when the second interaction information between each target user and the interface to be processed is generated can be obtained. That is, one target user corresponds to one first moment. Then, the time interval between each first moment and the current moment can be determined, that is, the second time interval.

[0152] It should be noted that each first component corresponds to a second time interval.

[0153] Step 42: Based on the second interaction information, determine the second complexity of the second operation corresponding to the second interaction information, and the availability parameters of the first component.

[0154] The availability parameter characterizes the degree of availability of the first component.

[0155] In the embodiments of this application, the second operation may refer to the operation performed by multiple target users on the first component of the interface to be processed during their interaction with the interface to be processed; the second complexity refers to the average complexity of multiple second operations on a certain first component.

[0156] Specifically, the second interaction information can be preprocessed, and the number of operation steps for the second operation of multiple target users on each first component can be determined based on the preprocessed information. The maximum number of operation steps is then determined from among these multiple operation steps. Next, each operation step is divided by the maximum number of operation steps, i.e., each operation step is normalized, to obtain the initial complexity of the second operation of multiple target users on each first component. Then, the average of the multiple initial complexities corresponding to the same first component can be calculated to obtain the average complexity of the second operation on that first component (i.e., the second complexity). In other words, one first component corresponds to one second complexity.

[0157] In this application embodiment, the availability parameters include the closure rate of the first component (i.e., the proportion of target users who close the first component), the misoperation rate of the first component (i.e., the proportion of target users who operate on the first component and then revert to the previous operation), and the negative voice feedback rate of the first component (i.e., the proportion of target users who express negative evaluations of the first component via voice).

[0158] It should be noted that the second interaction information can be directly analyzed to obtain three parameters: closure rate, misoperation rate, and negative voice feedback rate.

[0159] Step 43: Determine the second evaluation parameter based on the second complexity, availability parameter, first time step, and second coefficient.

[0160] In this embodiment, the second coefficient includes a fifth sub-coefficient, a sixth sub-coefficient, a seventh sub-coefficient, and an eighth sub-coefficient, and the sum of the fifth sub-coefficient (0.4), the sixth sub-coefficient (0.3), the seventh sub-coefficient (0.2), and the eighth sub-coefficient (0.1) is 1. Specifically, the closing rate of the first component can be multiplied by the fifth sub-coefficient to obtain a fifth value, the error rate of the first component can be multiplied by the sixth sub-coefficient to obtain a sixth value, the negative voice feedback rate of the first component can be multiplied by the seventh sub-coefficient to obtain a seventh value, and an exponential decay operation can be performed on the second time interval. The obtained calculated value can then be multiplied by the eighth sub-coefficient to obtain a fourth value. Finally, the fifth value, the sixth value, the seventh value, and the eighth value can be summed to obtain the second evaluation parameter.

[0161] It should be noted that one first component corresponds to one second evaluation parameter.

[0162] In one feasible approach, the formula for calculating the second evaluation parameter can be shown in formula (2) below.

[0163] Formula (2) Wherein, CFI represents the second evaluation parameter corresponding to a certain first component; The i-th parameter represents the closing rate of the first component, the misoperation rate of the first component, the negative voice feedback rate of the first component, and the second complexity, respectively. t1 represents the second coefficient (0.4, 0.3, 0.2, and 0.1 respectively); t2 represents the second time interval; k represents the attenuation coefficient (usually 0.05). This indicates exponential decay.

[0164] For example, if the closing rate of a certain first component is 0.82, the misoperation rate is 0.75, the negative voice feedback rate is 0.6, the second complexity is 0.5, and t2 is 0, then CFI = 0.4 × 0.82 × 1 + 0.3 × 0.75 × 1 + 0.2 × 0.6 × 1 + 0.1 × 0.5 × 1 = 0.723.

[0165] It is understood that in the interface optimization method provided in this application embodiment, the first moment when the second interaction information is generated is obtained, and a second time interval is determined based on the first moment and the current moment; then, based on the second interaction information, the second complexity of the second operation corresponding to the second interaction information and the availability parameter of the first component are determined; the availability parameter characterizes the availability of the first component; and then, based on the second complexity, availability parameter, second time interval, and second coefficient, the failure degree of the first component in the current scenario is comprehensively evaluated by combining the availability parameter of a certain first component under different dimensions, the complexity of the user operation component, and the number of days of data coverage, thereby ensuring that only components that consistently perform poorly are replaced, thus improving the accuracy and reliability of component optimization.

[0166] In some embodiments, determining candidate components from the target component library further includes the following steps 51 to 55: Step 51: Determine multiple candidate components from the target component library based on the functional information of the component to be replaced.

[0167] In this embodiment of the application, all components with the same function as the component to be replaced (i.e., matching function information) can be selected from the target component library, and these components can be identified as candidate components.

[0168] It should be noted that the components to be selected and the components to be replaced have the same function, only their styles are different.

[0169] Step 52: Determine the historical preference levels of multiple target users for each candidate component, and determine the fit between multiple target users and each candidate component.

[0170] In this embodiment of the application, the historical preference level represents the satisfaction level of multiple target users with the selected components over a period of time.

[0171] Specifically, the historical operation information of multiple target users regarding the candidate components is first obtained, and the historical closure rate of the candidate components is determined based on the historical operation information. Then, the historical preference level can be calculated as 1 - historical closure rate. In other words, the lower the historical closure rate, the higher the historical preference level, meaning that multiple target users are more satisfied with the candidate component. It should be noted that each candidate component corresponds to one historical preference level.

[0172] In the embodiments of this application, the degree of fit characterizes the matching between the features of the candidate component and the user group's preference features for the component, which in fact characterizes whether the candidate component is suitable for this user group.

[0173] Specifically, the attribute information (age, geographical location, etc.) of multiple target users and the interaction information between multiple target users and the interactive interfaces of various scenarios displayed on the central control display are analyzed to obtain the preference characteristics of this user group for components (e.g., support for voice interaction, etc.). Then, the similarity between the features of each candidate component and the preference features is calculated, which is the degree of fit between the user group and each candidate component.

[0174] It should be noted that the higher the similarity, the higher the degree of adaptation, and the more suitable the candidate component is for this user group.

[0175] Step 53: Determine the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed.

[0176] In this embodiment, the matching degree characterizes whether the function of the candidate component matches the scenario corresponding to the interface to be processed. For example, if the scenario corresponding to the interface to be processed is a campsite recommendation scenario, then the component displaying the clothing store in the vicinity will necessarily not match the campsite recommendation scenario.

[0177] Specifically, the scene information can be segmented to obtain multiple first words, and the function information can be segmented to obtain multiple second words. Then, the similarity between each first word and each second word can be calculated, and all similarities can be summed to obtain the matching degree.

[0178] Step 54: Based on historical preference level, matching level, fit level and third coefficient, determine the target score corresponding to each candidate component.

[0179] In this embodiment, the third coefficient may include a ninth sub-coefficient (0.3), a tenth sub-coefficient (0.6), and an eleventh sub-coefficient (0.1). Specifically, the historical preference level can be multiplied by the ninth sub-coefficient to obtain the ninth value, the matching level can be multiplied by the tenth sub-coefficient to obtain the tenth value, and the fit level can be multiplied by the eleventh sub-coefficient to obtain the eleventh value. Then, the ninth, tenth, and eleventh values ​​are summed to obtain the target score corresponding to each candidate component.

[0180] Step 55: Determine candidate components from multiple candidate components based on multiple target scores.

[0181] In this embodiment of the application, multiple first candidate components can be sorted in descending order of target score, and the candidate component with the highest target score (i.e. the highest target score) can be determined from the sorted multiple first candidate components as the candidate component.

[0182] It is understood that in the interface optimization method provided in this application embodiment, firstly, multiple candidate components are determined from the target component library based on the functional information of the component to be replaced; then, the historical preference of multiple target users for each candidate component is determined, as well as the degree of adaptation between multiple target users and each candidate component is determined; then, the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed is determined; then, based on the historical preference, matching degree, adaptation degree, and a third coefficient, the target score corresponding to each candidate component is determined; finally, candidate components are determined from multiple candidate components based on multiple target scores. In this way, by combining the weighted scores of three dimensions—historical preference (30%), scene matching degree (60%), and user adaptation degree (10%)—scientific sorting and precise selection of candidate components are achieved—prioritizing ensuring that the component is highly compatible with the current scene, secondly referring to the historical performance of the component, and finally matching the style preferences of the user group, thereby filtering out the candidate components most suitable for the user group in the current scene, thus improving the accuracy of component replacement.

[0183] In some embodiments, in addition to adaptive modifications to the interface based on the specific user's personal preferences (i.e. Figure 2 (Individual layer optimization), and replacing the relevant components in the corresponding sub-component layer of the interface (i.e. Figure 2 Component layer optimization can also optimize the interface generation model corresponding to the interface to be processed, thereby achieving... Figure 2 Optimization of the strategy layer in the middle, specifically, such as Figure 5 As shown, the specific process of strategy layer optimization may include the following steps 301 to 304: Step 301: Obtain the third interaction information between multiple users and the interface to be processed within the second cycle.

[0184] In this embodiment, the second period is a preset period, and the duration of the second period is longer than that of the first period. In one possible implementation, the second period is one month.

[0185] In this embodiment of the application, the multiple users are all users using the vehicle control system, while the multiple target users mentioned above are only a subset of these users.

[0186] It should be noted that the third interactive information is obtained in the same way as the first and second interactive information, that is, through... Figure 3 The feedback collection and analysis module collects the data. And it also needs to go through... Figure 3 The data preprocessing layer performs preprocessing operations.

[0187] Step 302: Based on the interface generation instructions, the fourth interaction information, and each user's satisfaction with the interface after the operation, construct multiple sample prompt messages.

[0188] The post-operation interface is obtained by adjusting the first component during the user's interaction with the interface to be processed.

[0189] In this embodiment of the application, the post-operation interface is the interface that the user is satisfied with, which is usually the interface that the user is satisfied with after making real-time adjustments to the interface to be processed; the user's satisfaction state refers to the state in which the user is willing to take the next action.

[0190] Specifically, such as Figure 6 As shown, the user's negative operation on the first component of the interface to be processed is first determined based on the third interaction information (e.g., closing the weather component, opening the estimated travel time component and then closing it after 1 second). The interface generation command, the third interaction information, and the interface content are then combined according to a preset format to obtain the initial sample prompt information. It should be noted that each user interaction with the interface to be processed generates third interaction information; therefore, one target user corresponds to one initial sample prompt information. Subsequently, multiple initial sample prompt information can be filtered for quality to obtain several high-quality sample prompt information.

[0191] For example, the sample prompt message is: [Interface generation instruction: nearby suitable places for weekend camping] -> [User's negative action: the weather card is expanded and then closed after 1 second, with no interaction] -> [User's satisfaction status: navigation is initiated after a deep browsing of the first campsite details page].

[0192] Step 303: Determine the confidence level corresponding to multiple sample prompts.

[0193] Confidence level represents the degree of consistency among the prompts from multiple samples.

[0194] In the embodiments of this application, the higher the consistency, the higher the confidence level, which means that there is the same tendency to make corrections among multiple users. That is, all users are dissatisfied with the same part of the interface to be processed (which may be a specific component or the overall layout of the component) in the same scenario. For example, if all users turn off the weather component in the campsite recommendation scenario, it means that all users are not interested in the weather component.

[0195] Specifically, the data volume of multiple sample prompts, the similarity between multiple sample prompts, and the scene coverage corresponding to the first component are determined. Then, the third time interval is determined based on the second moment generated by the third interaction information and the current moment. Finally, the confidence level is determined based on the data volume, similarity, scene coverage, and third time interval.

[0196] It should be noted that the process of determining the confidence level is through... Figure 2 The feedback analysis unit is implemented in the feedback acquisition and analysis module.

[0197] Step 304: If the confidence level is greater than or equal to the fourth threshold, adjust the interface generation model based on multiple sample prompts to obtain the adjusted model.

[0198] In this embodiment, the fourth threshold is 0.8. If the confidence level is greater than or equal to 0.8, it indicates that all users have a high tendency to modify a certain first component / component layout, etc. Since all users are involved, the interface generation model used to generate the interface according to user instructions can be fine-tuned directly to fundamentally solve the problem of all users being dissatisfied with the interface.

[0199] It should be noted that the interface to be processed is generated by the interface generation model. In other words, the interface generation model generates the interface based on user instructions. After the interface is generated, it is usually displayed on the central control display.

[0200] Specifically, multiple sample prompts can be analyzed to obtain the deviation information between the user's ideal interface and the interface to be processed. Then, the interface generation model can be adjusted based on the deviation information to obtain the adjusted model.

[0201] It should be noted that, if the confidence level is determined to be greater than or equal to the fourth threshold, a policy adjustment instruction (i.e., a model adjustment instruction) can be sent to [the relevant authority / organization]. Figure 3 The optimization execution and evaluation module in the middle, to trigger Figure 2 The strategy optimization process in the middle.

[0202] It is understood that in the interface optimization method provided in this application embodiment, firstly, third interaction information between multiple users and the interface to be processed within a second period is obtained; and based on the interface generation instruction, the third interaction information, and the satisfaction status of each user with the interface after operation, multiple sample prompt information is constructed; the interface after operation is obtained after the user adjusts the first component during the interaction with the interface to be processed; then, the confidence level corresponding to the multiple sample prompt information is determined; and if the confidence level is greater than or equal to a fourth threshold, the interface generation model is adjusted based on the multiple sample prompt information to obtain the adjusted model. In this way, by constructing structured sample prompt information containing "interface generation instruction, user correction behavior, and user satisfaction status" to incrementally fine-tune the interface generation model, the model can learn from the real correction behavior of a large number of users how to generate an interface that better meets user expectations, thereby improving the model's ability to continuously self-evolve based on group feedback, so that the model can accurately generate an interface that meets user expectations and improve the user experience.

[0203] In some embodiments, determining the confidence levels corresponding to multiple sample prompts further includes the following steps 61 to 63: Step 61: Determine the amount of data for multiple sample prompts and the similarity between multiple sample prompts.

[0204] In the embodiments of this application, the amount of data for multiple sample prompts can be determined, i.e., the total number of sample prompts and the similarity between multiple sample prompts.

[0205] Step 62: Determine the scene coverage corresponding to the target interface content in the interface to be processed.

[0206] Among them, the target interface content consists of multiple pieces of content in the interface to be processed that are not of interest to users; scene coverage represents the breadth of scenes covered by the target interface content.

[0207] In this embodiment, the target interface content is content that is uninteresting / unsatisfactory to all users in the interface to be processed. This content can be a specific first component, the layout information of a component, or other information. Specifically, the target interface content is determined based on multiple sample prompts, and all interfaces where this unsatisfactory content has appeared are identified. The number of scenes corresponding to these interfaces is then determined, and the proportion of this number of scenes to the total number of scenes is calculated, which is the scene coverage.

[0208] For example, if the target interface content is a weather component, then all interfaces where the weather component has appeared can be identified, and there are a total of 5 scenarios corresponding to these interfaces (campsite recommendation scenario, long-distance travel scenario, etc.), and the total number of scenarios is 13. Then the scenario coverage is 5 / 13=0.385.

[0209] Step 63: Obtain the second moment when the third interaction information is generated, and determine the third time interval based on the second moment and the current moment.

[0210] In this embodiment of the application, the second moment when the third interaction information is generated can be determined, and the time interval between each second moment and the current moment can be determined. Then, the average value of multiple time intervals can be calculated to obtain the third time interval.

[0211] Step 64: Determine the confidence level based on data volume, similarity, scene coverage, and third time interval.

[0212] In this embodiment of the application, a portion of the sample prompts that are more similar can be determined from multiple sample prompts based on multiple similarities. The number of such sample prompts is divided by the total number of all sample prompts to obtain the target sample proportion. Then, the target sample proportion, data volume, scene coverage, and third time interval are calculated to obtain the confidence level.

[0213] In one feasible approach, the confidence level can be calculated using the following formula (3).

[0214] Formula (3) Where GCC represents the second evaluation parameter corresponding to a certain first component; N represents the amount of data; The minimum sample threshold is 100; P represents the target sample percentage; R represents the scene coverage; and D represents the third time interval.

[0215] For example, if N is 1260, P is 0.89, R is 0.92, and D is 1, then GCC can be calculated as 0.4×126+0.3×0.89+0.2×0.92+0.1×1=0.94.

[0216] It is understood that in the interface optimization method provided in this application embodiment, the data volume of multiple sample prompts and the similarity between multiple sample prompts are first determined; the scene coverage corresponding to the target interface content in the interface to be processed is determined; wherein, the target interface content is content that multiple users are not interested in; scene coverage represents the breadth of the scenes covered by the target interface content; the second moment of the generation of the third interaction information is obtained, and the third time interval is determined based on the second moment and the current moment; the confidence level is determined based on the data volume, similarity, scene coverage and the third time interval. In this way, through the comprehensive quantitative evaluation of the four dimensions of data volume, similarity, scene coverage and time interval of the sample prompts, the group consensus confidence level between multiple sample prompts can be determined, and then when the sample quality meets the standard, incremental fine-tuning of the interface generation model is triggered, thereby avoiding the model being incorrectly optimized due to insufficient samples, disagreements, single scene or outdated data, and thus improving the accuracy of model adjustment.

[0217] In some embodiments, the interface generation model is adjusted based on multiple sample prompts to obtain an adjusted model, and the process further includes the following steps 71 to 73: Step 71: Process the prompts from multiple samples to obtain deviation information.

[0218] In this embodiment of the application, multiple sample prompts are analyzed to determine the deviation information between the user's ideal interface (i.e. the interface after operation) and the interface to be processed, thereby locating the defects of the interface generation model in terms of layout, components, and interaction logic.

[0219] Step 72: Determine the target parameter to be adjusted from multiple parameters of the model generated from the interface based on the deviation information.

[0220] In this embodiment, the interface generation model has many parameters, including parameters for controlling component size and parameters for controlling component layout. Specifically, the target parameter to be adjusted can be accurately determined from multiple parameters based on the deviation information obtained from the analysis.

[0221] For example, if the deviation information is: non-core components are too prominent and the component layout is unreasonable, then the determined target parameters are the parameters of the layout decoder. In other words, the adjusted model can be obtained by adjusting the parameters of the layout decoder in the interface generation model according to the deviation information, without needing to adjust all the parameters of the model.

[0222] Step 73: Adjust the target parameters of the interface-generated model based on the deviation information to obtain the adjusted model.

[0223] In this embodiment, after determining the target parameters, the target parameters can be directly adjusted based on the deviation information. That is, the interface generation model is adjusted using incremental fine-tuning to obtain the adjusted initial model. Then, the model can be pushed to 5% of the gray users among all users for 7 days of testing and verification. If the verification results show that the task completion time is shortened by 10% or more and the error rate is reduced by 8% or more, then the adjusted initial model is determined to be effective. At this time, the adjusted initial model is determined to be the final adjusted model and is fully released to all users.

[0224] It is understood that in the interface optimization method provided in this application embodiment, firstly, multiple sample prompts are processed to obtain deviation information; then, based on the deviation information, target parameters to be adjusted are determined from multiple parameters of the interface generation model; finally, the target parameters of the interface generation model are adjusted based on the deviation information to obtain the adjusted model. In this way, by comparing and analyzing user correction behavior with the model generation results, specific defects in the model's layout, components, and interaction logic are accurately located, and the target parameters corresponding to these defects are adjusted accordingly. Thus, model optimization can be achieved by fine-tuning only local parameters, without adjusting all model parameters. This significantly reduces computational resource consumption and training time while ensuring optimization effectiveness, achieving efficient and accurate model optimization.

[0225] This application provides embodiments such as Figure 6The complete self-evolving framework of "quantitative feedback-layered decision-making-closed-loop verification" shown includes three core layers: The individual layer calculates the Correction Intensity Index (CSI) by collecting real-time user interaction information with interface components, and stores preference information in short-term, medium-term, temporary, or long-term memory based on the degree of negativity. This allows for automatic interface adjustment when the user issues the same command again, achieving accurate extraction of long-term, highly personalized features and solving the problems of inaccurate preference judgment and susceptibility to interference in existing technologies. The component layer aggregates feedback data from multiple users according to a preset cycle, calculating the Comprehensive Failure Index (CFI) based on the closure rate, misoperation rate, negative voice feedback rate, and operational complexity. When both exceed the threshold for two consecutive cycles... The system triggers component replacement based on the given value, and ensures that the candidate component outperforms the original component through gray-scale A / B testing and statistical verification. This achieves automatic selection and self-verification of components, thereby solving the problems of blind component replacement and uncontrollable effects in existing technologies. The strategy layer constructs structured sample prompt information containing "interface generation instructions - user negative operation - user satisfaction status" over a longer second cycle. It calculates the group consensus confidence score (GCC) based on data volume, similarity, scene coverage, and time interval. When the confidence score reaches the threshold, it extracts deviation information to perform incremental hierarchical fine-tuning of the interface generation model, thus forming a complete closed loop from individual preference memory to group component evolution and model self-optimization. This solves the problems of low model update efficiency and easy introduction of noise in existing technologies.

[0226] Based on the foregoing embodiments, this application provides an interface optimization device that can be applied to... Figure 1-6 In the corresponding embodiment, the interface optimization method is provided with reference to Figure 7 As shown, the interface optimization device 4 may include: an acquisition unit 41, a first determination unit 42, a second determination unit 43, and an adjustment unit 44, wherein: The acquisition unit 41 is used to acquire the first interaction information between the target user and the interface to be processed with the first component; wherein the interface to be processed is generated based on the interface generation instructions input by the target user. The first determining unit 42 is used to determine, based on the first interaction information, the first evaluation parameter corresponding to the first component and the target user's preference information for the first component; wherein, the first evaluation parameter represents the degree of rejection of the first component by the target user; The second determining unit 43 is used to determine the target storage area corresponding to the preference information from multiple storage areas based on the first evaluation parameter, and to store the preference information in the target storage area; wherein the storage time of data in different storage areas is different; The adjustment unit 44 is used to obtain the preference information from the target storage area and adjust the first component in the interface to be processed based on the preference information if the interface generation instruction is received again within the storage time of the preference information to obtain the target interface.

[0227] In other embodiments of this application, the first determining unit 42 is further configured to perform the following steps: Based on the first interaction information, determine the first complexity of the first operation corresponding to the first interaction information, and the type of the first operation; wherein, the type represents the strength of the target user's intention to modify the first component; Determine the first number of times an interface generation command was received within a historical time period; If the first count is greater than the first threshold, determine the historical interface correction time closest to the current time, and determine the first time interval based on the historical interface correction time and the current time; where the historical interface correction time is the historical time when the interface to be processed was corrected. If the first count equals the first threshold, the first time interval is determined as the target value; The first evaluation parameter is determined based on the first complexity, type, first number, first time interval, and first coefficient.

[0228] In other embodiments of this application, the second determining unit 43 is further configured to perform the following steps: If the first evaluation parameter is less than the second threshold, the target storage area is determined to be the first storage area among multiple storage areas; If the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, the target storage region is determined to be the second storage region among multiple storage regions. If the first evaluation parameter is greater than or equal to the third threshold, the target storage region is determined from other storage regions based on the first number; wherein, other storage regions are regions other than the first storage region and the second storage region among multiple storage regions.

[0229] In other embodiments of this application, the second determining unit 43 is further configured to perform the following steps: If the first count is equal to the first threshold, or if the first count is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, the target storage area is determined to be the third storage area among other storage areas. If the first number is greater than the first threshold and there are multiple such numbers, multiple target historical evaluation parameters are determined from the multiple historical evaluation parameters based on the generation time of the historical evaluation parameters. If any historical evaluation parameter of a target is less than the third threshold, the target storage area is determined as the third storage area. If the historical evaluation parameters of each target are greater than or equal to the third threshold, the target storage area is determined to be the fourth storage area among the other storage areas.

[0230] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: The system acquires second interaction information between multiple target users and the interface to be processed within the current first period, and determines second evaluation parameters of multiple target users toward the first component within the current first period based on the second interaction information; wherein, the second evaluation parameters characterize the degree of collective rejection of the first component by multiple target users. Determine the third evaluation parameters for the first component for multiple target users within the target adjacent period; wherein, the target adjacent period is the first period adjacent to the current first period; If the sub-component library corresponding to the interface to be processed is determined to meet the update conditions based on the second evaluation parameter and the third evaluation parameter, the component to be replaced is determined from the sub-component library, and the candidate component is determined from the target component library; If a candidate component meets the target conditions, the component to be replaced will be replaced with the candidate component.

[0231] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: Obtain the first moment when the second interaction information is generated, and determine the second time interval based on the first moment and the current moment; Based on the second interaction information, determine the second complexity of the second operation corresponding to the second interaction information, and the availability parameter of the first component; wherein, the availability parameter characterizes the availability of the first component; The second evaluation parameter is determined based on the second complexity, the availability parameter, the second time interval, and the second coefficient. Correspondingly, the methods also include: If both the second and third evaluation parameters are greater than the third threshold, the sub-component library is determined to meet the update conditions.

[0232] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: Based on the functional information of the component to be replaced, multiple candidate components are determined from the target component library; Determine the historical preference levels of multiple target users for each candidate component, and determine the fit between multiple target users and each candidate component; Determine the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed; Based on historical preference level, matching degree, fit degree and third coefficient, determine the target score for each candidate component; Candidate components are determined from multiple candidate components based on multiple target scores.

[0233] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: Acquire third-party interaction information between multiple users and the interface to be processed within the second cycle; Based on the interface generation instructions, the third interaction information, and each user's satisfaction with the post-operation interface, multiple sample prompt messages are constructed; among them, the post-operation interface is obtained after the user adjusts the first component during the interaction with the interface to be processed. Determine the confidence level corresponding to multiple sample prompts; whereby the confidence level characterizes the degree of consistency among multiple sample prompts. If the confidence level is greater than or equal to the fourth threshold, the interface generation model is adjusted based on multiple sample prompts to obtain the adjusted model.

[0234] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: Determine the amount of data for multiple sample prompts and the similarity between multiple sample prompts; Determine the scene coverage corresponding to the target interface content in the interface to be processed; where the target interface content is multiple pieces of content in the interface to be processed that are not of interest to users; scene coverage represents the breadth of scenes covered by the target interface content. Obtain the second moment when the third interaction information is generated, and determine the third time interval based on the second moment and the current moment; The confidence level is determined based on the amount of data, similarity, scene coverage, and a third time interval.

[0235] In other embodiments of this application, the adjustment unit 44 is further configured to perform the following steps: Deviation information is obtained by processing multiple sample prompts; The target parameter to be adjusted is determined from multiple parameters of the interface-generated model based on the deviation information. The target parameters of the interface generation model are adjusted based on the deviation information to obtain the adjusted model.

[0236] It should be noted that a detailed explanation of the steps performed by each unit can be found in [reference needed]. Figure 1-6 The description of the interface optimization method provided in the corresponding embodiments will not be repeated here.

[0237] The interface optimization method provided in the embodiments of this application can automatically store the target user's preference information for the first component in the interface to be processed. Then, upon receiving the interface generation instruction again, it automatically adjusts the first component in the interface to be processed according to the preference information, eliminating the need for repeated manual operation by the user and thus improving the efficiency of interface optimization. Furthermore, by determining the degree of the target user's disapproval of the first component, the target storage area corresponding to the preference information can be identified. This allows the system to distinguish whether the preference is short-term, medium-term, or long-term based on the storage duration of the target storage area, thereby determining whether to temporarily adjust the interface to be processed or to make a permanent adjustment. This avoids the problem of the interface being permanently changed due to user error, thus improving the reliability of interface optimization.

[0238] Based on the foregoing embodiments, embodiments of this application provide an interface optimization device, which can be applied to... Figure 1-6 In the corresponding embodiment, the interface optimization method is provided with reference to Figure 8 As shown, the interface optimization device 5 may include: a processor 51, a memory 52, and a communication bus 53, wherein: Communication bus 53 is used to realize the communication connection between processor 51 and memory 52; Processor 51 is used to execute the interface optimization program in memory 52 to perform the following steps: Obtain first interaction information between the target user and the interface to be processed, which has a first component; wherein the interface to be processed is generated based on the interface generation instructions input by the target user; Based on the first interaction information, the first evaluation parameter corresponding to the first component and the target user's preference information for the first component are determined; wherein, the first evaluation parameter represents the degree of rejection of the first component by the target user; Based on the first evaluation parameter, the target storage area corresponding to the preference information is determined from multiple storage areas, and the preference information is stored in the target storage area; wherein, the storage time of data in different storage areas is different; If an interface generation instruction is received again within the storage period of the preference information, the preference information is retrieved from the target storage area, and the first component in the interface to be processed is adjusted based on the preference information to obtain the target interface.

[0239] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Based on the first interaction information, determine the first complexity of the first operation corresponding to the first interaction information, and the type of the first operation; wherein, the type represents the strength of the target user's intention to modify the first component; Determine the first number of times an interface generation command was received within a historical time period; If the first count is greater than the first threshold, determine the historical interface correction time closest to the current time, and determine the first time interval based on the historical interface correction time and the current time; where the historical interface correction time is the historical time when the interface to be processed was corrected. If the first count equals the first threshold, the first time interval is determined as the target value; The first evaluation parameter is determined based on the first complexity, type, first number, first time interval, and first coefficient.

[0240] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: If the first evaluation parameter is less than the second threshold, the target storage area is determined to be the first storage area among multiple storage areas; If the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, the target storage region is determined to be the second storage region among multiple storage regions. If the first evaluation parameter is greater than or equal to the third threshold, the target storage region is determined from other storage regions based on the first number; wherein, other storage regions are regions other than the first storage region and the second storage region among multiple storage regions.

[0241] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: If the first count is equal to the first threshold, or if the first count is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, the target storage area is determined to be the third storage area among other storage areas. If the first number is greater than the first threshold and there are multiple such numbers, multiple target historical evaluation parameters are determined from the multiple historical evaluation parameters based on the generation time of the historical evaluation parameters. If any historical evaluation parameter of a target is less than the third threshold, the target storage area is determined as the third storage area. If the historical evaluation parameters of each target are greater than or equal to the third threshold, the target storage area is determined to be the fourth storage area among the other storage areas.

[0242] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: The system acquires second interaction information between multiple target users and the interface to be processed within the current first period, and determines second evaluation parameters of multiple target users toward the first component within the current first period based on the second interaction information; wherein, the second evaluation parameters characterize the degree of collective rejection of the first component by multiple target users. Determine the third evaluation parameters for the first component for multiple target users within the target adjacent period; wherein, the target adjacent period is the first period adjacent to the current first period; If the sub-component library corresponding to the interface to be processed is determined to meet the update conditions based on the second evaluation parameter and the third evaluation parameter, the component to be replaced is determined from the sub-component library, and the candidate component is determined from the target component library; If a candidate component meets the target conditions, the component to be replaced will be replaced with the candidate component.

[0243] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Obtain the first moment when the second interaction information is generated, and determine the second time interval based on the first moment and the current moment; Based on the second interaction information, determine the second complexity of the second operation corresponding to the second interaction information, and the availability parameter of the first component; wherein, the availability parameter characterizes the availability of the first component; The second evaluation parameter is determined based on the second complexity, the availability parameter, the second time interval, and the second coefficient. Correspondingly, the methods also include: If both the second and third evaluation parameters are greater than the third threshold, the sub-component library is determined to meet the update conditions.

[0244] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Based on the functional information of the component to be replaced, multiple candidate components are determined from the target component library; Determine the historical preference levels of multiple target users for each candidate component, and determine the fit between multiple target users and each candidate component; Determine the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed; Based on historical preference level, matching degree, fit degree and third coefficient, determine the target score for each candidate component; Candidate components are determined from multiple candidate components based on multiple target scores.

[0245] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Acquire third-party interaction information between multiple users and the interface to be processed within the second cycle; Based on the interface generation instructions, the third interaction information, and each user's satisfaction with the post-operation interface, multiple sample prompt messages are constructed; among them, the post-operation interface is obtained after the user adjusts the first component during the interaction with the interface to be processed. Determine the confidence level corresponding to multiple sample prompts; whereby the confidence level characterizes the degree of consistency among multiple sample prompts. If the confidence level is greater than or equal to the fourth threshold, the interface generation model is adjusted based on multiple sample prompts to obtain the adjusted model.

[0246] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Determine the amount of data for multiple sample prompts and the similarity between multiple sample prompts; Determine the scene coverage corresponding to the target interface content in the interface to be processed; where the target interface content is multiple pieces of content in the interface to be processed that are not of interest to users; scene coverage represents the breadth of scenes covered by the target interface content. Obtain the second moment when the third interaction information is generated, and determine the third time interval based on the second moment and the current moment; The confidence level is determined based on the amount of data, similarity, scene coverage, and a third time interval.

[0247] In other embodiments of this application, processor 51 is used to execute an interface optimization program in memory 52 to perform the following steps: Deviation information is obtained by processing multiple sample prompts; The target parameter to be adjusted is determined from multiple parameters of the interface-generated model based on the deviation information. The target parameters of the interface generation model are adjusted based on the deviation information to obtain the adjusted model.

[0248] It should be noted that a detailed description of the steps performed by processor 51 can be found in [reference needed]. Figure 1-6 The interface optimization methods provided in the corresponding embodiments will not be elaborated here.

[0249] The interface optimization method provided in the embodiments of this application can automatically store the target user's preference information for the first component in the interface to be processed. Then, upon receiving the interface generation instruction again, it automatically adjusts the first component in the interface to be processed according to the preference information, eliminating the need for repeated manual operation by the user and thus improving the efficiency of interface optimization. Furthermore, by determining the degree of the target user's disapproval of the first component, the target storage area corresponding to the preference information can be identified. This allows the system to distinguish whether the preference is short-term, medium-term, or long-term based on the storage duration of the target storage area, thereby determining whether to temporarily adjust the interface to be processed or to make a permanent adjustment. This avoids the problem of the interface being permanently changed due to user error, thus improving the reliability of interface optimization.

[0250] Based on the foregoing embodiments, embodiments of this application provide a computer-readable storage medium storing one or more programs, which can be executed by one or more processors to implement... Figure 1-6 The corresponding implementation provides the steps of the interface optimization method.

[0251] Based on the foregoing embodiments, embodiments of this application provide a computer program product, which includes a computer program that is implemented when executed by processor 71. Figure 1-6 The corresponding implementation provides the steps of the interface optimization method.

[0252] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0253] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. 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 can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, 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, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0254] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0255] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0256] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. An interface optimization method, characterized in that, The method includes: Obtain first interaction information between the target user and the interface to be processed, which has a first component; wherein the interface to be processed is generated based on the interface generation instructions input by the target user; Based on the first interaction information, a first evaluation parameter corresponding to the first component and the target user's preference information for the first component are determined; wherein, the first evaluation parameter represents the degree of the target user's negative attitude towards the first component; Based on the first evaluation parameter, the target storage region corresponding to the preference information is determined from multiple storage regions, and the preference information is stored in the target storage region; wherein, the storage time of data in different storage regions is different; If the interface generation instruction is received again within the storage period of the preference information, the preference information is retrieved from the target storage area, and the first component in the interface to be processed is adjusted based on the preference information to obtain the target interface.

2. The method according to claim 1, characterized in that, The step of determining the first evaluation parameter corresponding to the first component based on the first interaction information includes: Based on the first interaction information, a first complexity of the first operation corresponding to the first interaction information and a type of the first operation are determined; wherein, the type characterizes the strength of the target user's intention to modify the first component; Determine the first number of times the interface generation instruction was received within a historical time period; If the first number of times is greater than the first threshold, the most recent historical interface correction time is determined, and a first time interval is determined based on the historical interface correction time and the current time; wherein, the historical interface correction time is the historical time when the interface to be processed was corrected. If the first number of times is equal to the first threshold, the first time interval is determined to be the target value; The first evaluation parameter is determined based on the first complexity, the first type, the first number of times, the first time interval, and the first coefficient.

3. The method according to claim 2, characterized in that, The step of determining the target storage region corresponding to the preference information from multiple storage regions based on the first evaluation parameter includes: If the first evaluation parameter is less than the second threshold, the target storage area is determined to be the first storage area among the plurality of storage areas; If the first evaluation parameter is greater than or equal to the second threshold and less than the third threshold, the target storage area is determined to be the second storage area among the plurality of storage areas; If the first evaluation parameter is greater than or equal to the third threshold, the target storage region is determined from other storage regions based on the first number of times; wherein, the other storage regions are regions other than the first storage region and the second storage region among the plurality of storage regions.

4. The method according to claim 3, characterized in that, Determining the target storage region from other storage regions based on the first number of times includes: If the first number of times is equal to the first threshold, or if the first number of times is greater than the first threshold and the number of historical evaluation parameters corresponding to the first component is one, the target storage area is determined to be the third storage area among the other storage areas; If the first number is greater than the first threshold and the number is multiple, multiple target historical evaluation parameters are determined from the multiple historical evaluation parameters based on the generation time of the historical evaluation parameters. If any historical evaluation parameter of a target is less than the third threshold, the target storage area is determined to be the third storage area. If the historical evaluation parameters of each target are greater than or equal to the third threshold, the target storage area is determined to be the fourth storage area among the other storage areas.

5. The method according to claim 1, characterized in that, The method further includes: The system acquires second interaction information between multiple target users and the interface to be processed within the current first period, and determines second evaluation parameters of multiple target users toward the first component within the current first period based on the second interaction information; wherein the second evaluation parameters characterize the degree of collective rejection of the first component by multiple target users. A third evaluation parameter for the first component is determined for multiple target users within a target adjacent period; wherein, the target adjacent period is the first period adjacent to the current first period; If the sub-component library corresponding to the interface to be processed is determined to meet the update conditions based on the second evaluation parameter and the third evaluation parameter, the component to be replaced is determined from the sub-component library, and the candidate component is determined from the target component library; If the candidate component meets the target condition, the component to be replaced is replaced with the candidate component.

6. The method according to claim 5, characterized in that, The step of determining the second evaluation parameters of multiple target users for the first component within the current first period based on the second interaction information includes: Obtain the first moment when the second interaction information is generated, and determine the second time interval based on the first moment and the current moment; Based on the second interaction information, the second complexity of the second operation corresponding to the second interaction information is determined, as well as the availability parameter of the first component; wherein, the availability parameter characterizes the availability of the first component; The second evaluation parameter is determined based on the second complexity, the availability parameter, the second time interval, and the second coefficient. Accordingly, the method further includes: If both the second evaluation parameter and the third evaluation parameter are greater than the third threshold, it is determined that the sub-component library meets the update condition.

7. The method according to claim 5, characterized in that, The step of determining candidate components from the target component library includes: Based on the functional information of the component to be replaced, a number of candidate components are determined from the target component library; Determine the historical preference level of the multiple target users for each candidate component, and determine the degree of fit between the multiple target users and each candidate component; Determine the degree of matching between the functional information of each candidate component and the scene information corresponding to the interface to be processed; Based on the historical preference level, the matching degree, the fit degree, and the third coefficient, the target score corresponding to each candidate component is determined; The candidate component is determined from the plurality of candidate components based on multiple target scores.

8. The method according to claim 1, characterized in that, The method further includes: Acquire third interaction information between multiple users and the interface to be processed within the second cycle; Based on the interface generation instructions, the third interaction information, and each user's satisfaction with the post-operation interface, multiple sample prompt messages are constructed; wherein, the post-operation interface is obtained by the user adjusting the first component during the interaction with the interface to be processed; Determine the confidence level corresponding to the plurality of sample prompts; wherein, the confidence level characterizes the degree of consistency among the plurality of sample prompts; If the confidence level is greater than or equal to the fourth threshold, the interface generation model is adjusted based on the multiple sample prompts to obtain the adjusted model.

9. The method according to claim 8, characterized in that, Determining the confidence level corresponding to the plurality of sample prompts includes: Determine the amount of data in the plurality of sample prompts and the similarity between the plurality of sample prompts; Determine the scene coverage corresponding to the target interface content in the interface to be processed; wherein, the target interface content is multiple pieces of content in the interface to be processed that are not of interest to users; the scene coverage represents the breadth of the scenes covered by the target interface content; Obtain the second moment when the third interaction information is generated, and determine the third time interval based on the second moment and the current moment; The confidence level is determined based on the amount of data, the similarity, the scene coverage, and the third time interval.

10. The method according to claim 8, characterized in that, The interface generation model is adjusted based on the multiple sample prompts to obtain the adjusted model, including: The deviation information is obtained by processing the multiple sample prompts. Based on the deviation information, the target parameter to be adjusted is determined from multiple parameters of the model generated by the interface; The target parameters of the interface generation model are adjusted based on the deviation information to obtain the adjusted model.

11. An interface optimization device, characterized in that, The device includes: The acquisition unit is used to acquire first interaction information between the target user and the interface to be processed, which has a first component; wherein the interface to be processed is generated based on the interface generation instruction input by the target user. The first determining unit is configured to determine, based on the first interaction information, a first evaluation parameter corresponding to the first component and the target user's preference information for the first component; wherein, the first evaluation parameter characterizes the degree of rejection of the target user towards the first component; The second determining unit is configured to determine the target storage region corresponding to the preference information from multiple storage regions based on the first evaluation parameter, and store the preference information in the target storage region; wherein the storage time of data in different storage regions is different; An adjustment unit is configured to, if the interface generation instruction is received again within the storage period of the preference information, retrieve the preference information from the target storage area and adjust the first component in the interface to be processed based on the preference information to obtain the target interface.

12. An interface optimization device, characterized in that, The device includes: a processor, a memory, and a communication bus; The communication bus is used to realize the communication connection between the processor and the memory; The processor is used to execute an interface optimization program in memory to implement the steps of the interface optimization method as described in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, which can be executed by one or more processors to implement the steps of the interface optimization method as described in any one of claims 1 to 10.

14. A computer program product, the computer program product comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the interface optimization method according to any one of claims 1 to 10.

Images