Data processing method and data processing device
Through multiple rounds of data processing and guidance information feedback optimization, the problem of poor accuracy of automatic output files for user input description information in intelligent application scenarios is solved, and more accurate user intention matching is achieved, improving the data processing accuracy and user satisfaction of intelligent applications Spend.
Patent Information
- Application Number
- CN202411203847.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-12-03
AI Technical Summary
In the intelligent application scenarios of the prior art, the files automatically output by the description information input by the user are poor in accuracy and cannot meet the actual needs of the user.
By responding to user input data, output response results and guidance information, update the attribute parameters of the response results, and generate the second input data based on the feedback information until a target file matching the user's intention is generated. The method includes multiple rounds of data processing and guided information feedback optimization, and improve the accuracy of output using technologies such as processing models and user portraits.
It significantly improves the accuracy of output files and user satisfaction, can more accurately understand and meet user intentions, and improves the accuracy of data processing of intelligent applications.
Smart Images

Figure CN119065567A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and more specifically to a data processing method and a data processing device. Background Art
[0002] In intelligent application scenarios, users can input relevant descriptive information and obtain files that match the user's intentions through intelligent models or related processing methods. For example, users can enter keywords describing images into drawing tools, and the drawing tools automatically generate images based on the input description keywords, and finally generate images that meet the user's description. However, the files that are automatically output based on the description information entered by the user are often far from the files that the user actually expects to obtain, and their accuracy is poor, which cannot meet the actual needs of users. Summary of the invention
[0003] In view of this, this application provides the following technical solutions:
[0004] The first aspect of the present application provides a data processing method, comprising:
[0005] In response to obtaining first input data, outputting a response result for the first input data and guidance information for the response result, wherein the guidance information is used to update an attribute parameter of the response result;
[0006] generating second input data based on feedback information of the guidance information;
[0007] A target file matching the user intention is generated based on the user intention represented by the second input data.
[0008] Optionally, the outputting a response result to the first input data and guiding information to the response result includes at least one of the following:
[0009] Obtaining a first processing instruction for the first input data, and performing generation processing on the first input data using a first processing model matching the first processing instruction to output the response result and guidance information for the response result;
[0010] Obtaining a first processing instruction for the first input data, performing generation processing on the first input data using a first processing model matching the first processing instruction to output the response result, and outputting guidance information for the response result using a second processing model matching the first processing instruction;
[0011] Generate and process first target data in the first input data based on the type of the first input data and the represented user intention, so as to output the response result and guidance information for the response result;
[0012] The first input data and context data of its source application are integrated to obtain the third input data, and the third input data is generatively processed using a generative model to output the response result and guidance information for the response result.
[0013] Optionally, obtaining a first processing instruction for the first input data includes at least one of the following:
[0014] generating the first processing instruction in response to obtaining a trigger operation of a first type of control acting on a first type of application, wherein the first type of application is an application capable of responding to the first input data;
[0015] generating the first processing instruction in response to obtaining a trigger operation for first input data input to a second type of application, the second type of application being an application that can respond to or cannot respond to the first input data;
[0016] In response to obtaining first input data input to a third type of application, generating a corresponding first processing instruction based on application information of the third type of application;
[0017] A corresponding first processing instruction is generated based on the type of the first input data and / or application information of a source application.
[0018] Optionally, the generating the second input data based on the feedback information of the guidance information comprises at least one of the following:
[0019] Integrate the first input data and the second target data in the guidance information determined based on the feedback information to obtain the second input data;
[0020] integrating feedback information different from the guidance information and the first input data to obtain the second input data;
[0021] Second input data is generated based on feedback information of the guide information and a conversion result of the response result.
[0022] Optionally, the generating the second input information based on the feedback information of the guidance information includes:
[0023] generating fourth input data based on first feedback information of first guidance information of a first response result to the first input data, and outputting a second response result to the fourth input data and second guidance information to the second response result;
[0024] generating fifth input data based on second feedback information to the second guidance information, and outputting a third response result to the fifth input data and third guidance information to the third response result;
[0025] The second input data is generated based on third feedback information of the third guidance information.
[0026] Optionally, the generating the second input data based on the feedback information of the guidance information comprises at least one of the following:
[0027] Obtaining target user portrait information corresponding to the first input data, and generating the second input data based on the user portrait information and feedback information on the guidance information, wherein the target user portrait includes group user portrait information or user portrait information of the target user;
[0028] The current environment information of the electronic device is obtained, and the second input data is generated with reference to the environment information and feedback information of the guide information.
[0029] Optionally, the generating the second input data based on the feedback information of the guidance information comprises at least one of the following:
[0030] generating second input data based on the obtained selection operation of the recommended data for the third target data in the response result and / or the edit operation for the second target data;
[0031] Second input data is generated based on the configuration operation obtained for the configuration options in the configuration window of the response result, where the configuration options are used to update the target data in the response result.
[0032] Optionally, the generating a target file matching the user intention based on the user intention represented by the second input data includes:
[0033] generating a second processing instruction based on the attribute information of the target application and / or the user intention represented by the second input data;
[0034] The second input data is generated and processed using a processing model or a model service corresponding to the second processing instruction to obtain a target file that matches the attribute information and the user's intention.
[0035] Optionally, at least one of the following is also included:
[0036] In response to obtaining an adjustment operation on the target file, updating configuration parameters of the target file and model parameters of a processing model for generating the target file;
[0037] In the case where the target file is a human-computer interaction interface, configuration parameters of the human-computer interaction interface and model parameters of a processing model for generating the human-computer interaction interface are updated based on the obtained user behavior data generated by using the human-computer interaction interface.
[0038] A second aspect of the present application provides a data processing device, comprising:
[0039] an output unit, configured to output a response result for the first input data and guidance information for the response result in response to obtaining the first input data, wherein the guidance information is used to update an attribute parameter of the response result;
[0040] A first generating unit, configured to generate second input data based on feedback information of the guiding information;
[0041] The second generating unit is used to generate a target file matching the user intention based on the user intention represented by the second input data.
[0042] A third aspect of the present application provides an electronic device, including:
[0043] A memory, used to store applications and data generated by the operation of the applications;
[0044] The processor is used to execute the application program to implement the data processing method as described in any one of the above. BRIEF DESCRIPTION OF THE DRAWINGS
[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.
[0046] Figure 1 A flowchart of a data processing method provided in an embodiment of the present application;
[0047] Figure 2 A schematic diagram of processing an image generation scenario provided in an embodiment of the present application;
[0048] Figure 3 A schematic diagram of a UI design architecture provided in an embodiment of the present application;
[0049] Figure 4 A data processing device is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION
[0050] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0051] The terms "first", "second", etc. in this application are used to distinguish different objects, rather than to describe a specific order. For example, the first audio signal and the second audio signal in the embodiment of the present application are only audio signals that characterize different audio signals. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes steps or units that are not listed, or optionally includes other steps or units inherent to these processes, methods, products or devices. Mentioning "embodiment" in this application means that the specific features, structures or characteristics described in conjunction with the embodiment may be included in at least one embodiment of the present application. The appearance of this phrase at various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application can be combined with other embodiments.
[0052] The present application embodiment provides a data processing method, which is applied to the scenario of identifying user intentions. By processing the user's input data, a target file matching the user's intentions can be output, thereby improving the accuracy of the output target file. Figure 1 , is a flow chart of a data processing method provided in an embodiment of the present application, the method may include the following steps:
[0053] S101 . In response to obtaining first input data, output a response result corresponding to the first input data and guidance information corresponding to the response result.
[0054] The first input data is the data input by the user in the current application scenario. For example, if the current application scenario is an intelligent question and answer scenario, the first input data may be the question data input by the user. If the current application scenario is an intelligent search scenario, the first input data may be the search data input by the user, etc. Specifically, depending on the application scenario and the application requirements, the data type and data format of the first input data input by the user may be different. For example, the first input data may be text string data, or multimedia data such as audio data, image data, and video data; or, the first input data may also be various types of demand data or instructions for generating a user interface (User Interface, UI, also referred to as a human-computer interaction interface), such as demand instructions for generating a specific display interface or configuration parameters for configuring the display interface and control functions.
[0055] The response result for the first input data refers to the result data obtained by processing the first input data. Specifically, the first input data can be processed based on a specific processing algorithm (or processing model) to obtain a response result. For example, the first input data is the image description information that needs to be obtained, and the processing algorithm (or processing model) can be an image generation algorithm (or raw image model). The response result can be an image obtained by processing the image description information based on the image generation algorithm (or raw image model). Specifically, the response result can be an image, audio, text, video, or user interface information generated according to the processing requirements corresponding to the first input data. In the embodiment of the present application, while outputting the response result for the first input data, guidance information for the response result is also output, and the guidance information is used to update the attribute parameters of the response result. The guidance information is mainly used to optimize the guidance prompts of the response result, such as guidance prompts for the input data, or optimization suggestions or design suggestion information for the generated user interface information. In the process of updating the attribute parameters of the response result, the guidance information can be used to update the parameters that affect the generation of the response result in the corresponding processing algorithm (or processing model), so that the processing algorithm (or processing model) can better adapt to the current application scenario, better understand the input information or user instructions, and output a response result that is more accurate with the input data. Among them, the processing algorithm or processing model is an algorithm or model for processing the input data to obtain a response result for the input data. Through the updated attribute parameters, a more accurate response result can be generated. For example, updating the attribute parameters of the response result through the guidance information can optimize the display effect of the image, the display content or the display effect of the user interface, the interactive effect, the interactive design, etc. It can also be to optimize the display effect of the text, the display mode of the video or the output effect of the audio, or the output content, etc.
[0056] Different types of response results correspond to different configuration parameters. For example, when the response result is an image, the configuration parameters corresponding to the response result may be image parameters, which may be parameters that characterize pixels in the image, such as resolution, pixel size, color depth, contrast, brightness, etc. The image parameters may also be parameters that characterize image content, such as style, color, composition, theme, etc. When the image parameters of the image response result are updated based on the guidance information, the relevant image parameters may be adjusted by an image processing algorithm or an image processing model to achieve the purpose of updating the image parameters, thereby generating a more accurate image response result according to the updated image parameters; if the response result is audio data, the configuration parameters corresponding to the response result may be audio parameters, which may include sampling rate, number of channels, bit rate, frequency response parameters, etc., and may also include the timbre and voice style of the audio object included in the audio; if the response result is text data, the corresponding configuration parameters of the response result may be string parameters, which may include parameters such as string length, string encoding method, string type, etc., and may also be text type, text conversion conditions, or text application parameters; if the response result is user interface information, the corresponding attribute parameters may include user interface display parameters and interaction logic, etc. Therefore, in the embodiment of the present application, while outputting the response result corresponding to the first input data, guidance information for the response result is also output, and the attribute parameters of the response result can be updated through the guidance information so that a more accurate response result can be obtained based on the updated attribute parameters.
[0057] S102: Generate second input data based on feedback information of the guidance information.
[0058] In order to more accurately meet the user's intention and output accurate information, feedback information corresponding to the guidance information can be obtained after the guidance information is output. Among them, the feedback information can be the user's selection information on the content contained in the guidance information, such as selecting the recommended option in the guidance information, and using the selected option as feedback information. In the application scenario of generating images based on the description information input by the user, the guidance information can be the candidate image to be selected by the user, etc., and the feedback information can be the candidate image selected by the user; the feedback information can also be input information other than the guidance information options, such as the user denies the recommended option included in the current guidance information, and uses the denied input data as the feedback information, or the user re-enters new prompt information, etc., and uses the prompt information input by the user as feedback information. Correspondingly, in the application scenario of user interface generation, feedback from users or corresponding user groups on existing user interface design suggestions can also be used as feedback information, or information after user interface testing can be used as feedback information. Generating the second input data based on the feedback information of the guidance information can make the input data more reflective of the user's input intention. For example, the second input information can be obtained by optimizing the first input data based on the feedback information. For example, the first input data is "querying the audio of a certain singer's song", the guidance information can be the different time intervals when the singer frequently releases songs, and the feedback information can be the target time interval selected by the user. Then the second input data can be used to query the audio of the singer's song in the target time interval. When generating the second input data, in addition to the first input data and the feedback information, the second input data can also be determined based on the user feature information corresponding to the input first input data, or the feature information of a user group similar to the user. In addition, the second input data can also be determined based on the current application scenario features, such as the use scenario environment and other information. This can more accurately reflect the user's intentions and current application needs, making the final output result more accurate.
[0059] S103: Generate a target file matching the user intention based on the user intention represented by the second input data.
[0060] After obtaining the second input data, the second input data can be analyzed for intent, such as calling the corresponding intent understanding model according to the current application scenario, processing the second input data through the intent understanding model to obtain the user intent, and then generating a target file of the user intent that matches the user intent, the file type of the target file matches the data type of the input data, such as the target file can be a text file, an image file, a video file, an animation file or a multimodal media data file. Specifically, when generating a target file based on the user intent represented by the second input data, it can be processed based on the target application, for example, the second input data is processed based on the processing instructions input to the target application to generate the target file, and the target application can be an information search application, an application that generates images based on text, or an application that generates a user interface, etc. The type of the target file and the response result can be the same or different, such as the response result can be an image, and the target file can be a video. For example, in the application scenario of generating a video, the current response result can be an image frame, and the target file can be a video generated based on multiple rounds of image frames. Alternatively, in the application scenario of generating a user interface, the response result can be a partial area or user interface element of the generated user interface, and correspondingly, the target file can be a user interaction interface generated based on multiple rounds of user interface elements, etc.
[0061] In the embodiment of the present application, processing can also be performed based on multiple rounds of input data until a target file that meets the user's intention is output. For example, based on the second input data, a response result of the second input data and guidance information for the response result are output, and then the third input data is generated based on the feedback information corresponding to the guidance information. The file is output based on the third input data. If the user has not selected the file, it means that the file does not fully meet the user's intention. The guidance information corresponding to the file can continue to be output, thereby obtaining the fourth input data, and executing to obtain the target file that matches the user's intention. In the embodiment of the present application, the corresponding guidance information can be output while outputting the response result corresponding to the user's input data. In this way, the current data processing mode can be continuously optimized based on the feedback information of the guidance information during the interaction with the user, so that the user's intention can be understood and met more accurately, and the accuracy of the generated file and the user's satisfaction can be improved.
[0062] The following describes the relevant technical features of the data processing method in the embodiment of the present application in combination with actual application scenarios.
[0063] In one implementation of the embodiment of the present application, outputting a response result for the first input data and guidance information for the response result includes at least one of the following:
[0064] (1) Obtaining a first processing instruction for first input data, and generating and processing the first input data using a first processing model matching the first processing instruction to output a response result and guidance information for the response result.
[0065] The first processing instruction represents the processing requirement instruction for the first input data, such as the first processing instruction can be an instruction to generate a corresponding image based on the input text data, or a text generation instruction to generate a target article format according to the corresponding description information, or an instruction to generate a user interface, an instruction to generate audio or video, etc. After obtaining the first processing instruction, the processing requirement of the current input data can be determined, so the processing model matching it can be called based on the first processing instruction, that is, the first processing model is obtained, and then the first processing model is used to generate and process the first input data to obtain a response result and guidance information for the response result. For example, the first processing instruction is generated based on the trigger operation of the relevant image generation control in the image generation application, so the corresponding image generation model can be called based on the image generation instruction, and the current first input data is subjected to image generation processing to obtain a preliminarily generated image and corresponding guidance information. Or the instruction generated by the first processing instruction on the text generation application can call the text generation model corresponding to the text generation instruction, perform text generation processing on the first input information, and obtain the corresponding text result and guidance information. In the process of generating guidance information using the first processing model in this embodiment, the relevant configuration items or configuration weights missing in the response result can be determined based on the corresponding configuration parameters and weights of the first input data in the first processing model, thereby generating guidance information. For example, the first processing instruction is an image generation instruction, and the first input data can be "please generate a picture of a dog". The first processing model obtains the user's intention to generate a picture of a dog based on the first input data, and can generate a simple dog image as the current response result. Then, since the configuration items such as the type, action, environment, style, etc. of the dog are not recognized by the current processing model, this information can be used as guidance information to guide the user to enrich this information. In this embodiment, by calling the processing model corresponding to the processing instruction to process the input data, the current processing requirements can be better met and the accuracy of the processing can be improved.
[0066] (2) obtaining a first processing instruction for first input data, performing generation processing on the first input data using a first processing model matching the first processing instruction to output a response result, and outputting guidance information for the response result using a second processing model matching the first processing instruction,
[0067] In this embodiment, the first processing instruction may be an image generation instruction, a text generation instruction, a video generation instruction, a user interface generation instruction, an instruction for generating audio or video, and the like. Therefore, the first processing model corresponding to the first processing instruction may be called based on the first processing instruction. For example, if the first processing instruction is an image generation instruction, the image generation model may be called to process the first input data to obtain an initial image. For another example, if the first processing instruction is generated on a text generation application, the corresponding generated text model may be called to process the input data, and the obtained initial text information may be used as a response result to the first input data. In this embodiment, the first processing model may be used only to process the first input data and output a response result. Based on the above example, the first processing model may be used only to generate images, text, video or audio, user interface, and the like. The second processing model may be a sub-model of the first processing model, or may be a model configured to monitor or test the processing effect of the first processing model. For example, the first processing model is an image generation model, and the second processing model can be configured with standard image generation logic flow and configuration parameters, because it is possible to determine the difference information between the output response result and the standard image generation logic or configuration parameters, and then output corresponding guidance information based on the difference information. The guidance information may be suggestion information for adjusting the output image, etc.
[0068] (3) Generate and process the first target data in the first input data based on the type of the first input data and the represented user intention, so as to output a response result and guidance information for the response result.
[0069] Different types of first input data can obtain different response results and processing methods, and the first input data input by the user is not necessarily all key information for processing. Therefore, the first input data can be pre-processed according to the type of the current first input data to obtain the corresponding first target data. For example, when the first input data is text data, the text data can be semantically analyzed or understood, so as to extract keywords by word segmentation, and the keywords are used as the first target data, and then an image or video corresponding to the text data is generated according to the keywords. Correspondingly, if the first input data is image data, the corresponding processing method can first convert the image into text data, and use the text data as the first target data, so that the corresponding response result is obtained by processing according to the text data. Specifically, the first target data can be all or part of the first input data, such as the first target data is only generated for a specific type of word in the first input data, such as generating the first target data only based on nouns, adjectives or verbs extracted from the first input data. In this embodiment, the first target data in the first input data can be obtained first, and then the generation processing is performed based on the first target data, which can avoid the influence of other interference data in the first input data on the processing process, making the processing process more accurate.
[0070] (4) Integrate and process the first input data and context data of its source application to obtain third input data, and use the generative model to generate and process the third input data to output a response result and guidance information for the response result.
[0071] In application scenarios where the first input data has a context, such as in an intelligent question-answering system, the first input data may be the data currently input by the user, and the context data may be the data previously input by the user and the data output by the intelligent question-answering system. In order to more accurately understand the user's intention, the context data and the first input data may be integrated, such as using the deduplicated data as the third input data. It is also possible to obtain the context corresponding to the first input data within a certain period. For example, in an application scenario where an image is generated based on user description information, the first input data may be the description information currently input by the user, and the context data may be the relevant text data in the process of successfully generating an image. These text data and the description information currently input by the user are determined as the third input data. Such processing based on the third input data can better meet the current application scenario and the user's application needs, making the output response results and guidance information more accurate.
[0072] In one implementation of the embodiment of the present application, the obtaining of the first processing instruction for the first input data includes at least one of the following:
[0073] (1) In response to obtaining a trigger operation acting on a first type of control of a first type of application, generating a first processing instruction.
[0074] Among them, the first type of application is an application that can respond to the first input, which can be an application applied to an electronic device, such as a raw image application that generates images, a raw text application that generates text, or an application that generates videos, etc. These applications can be applications generated in combination with artificial intelligence algorithms or artificial intelligence models; correspondingly, the first type of application can respond to the first input to implement corresponding functional services, such as generating corresponding multimedia data such as images, texts, videos, and audios, and can also be an application that implements artificial intelligence image processing, optimization, etc., or implements control of the device, etc. The first type of control can be a control that can be triggered by the user in the current first type of application. For example, when the first type of application is a raw image application, the first type of control can be an image creation control. Specifically, the first type of control can also be a text generation control, a video creation space, an audio generation control, a user interface generation control, etc. The trigger operation acting on the first type of control can match the trigger type of the first type of control, such as the trigger operation can include a click operation, a selection operation, and can also include a drag operation, a path input operation, etc. For example, if the user expects to generate a picture similar to the current picture, the current picture can be dragged to the area corresponding to the first type of control, or the path information of the current picture can be selected in the path import area of the first type of control. Correspondingly, the first processing instruction may be an instruction generated according to the corresponding function of the first type of control, such as a text generation instruction, or an instruction to optimize the current information, or an image generation instruction, a video generation instruction, a user interface generation instruction, and the like.
[0075] (2) In response to obtaining a trigger operation for first input data input to the second type application, generating a first processing instruction.
[0076] The second type of application is an application that can or cannot respond to the first input data. Correspondingly, the second type of application can be the same as the first type of application described above, or it can be different. If the second type of application is the same as the first type of application, it can directly respond to the first input data. The second type of application can also be a social application or a conference application, such as an instant messaging application, an online conference application, etc. For example, a user can input the first input data in a chat application to make the chat application call a data processing model, so that the data processing model processes the first input data. Specifically, the trigger operation of the first input data input to the second type of application can be an operation that the second type of application can support, such as a copy and paste operation, a data migration operation, a screenshot operation, a text selection operation, a text marking operation, an editing operation, etc. When these trigger operations are detected, a first processing instruction will be generated so that the first input data is processed based on the first processing instruction. If it is detected that the first input data input to the second type of application is copied or intercepted to the first type of application, the generation of the first processing instruction can be triggered so that the first type of application performs a corresponding processing operation on the first input data based on the first processing instruction.
[0077] (3) In response to obtaining first input data input to the third-category application, generating a corresponding first processing instruction based on the application information of the third-category application.
[0078] In this implementation, the corresponding first processing instruction may be generated according to the application type of the third category application. For example, if the third category application is an image generation application, the first processing instruction may be generated according to the image generation rule of the image generation application; or if the third application is a text generation application, the first processing instruction may be generated according to the text generation rule, so as to input the input first input data into the corresponding text generation model for processing to obtain a text result.
[0079] (4) Generate a corresponding first processing instruction based on the type of the first input data and / or application information of the source application.
[0080] The type of the first input data may be text, image, audio, video, etc. The first processing instruction corresponding to the text type may be a text generation instruction, and the first processing instruction corresponding to the image type may be a text-to-image conversion instruction or a text generation instruction, etc. Furthermore, in order to improve the accuracy of the output, the first processing instruction may be generated in combination with the application information of the application receiving the first input data, wherein the source application of the first input information may refer to which application the first input new information is input to, that is, which application the first input data comes from. The application information may include the application type and function type of the source application of the first input data, such as the application type may be an image generation application type, and the function type may be a function for realizing accurate configuration of device configuration parameters, etc. In this way, a more accurate first processing instruction may be generated, thereby realizing accurate information processing.
[0081] In one implementation of the embodiment of the present application, generating the second input data based on the feedback information of the guidance information includes at least one of the following:
[0082] (1) The first input data and the second target data in the guidance information determined based on the feedback information are integrated and processed to obtain the second input data.
[0083] The guidance information may include optimization or prompt information for the current processing. The user can generate feedback information on the guidance information through operations such as selection and marking, thereby determining the target selection information in the guidance information corresponding to the feedback information as the second target data, and integrating the second target data with the first input data to obtain the second input data that can more accurately reflect the user's intention.
[0084] (2) The feedback information different from the guidance information and the first input data are integrated to obtain the second input data.
[0085] If the optional information included in the guidance information does not meet the user's needs, the user can provide feedback information different from the guidance information, such as the user can provide guidance prompts or recommended new options or inputs that are different from the processing model conditions of the current application. For example, the guidance information provided by the processing model allows the user to choose whether the action of the character in the image is running or walking, and the feedback information provided by the user can be the action of sitting, then it is necessary to generate the second input data based on the feedback information provided, such as the first input data is to draw the characters in the house, and the second input data can be to draw the characters sitting in the house, etc. For another example, in the user interface generation scenario, the processing model provides three layout modes, and then the user can provide other layout modes different from these three layout modes, so as to generate the corresponding second input information according to the layout mode provided by the user. In this way, more accurate input information can be generated based on user feedback information, thereby improving the accuracy of information processing.
[0086] (3) Generate second input data based on the feedback information of the guidance information and the conversion result of the response result.
[0087] The response result can be an image, audio, video or text, etc. The response result can be converted according to the information type of the guidance information so that the information type of the converted response result is consistent with the information type of the guidance information, so that when the same type of information is integrated and processed, the accuracy of information processing can be improved, and more detailed information can be obtained. For example, if the guidance information is text information and the response result is image information, the response result can be converted, that is, the image information is converted into text information, and then integrated with the user feedback information, such as deduplication or information detailing, to obtain the second input data.
[0088] In one implementation of the embodiment of the present application, generating the second input information based on the feedback information of the guidance information includes:
[0089] Based on the first feedback information of the first guidance information of the first response result to the first input data, the fourth input information is generated, and the second response result to the fourth input data and the second guidance information for the second response result are output; based on the second feedback information to the second guidance information, the fifth input data is generated, and the third response result to the fifth input data and the third guidance information for the third response result are output; based on the third feedback information of the third guidance information, the second input data is generated.
[0090] In order to accurately output the target file, the processing algorithm of the processing model for processing the input information and the attribute parameters of the corresponding response result can be repeatedly adjusted. For example, if the user inputs the first input data for the first time, the first input data can be processed to output the corresponding first response result and the first guidance information, and then the first feedback information and the first input data are integrated to obtain the fourth input data according to the first feedback information. Correspondingly, the fourth input data can also be data generated based on the first response result and the first feedback information. The fourth input data is processed in the second round to obtain the corresponding second response result and the second guidance information. When the user determines the second feedback information according to the second guidance information, the fifth input data can be generated based on the second feedback information. The fifth input data can be the data converted and integrated from the first input data, the fourth input data or the second response result. The fifth input data is processed in the third round to obtain the corresponding third response result and the third guidance information until the conditions for outputting the most accurate result are met, the corresponding guidance information is stopped, and the target file is output.
[0091] For example, the first input data input by the user for the first time can be "play a cheerful song", the first response result can be a list of candidate songs, and the first guide information can be "please select a male singer or a female singer"; the corresponding first feedback information can be the information of the female singer selected by the user, then the fourth input data is to play a cheerful song sung by a female singer, and then the second response result is a list of candidate songs for the input data, and the second guide information can be the geographical area information of the specific female singer. Then the second feedback information is the target geographical area information selected by the user, and the fifth input data can be to play a cheerful song sung by a female singer from the target geographical area. The third response structure is a list of candidate songs, and the third guide information can be information such as the name of a specific female singer. When the user selects a specific female singer, the second input data can be to play a cheerful song sung by the female singer, so that the target file output by the electronic device can be a cheerful song with the highest click rate sung by the female singer, or automatically played according to a list of songs that meet the conditions. In the embodiment of the present application, multiple rounds of data processing are performed, so that the processing algorithm can better understand the user's intention, output more accurate data, and improve the user experience. Specifically, how many rounds of data processing are performed may be determined by the weight parameters or the number of configuration items configured in the current processing model, or may be determined based on the user's final selection information.
[0092] In one implementation of the embodiment of the present application, generating the second input data based on the feedback information of the guidance information includes at least one of the following:
[0093] (1) Obtain target user portrait information corresponding to the first input data, and generate second input data based on the user portrait information and feedback information on the guidance information.
[0094] The target user portrait includes group user portrait information or user portrait information of the target user. The target user may be the user who inputs the first input data, and the group may be a set of users with associated attributes with the target user. Specifically, the target user portrait may be the characteristic information of the target user, such as age, occupation, geographic location, or frequently used application information, etc. It should be noted that the user characteristic information applied when generating the target user portrait is characteristic information that the corresponding user allows to be disclosed. The user habits or preferences can be obtained through the target user portrait information, so that the weight parameters of the relevant features in the first input data can be adjusted according to this information, and then the response information that meets the user preferences, such as text, image or video, can be generated. In the actual application scenario, it can be distinguished based on the first input data whether it is a user interface design scenario or a personalized image generation or video generation scenario. In the user interface generation scenario, the group user portrait information is usually used to generate the corresponding second input information to better meet the needs of the group customers. In the personalized image generation scenario, the user portrait information of the target user who inputs the first input data is usually used to generate the second input data, which can better meet the needs of the target user. In the embodiment of the present application, after generating the second input data, the weight parameters of the relevant features in the second input data can be adjusted based on the target user portrait information. For example, if the second input data includes text information such as time, place, and person, the weight of the relevant information can be adjusted according to the target user portrait information. For example, if the weight of the location information is adjusted to the maximum, the processing process can pay more attention to the location information to better meet the user's habits or preferences. It is also possible to obtain a more accurate target file by referring to the target user portrait information during the process of generating the target file.
[0095] (2) Obtaining current environment information of the electronic device, and generating second input data with reference to the environment information and feedback information on the guidance information.
[0096] The current environment information of the electronic device can be the current use environment information of the electronic device, such as what kind of information the user usually processes through the electronic device, and can also be information about the external environment in which the electronic device is located, such as brightness parameters. For example, in an image generation scenario, the brightness attribute parameters of the pixels in the generated candidate image can be adjusted according to the brightness parameters of the current environment, so that the output image can better match the current environment and meet the needs of the user. Correspondingly, the environmental parameters can also be the current operating environment of the electronic device. If it is detected that the current network transmission rate of the electronic device is low, when outputting the corresponding image, a compressed version of the image can also be output, which is convenient for the user to quickly download and save. In this way, based on the current environmental information, a second input data that is more in line with the current environment can be generated, so that the target file obtained after the processing model processes the second input data can also better meet the current application scenario.
[0097] In one implementation of the embodiment of the present application, generating the second input data based on the feedback information of the guidance information includes at least one of the following:
[0098] (1) Based on the obtained selection operation of the recommended data for the third target data in the response result and / or the edit operation for the second target data, the second input data is generated.
[0099] The third target data in the response result may be key information in the response result. For example, when the response result is text data, the third target data may be keywords in the text data; if the response result is image data, the third target data may be the main image object in the image data, such as people, scenery, etc. The recommendation data of the corresponding third target data may be recommendation words for keywords in the response result, or recommendation objects for image objects, etc. The selection operation for the recommendation data may be selection, replacement, change, etc. of the recommendation words or recommended objects. The second target data may be all the data in the response result, or part of the data in the response result, such as specific keywords or key objects in the response result. The editing operation for the second target data may be selection, replacement, change, etc. of the response result or key data in the response result. In this way, the second input data is generated based on the data obtained after these operations, which can provide more detailed information in combination with the current application scenario and improve the accuracy of subsequent data processing.
[0100] For example, when the response result is image data, in response to obtaining a trigger operation for a target image object in the image data, a recommended object for the target image object may be output, and the target image object may be replaced with a target recommended object determined by a selection operation on the recommended object, to obtain second input data. When the response result is audio data, in response to obtaining a trigger operation for a target audio segment in the audio data, a recommended audio segment for the target audio segment may be output, and the target audio segment may be replaced with a target recommended audio segment determined by a selection operation on the recommended audio segment, to obtain second input data. When the response result is video data, in response to obtaining a trigger operation for a target image frame in the video data, a recommended image frame for the target image frame may be output, and the target image frame may be replaced with a recommended image frame determined by a selection operation on the recommended image frame, to obtain second input data.
[0101] (2) generating second input data based on the configuration operation obtained for the configuration option in the configuration window of the response result, wherein the configuration option is used to update the target data in the response result.
[0102] In this implementation, a configuration window can be set to assist the user in performing configuration operations to avoid problems such as user input errors. Specifically, the target data in the response result can be a target keyword, a target image frame, a target object in an image, a target audio segment in audio data, a user interface control in a user interface, etc. For example, a configuration operation of replacing or changing the key tag, control, etc. of the key word or key object in the corresponding response result can be performed through a label window. Thus, the second input data that better meets the user's intention is generated according to the information corresponding to the configuration operation.
[0103] In one implementation of the embodiment of the present application, generating a target file matching the user intention based on the user intention represented by the second input data includes:
[0104] Generate a second processing instruction based on the attribute information of the target application and / or the user intention represented by the second input data; use the processing model or model service corresponding to the second processing instruction to generate and process the second input data to obtain a target file that matches the attribute information and the user intention.
[0105] Among them, the attribute information of the target application can be information such as the type, function and usage conditions of the target application. Therefore, the attribute information of the target application can be used to determine whether the current target application is an image generation application, a text generation application, a video generation application, or a user interface generation application, so as to determine whether to call a processing model or a third-party processing service. Further, if the information of the corresponding processing model cannot be determined based on the attribute information of the target application, such as when the input information is from a chat application or a conference application, it may be impossible to determine the target processing model used to process the input data. At this time, the user intention represented by the input data can be used to generate a second processing instruction, and the corresponding processing model or a third-party model service can be called according to the second processing instruction, so as to process the input data and obtain a target file that matches the attribute information of the target application and the user intention. If the attribute information of the target application is represented as an image generation application, the target file is an image file that matches the user intention. If the target application is a text generation application, the target file can be a text file. Specifically, the processing model can be a local model, such as an image generation model, a text generation model, a video generation model, an audio generation model, etc. configured in the current electronic device. The model service may be a service interface provided by the cloud that can be used to generate text, generate images, or generate videos, so that the cloud model can be called through the service interface to perform the corresponding processing tasks.
[0106] In one implementation of the embodiment of the present application, the data processing method further includes at least one of the following:
[0107] (1) In response to obtaining an adjustment operation on a target file, updating configuration parameters of the target file and model parameters of a processing model for generating the target file.
[0108] The target file is a file that matches the user's intention. For example, if the user's intention is to generate a corresponding image through the input description text, the target file is an image file. When outputting the image file, the user can optimize and adjust the image included in the image file, such as adjusting image parameters such as color and brightness in the image, and adjusting the character movements in the image. At this time, adjustment information corresponding to these adjustment operations will be recorded, and the image parameters in the image file will be updated based on the adjustment information, or the model parameters or algorithms of the image generation model used to generate the image file will be updated, so that when the updated image generation model is subsequently applied, an image file that better meets user needs can be output, thereby improving the user experience. Specifically, after obtaining the adjustment operation of the target file, the attribute parameters corresponding to the adjustment operation of the target file can be recorded, and the relevant attribute parameters in the feedback information of the user's corresponding guidance information can be combined to adjust these attribute parameters to the parameter values of the corresponding attribute parameters (such as adjusting the weight value of the attribute parameter in the entire processing model, or adjusting the value of the attribute parameter itself), so that when the subsequent user inputs similar input information, these adjusted attribute parameters are directly called and applied to the corresponding processing model, so that the subsequent processing can be processed based on the user's previous information, and the user can avoid repeated input and adjustment. For example, in the raw image scene, after obtaining the target image corresponding to the target file, if the user adjusts the image brightness parameter in the image and the image style in the image content, when the subsequent user inputs the information for generating other images again, the processing model matching the image brightness parameter and the image style can also be used for processing. For example, if the user selects the natural image style, the next time there is a need for raw images, a candidate image matching the natural image style will be directly generated, or the proportion of candidate images with natural image style will be increased, such as when four candidate images are generated, three of which can be images with natural image style.
[0109] (2) When the target file is a human-computer interaction interface, based on the user behavior data generated by using the human-computer interaction interface, the configuration parameters of the human-computer interaction interface and the model parameters of the processing model for generating the human-computer interaction interface are updated.
[0110] Among them, in the case of generating a human-computer interaction interface, the configuration parameters of the human-computer interaction interface and the model parameters of the processing model used to generate the human-computer interaction interface can be updated according to the user's usage data of the current human-computer interaction interface. For example, if the information input method of the currently generated human-computer interaction interface is keyboard input, and the user uses handwriting input when using the human-computer interaction interface, the information input method of the human-computer interaction interface can be updated to handwriting input, and the model parameters in the processing algorithm of the input mode in the processing model can be changed to relevant parameters of handwriting input, so that the input method of the subsequently generated human-computer interaction interface better meets the user's usage habits and improves the user experience.
[0111] The data processing method of the embodiment of the present application is described below using an application scenario of image generation.
[0112] In the application scenario of image generation, the user will input the description information of the image to be generated into the image generation application, and the image generation application will output the target image that matches the user's description information. In the process of processing the input data by the image generation application, multiple rounds of dialogue with the user can be conducted to output an image that matches the user's intention. In each round of human-computer dialogue, the user's input information is effectively associated to ensure that the processing model can understand the user's contextual input information and continuity requirements. At the same time, in the process of multiple rounds of dialogue, by analyzing the user's input information and selected information in multiple rounds of dialogue, a detailed user portrait is gradually constructed, including the user's preferences, habits, and needs, so that the user's intention can be analyzed based on the user portrait and the content of multiple rounds of dialogue to obtain a more accurate image.
[0113] See also Figure 2, which is a processing diagram of an image generation scenario provided in an embodiment of the present application. If the user first inputs a simple input message "Please draw a puppy", the output information at this time may be the response result, namely candidate image 1 and candidate image 2, and guidance information such as "What theme do you like? Nature or city" will also be output. If the user chooses a natural theme, the weight value of the attribute parameter of the natural theme in the image generation model will be increased, and the generated input information "Please draw a puppy under a natural theme" will be input into the adjusted image generation model. At this time, candidate image 3 and candidate image 4 will be output, and the current guidance information "What image style do you like? Realistic or abstract" will be output. If the user selects a realistic image style, the weight value of the attribute parameter of the realistic image style in the image generation model will be increased, and the input information of the puppy drawn in a realistic image style under a natural theme will be input into the adjusted image generation model. At this time, candidate image 5 and candidate image 6 will be output. If the user selects candidate image 5 at this time, the guidance information at this time may be "What is the puppy doing? Running or walking". If the user inputs the information of "running" representing the puppy's action, the image object in candidate image 5 will be analyzed, and combined with the puppy's action input by the user, the final information input into the image generation model may be "A golden retriever is running in a lush green park, smiling in the bright sunshine, with its glittering golden fur. In the background, people are having a picnic. Realistic style", so that the image generation model processes this and outputs the target images, namely image 7 and image 8. If the user uses the image generation application next time, even if he only inputs "please draw a horse", the image generation application will generate detailed input information based on the user's portrait and historical preferences, such as "a horse running on a lush green grassland, in the bright sunshine, smiling, with glittering golden fur, realistic style", and input the detailed information into the processing model to obtain the corresponding image. This method uses the system's understanding of the user, combined with the user's preferences and historical behavior, to generate a more personalized image that meets the user's expectations.
[0114] In this embodiment, through multiple rounds of dialogue with the user, in each round of dialogue, the processing algorithm can be combined to gradually guide the user to select a more detailed image description option, and at the same time quickly generate candidate images for the user to provide feedback on the preliminary image, so that the user can point out the image area that is satisfactory or dissatisfied with the current candidate image, so that the processing system adjusts the weights of the image parameters in the image feature library used for image generation according to the user's feedback information, optimizes the generation target parameters, and in each subsequent round of dialogue, the system will further optimize the guidance information according to the user's choice, and finally generate a detailed information describing the user's intention to draw the image. The user can also use the corresponding guidance information to generate the image, and can also modify the relevant content in the guidance information, so that the processing system can continue to optimize and process until a target image that meets the user's needs is generated.
[0115] Specifically, in the above processing, a feature library of different dimensions for image generation can be constructed, and initialization weight information is set for each image feature in the feature library. Specifically, the feature library can contain a variety of image features related to image generation, and these image features represent the attribute parameters of the image. For example, it can include style, color, composition, theme, etc. The initial weights of these features are set during initialization. According to the content of each round of dialogue, guidance information, and user feedback on the guidance information, the feature weights of the above features are dynamically adjusted. For example, if the user has a greater intention for realistic style images, the weight of realistic style image features will be increased. Correspondingly, in order to generate accurate images in this application scenario, the image generation model can be called to perform in-depth analysis of user input information and perform context perception to ensure that the generated content in each round of dialogue is more in line with user needs. For example, when the user describes a complex scene, the large model can understand and parse the key elements and generate corresponding images. The image processing model can also be used to clarify the user's intention, and the user's intention can be accurately perceived through the model, and feedback and suggestions that are more in line with user needs can be provided in each round of dialogue to improve the effectiveness of the dialogue and user satisfaction.
[0116] After the user enters the relevant input information, the large model can also be used to analyze the behavior and preferences of similar users and provide personalized guidance information. This guidance information can be based on the experience of successful image drawing of similar users, which helps to improve the satisfaction of the current user. And the guidance information can be optimized by combining the user portrait and the user's historical application data to ensure that it is highly relevant and personalized. In this application scenario, after the image is generated, the user's subsequent operations, such as modification, evaluation or use of the image, can be recorded. Through these data, the user portrait and dialogue process can be further optimized. Through continuous data analysis and feedback loops, the system's ability to understand user intentions is improved, ensuring that the generated images are more in line with user needs and improving user satisfaction.
[0117] In the design of user interface and human-computer interaction interface (abbreviated as UI design for convenience), a UI interface matching the user's intention can also be obtained based on the data processing method provided by this application. Figure 3 , which is a schematic diagram of a UI design architecture provided in an embodiment of the present application, showing the various functional modules and their interactive relationships in the UI design architecture, and can form a complete UI design and optimization process through modules such as large models, user data analysis, design and testing.
[0118] Specifically, users can input description information of the UI interface they need, and can also obtain relevant demand information about the user or a group of users similar to the user to obtain user portrait information, which can make the user's design intention clearer. User demand information can be obtained through user interviews, questionnaires, behavior tracking, etc. in the user demand module. User interviews are conducted through face-to-face or remote interviews with users to understand their needs, expectations, pain points, and other information regarding UI interface design; questionnaires are conducted using questionnaire tools to collect user feedback and opinions and obtain quantitative data support; behavior tracking is to track user operations and behaviors in the application and collect actual usage data.
[0119] In the application scenario of this UI design, the large model processing module can also be used to process the acquired user demand information to obtain user intentions, such as using the large model to analyze user feedback and behavior data, deeply understand user needs, and provide accurate demand analysis and establishment. The large model can also be used for context perception, such as processing user demand information for UI design and feedback information on the use of existing UI interfaces, as well as current application scenarios and usage environment data, to understand the complex scenarios and needs input by users and generate UI design solutions that better meet user intentions.
[0120] After obtaining user intention data based on the large model, the above data can be analyzed based on the user data analysis module. For example, the intelligent feedback analysis function can be used to automatically integrate and classify user counterexample information, and the key issues and demand information in the user feedback information can be extracted using natural language processing technology to generate a data report about the user. In this way, if you need to obtain user-related data, you can directly call the user's data report to obtain the user's intention information, as well as optimization suggestions for the UI interface, so that the relevant UI interface can be adjusted. The user behavior data can also be collected and analyzed through the quantitative data analysis function to assist the UI design end in better understanding user needs and behaviors. The user data can be monitored through the real-time demand monitoring function to obtain timely information on changes in user needs or intentions, quickly respond to changes in user intentions, and adjust the design of the UI interface. The feedback management function can also be used to automatically integrate user feedback information on existing UI interfaces or related information, and these feedback information can be generated into corresponding visual reports, that is, different feedback dimensions are marked in different formats to help the UI design end efficiently process and analyze the feedback data.
[0121] In the design module, UI interface design can be implemented based on the UI interface description information input by the user and the user intention data obtained by the above analysis. Based on the UI prototype generation function, a large model is used to quickly generate a UI prototype design plan, in which a variety of candidate plans can be generated for reference by relevant designers to improve design efficiency and quality. The personalized optimization suggestion function can also provide personalized design optimization suggestions matching the user based on the user portrait and phase data to improve the relevance of the design and user satisfaction.
[0122] Correspondingly, the design architecture also includes a test module. Based on the rapid user testing function in the test module, user testing can be quickly performed in the UI design stage to collect user feedback and verify the effectiveness of the design scheme. The A / B test function can also be used to compare the user performance of different design schemes and select the best UI design scheme for implementation. Finally, the results of all the above modules can be integrated into the UI design module in the design architecture to form a final UI design scheme, and continuously optimized and improved. Specifically, each time a UI interface candidate scheme is generated, corresponding guidance information can be provided to assist users in optimizing and adjusting the current candidate scheme, and the weight value of the current design element can be adjusted according to the user's feedback information on the guidance information. For example, if the user's feedback information indicates that the user likes brightly colored elements, the weight value of the warm-toned elements can be increased, and the weight value of the cold-toned elements can be reduced accordingly, so that the subsequent generated UI design scheme can better meet the needs of users and improve the user experience.
[0123] In the embodiment of the present application, a data processing device is also provided. Figure 4 , the device comprises:
[0124] An output unit 401 is used for outputting a response result for the first input data and guidance information for the response result in response to obtaining the first input data, wherein the guidance information is used to update an attribute parameter of the response result;
[0125] A first generating unit 402, configured to generate second input data based on feedback information of the guidance information;
[0126] The second generating unit 403 is configured to generate a target file matching the user intention based on the user intention represented by the second input data.
[0127] Optionally, the output unit includes at least one of the following subunits:
[0128] a first processing subunit, configured to obtain a first processing instruction for the first input data, and generate and process the first input data using a first processing model matching the first processing instruction, so as to output the response result and guidance information for the response result;
[0129] a second processing subunit, configured to obtain a first processing instruction for the first input data, generate and process the first input data using a first processing model matching the first processing instruction to output the response result, and output guidance information for the response result using a second processing model matching the first processing instruction;
[0130] a third processing subunit, configured to generate and process first target data in the first input data based on the type of the first input data and the represented user intention, so as to output the response result and guidance information for the response result;
[0131] The fourth processing sub-unit is used to integrate the first input data and the context data of its source application to obtain the third input data, and use the generative model to generate the third input data to output the response result and guidance information for the response result.
[0132] Optionally, obtaining a first processing instruction for the first input data includes at least one of the following:
[0133] generating the first processing instruction in response to obtaining a trigger operation of a first type of control acting on a first type of application, wherein the first type of application is an application capable of responding to the first input data;
[0134] generating the first processing instruction in response to obtaining a trigger operation for first input data input to a second type of application, the second type of application being an application that can respond to or cannot respond to the first input data;
[0135] In response to obtaining first input data input to a third type of application, generating a corresponding first processing instruction based on application information of the third type of application;
[0136] A corresponding first processing instruction is generated based on the type of the first input data and / or application information of a source application.
[0137] Optionally, the first generation unit includes at least one of the following subunits:
[0138] a first integration subunit, configured to integrate the first input data and second target data in the guidance information determined based on the feedback information to obtain the second input data;
[0139] a second integration subunit, configured to integrate the feedback information different from the guidance information and the first input data to obtain the second input data;
[0140] The first generating subunit is used to generate second input data based on feedback information of the guide information and a conversion result of the response result.
[0141] Optionally, the first generating unit includes:
[0142] a first output subunit, configured to generate fourth input data based on first feedback information of first guidance information of a first response result to the first input data, and output a second response result to the fourth input data and second guidance information to the second response result;
[0143] a second generating subunit, configured to generate fifth input data based on second feedback information to the second guiding information, and output a third response result to the fifth input data and third guiding information to the third response result;
[0144] The third generating subunit is configured to generate the second input data based on the third feedback information of the third guiding information.
[0145] Optionally, the first generating subunit includes at least one of the following subunits:
[0146] a fourth generating subunit, configured to obtain target user portrait information corresponding to the first input data, and generate the second input data based on the user portrait information and feedback information on the guidance information, wherein the target user portrait includes group user portrait information or user portrait information of the target user;
[0147] The fifth generating subunit is used to obtain current environment information of the electronic device, and generate the second input data with reference to the environment information and feedback information of the guiding information.
[0148] Optionally, the first generating unit includes at least one of the following subunits:
[0149] A sixth generating subunit, configured to generate second input data based on the obtained selection operation of the recommended data for the third target data in the response result and / or the edit operation for the second target data;
[0150] The seventh generating subunit is used to generate second input data based on the obtained configuration operation of the configuration option in the configuration window of the response result, where the configuration option is used to update the target data in the response result.
[0151] Optionally, the second generating unit includes:
[0152] an eighth generating subunit, configured to generate a second processing instruction based on the attribute information of the target application and / or the user intention represented by the second input data;
[0153] The ninth generation sub-unit is used to generate and process the second input data using a processing model or a model service corresponding to the second processing instruction to obtain a target file that matches the attribute information and the user's intention.
[0154] Optionally, the device further comprises at least one of the following:
[0155] A first updating unit, configured to update, in response to obtaining an adjustment operation on the target file, configuration parameters of the target file and model parameters of a processing model for generating the target file;
[0156] When the target file is a human-computer interaction interface, the second updating unit updates configuration parameters of the human-computer interaction interface and model parameters of a processing model for generating the human-computer interaction interface based on the obtained user behavior data generated by using the human-computer interaction interface.
[0157] It should be noted that the specific implementation of each unit and sub-unit in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.
[0158] In another embodiment of the present application, a readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the data processing method as described in any one of the above items is implemented.
[0159] In another embodiment of the present application, an electronic device is further provided, and the electronic device may include:
[0160] A memory, used to store applications and data generated by the operation of the applications;
[0161] The processor is used to execute the application program to implement the data processing method as described in any one of the above.
[0162] It should be noted that the specific implementation of the processor in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.
[0163] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.
[0164] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.
[0165] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
[0166] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A data processing method, comprising: In response to obtaining first input data, outputting a response result for the first input data and guidance information for the response result, wherein the guidance information is used to update an attribute parameter of the response result; generating second input data based on feedback information of the guidance information; A target file matching the user intention is generated based on the user intention represented by the second input data.
2. The method according to claim 1, wherein the outputting of the response result to the first input data and the guidance information to the response result comprises at least one of the following: Obtaining a first processing instruction for the first input data, and performing generation processing on the first input data using a first processing model matching the first processing instruction to output the response result and guidance information for the response result; Obtaining a first processing instruction for the first input data, performing generation processing on the first input data using a first processing model matching the first processing instruction to output the response result, and outputting guidance information for the response result using a second processing model matching the first processing instruction; Generate and process first target data in the first input data based on the type of the first input data and the represented user intention, so as to output the response result and guidance information for the response result; The first input data and context data of its source application are integrated to obtain the third input data, and the third input data is generatively processed using a generative model to output the response result and guidance information for the response result.
3. The method according to claim 2, wherein obtaining a first processing instruction for the first input data comprises at least one of the following: generating the first processing instruction in response to obtaining a trigger operation of a first type of control acting on a first type of application, wherein the first type of application is an application capable of responding to the first input data; generating the first processing instruction in response to obtaining a trigger operation for first input data input to a second type of application, the second type of application being an application that can respond to or cannot respond to the first input data; In response to obtaining first input data input to a third type of application, generating a corresponding first processing instruction based on application information of the third type of application; A corresponding first processing instruction is generated based on the type of the first input data and / or application information of a source application.
4. The method according to claim 1, wherein generating the second input data based on the feedback information of the guidance information comprises at least one of the following: Integrate the first input data and the second target data in the guidance information determined based on the feedback information to obtain the second input data; integrating feedback information different from the guidance information and the first input data to obtain the second input data; Second input data is generated based on feedback information of the guide information and a conversion result of the response result.
5. The method according to claim 1 or 4, wherein generating the second input information based on the feedback information of the guidance information comprises: generating fourth input data based on first feedback information of first guidance information of a first response result to the first input data, and outputting a second response result to the fourth input data and second guidance information to the second response result; generating fifth input data based on second feedback information to the second guidance information, and outputting a third response result to the fifth input data and third guidance information to the third response result; The second input data is generated based on third feedback information of the third guidance information.
6. The method according to claim 1 or 4, wherein generating the second input data based on the feedback information of the guidance information comprises at least one of the following: Obtaining target user portrait information corresponding to the first input data, and generating the second input data based on the user portrait information and feedback information on the guidance information, wherein the target user portrait includes group user portrait information or user portrait information of the target user; The current environment information of the electronic device is obtained, and the second input data is generated with reference to the environment information and feedback information of the guide information.
7. The method according to claim 1, wherein generating the second input data based on the feedback information of the guidance information comprises at least one of the following: generating second input data based on the obtained selection operation of the recommended data for the third target data in the response result and / or the edit operation for the second target data; Second input data is generated based on the configuration operation obtained for the configuration options in the configuration window of the response result, where the configuration options are used to update the target data in the response result.
8. The method according to claim 1, wherein generating a target file matching the user intention based on the user intention represented by the second input data comprises: generating a second processing instruction based on the attribute information of the target application and / or the user intention represented by the second input data; The second input data is generated and processed using a processing model or a model service corresponding to the second processing instruction to obtain a target file that matches the attribute information and the user's intention.
9. The method according to claim 1, further comprising at least one of the following: In response to obtaining an adjustment operation on the target file, updating configuration parameters of the target file and model parameters of a processing model for generating the target file; In the case where the target file is a human-computer interaction interface, configuration parameters of the human-computer interaction interface and model parameters of a processing model for generating the human-computer interaction interface are updated based on the obtained user behavior data generated by using the human-computer interaction interface.
10. A data processing device, comprising: an output unit, configured to output a response result for the first input data and guidance information for the response result in response to obtaining the first input data, wherein the guidance information is used to update an attribute parameter of the response result; A first generating unit, configured to generate second input data based on feedback information of the guiding information; The second generating unit is used to generate a target file matching the user intention based on the user intention represented by the second input data.
