Interaction processing method and device
By identifying and adjusting user intentions, using target reference information to process user input, solve the problem that large language model output does not meet the needs and improve user experience.
Patent Information
- Application Number
- CN202411194576.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-03
AI Technical Summary
When the large language model processes user input, the output instruction execution results may not meet user needs, resulting in poor user experience.
By identifying the target input data, a first user intention is obtained, and processing is performed based on the target reference information, a second user intention is obtained, and a matching response operation is performed. The method includes obtaining content information, attribute information, user portrait, long-term memory, short-term memory, evaluation information, etc. to adjust user intentions and ensure that it is more in line with user needs.
Improve user interaction experience, ensure that responsive operations are more in line with user needs, and improve user satisfaction.
Smart Images

Figure CN119065558A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to an interactive processing method and device. Background Art
[0002] Currently, a large language model (LLM) is usually used to process user input data to obtain corresponding instructions, such as starting an application or pushing related messages.
[0003] However, the command execution results output by this large language model may not meet the needs of users, resulting in a poor user experience. Summary of the invention
[0004] In view of this, the present application provides an interactive processing method and device as follows:
[0005] An interactive processing method, comprising:
[0006] Identify the obtained target input data to obtain a first user intention;
[0007] Based on the target reference information, the first user intention is processed to obtain a second user intention;
[0008] executing a response operation matching the second user intention;
[0009] The target reference information corresponding to different target input data is different or the same.
[0010] The above method preferably further comprises at least one of the following:
[0011] Obtaining content information and / or attribute information of the target input data, and acquiring the target reference information based on the content information and / or the attribute information, wherein the target reference information is related to or unrelated to the target input data;
[0012] Evaluation information for the first user intention is obtained, and corresponding target reference information is acquired based on the evaluation information.
[0013] In the above method, preferably, obtaining the target reference information based on the content information and / or the attribute information includes at least one of the following:
[0014] Determine the interaction scenario represented by the content information, and obtain target reference information matching the interaction scenario;
[0015] Determine an interactive object in the content information, and obtain target reference information matching the interactive object;
[0016] Determine a data type of the target input data represented by the attribute information, and obtain target reference information matching the data type;
[0017] Determine source information of the target input data represented by the attribute information, and obtain target reference information matching the source information;
[0018] Determine an input position of the target input data represented by the attribute information, and obtain target reference information matching the input position;
[0019] Determine the data volume of the target input data represented by the attribute information, and obtain target reference information matching the data volume.
[0020] In the above method, preferably, obtaining the target reference information based on the content information and / or the attribute information includes at least one of the following:
[0021] Acquire user portrait information associated with the content information and / or the attribute information, and use the user portrait information as the target reference information;
[0022] Acquire long-term memory associated with the content information and / or the attribute information, and use the long-term memory as the target reference information;
[0023] Acquire short-term memory associated with the content information and / or the attribute information, and use the short-term memory as the target reference information.
[0024] In the above method, preferably, obtaining corresponding target reference information based on the evaluation information includes at least one of the following:
[0025] In a case where the evaluation information includes first feedback information of a target user on the first user intention, using preset reference information matching the first user intention as the target reference information;
[0026] In a case where the evaluation information includes second feedback information of the target user on the first user's intention, using the second feedback information directly as the target reference information, or generating the target reference information based on the second feedback information;
[0027] In a case where the evaluation information includes supplementary information of the target user's intention to the first user, the target reference information is generated based on the supplementary information.
[0028] In the above method, preferably, the step of processing the first user intent based on the target reference information to obtain the second user intent includes at least one of the following:
[0029] adding an attribute tag corresponding to the target reference information or a prompt word corresponding to the target reference information to the first user intention, so as to obtain the second user intention;
[0030] Using the target reference information, generating or replacing the first user intention to obtain the second user intention;
[0031] The first user intention is processed based on a processing strategy corresponding to the type and / or source of the target reference information to obtain the second user intention.
[0032] The above method, preferably, wherein the first user intention is processed with a corresponding processing strategy based on the type and / or source of the target reference information to obtain the second user intention, includes at least one of the following:
[0033] In a case where the target reference information is user portrait information of a target user, generating or updating the first user intent based on the user portrait information to obtain the second user intent;
[0034] In a case where the target reference information only includes long-term memory, adjusting or editing the first user intention based on the long-term memory to obtain the second user intention;
[0035] In the case where the target reference information includes short-term memory and long-term memory, the first user intention is updated based on the short-term memory, or, in the case where the first user intention cannot be updated using the short-term memory, the first user intention is updated using the long-term memory to obtain the second user intention;
[0036] In a case where the target reference information is evaluation information for the first user intention, replacing the first user intention based on the evaluation information or performing intent understanding processing on the evaluation information and the first user intention again to obtain the second user intention;
[0037] In a case where the target reference information comes from a target application, updating the first user intent based on application information of the target application to obtain the second user intent;
[0038] In a case where the target reference information comes from a target space environment, a model generation process is performed on the first user intention based on environmental parameters of the target space environment to obtain the second user intention.
[0039] In the above method, preferably, executing a response operation matching the second user intention includes at least one of the following:
[0040] Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding application tool to perform a target response operation based on the execution strategy;
[0041] Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding agent to perform a target response operation based on the execution strategy;
[0042] A corresponding execution strategy is determined based on the second user intention and the operating status information of the electronic device, so as to call a corresponding application tool or agent to perform a target response operation based on the execution strategy.
[0043] The above method preferably further comprises at least one of the following:
[0044] updating the second user intention based on the obtained third feedback information of the execution result of the response operation;
[0045] The second user intention is updated based on the input operation acting on the interactive interface of the target agent.
[0046] An interactive processing device, comprising:
[0047] A first intention obtaining unit, used for identifying the obtained target input data to obtain a first user intention;
[0048] a second intention obtaining unit, configured to process the first user intention based on target reference information to obtain a second user intention;
[0049] an operation execution unit, configured to execute a response operation matching the second user's intention;
[0050] The target reference information corresponding to different target input data is different or the same.
[0051] A computer device / equipment / system comprises: a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement any one of the above-mentioned interaction processing methods.
[0052] A computer-readable storage medium stores a computer program / instruction, which implements any one of the above-mentioned interactive processing methods when executed by a processor.
[0053] A computer program product comprises a computer program / instruction, wherein when the computer program / instruction is executed by a processor, the computer program / instruction implements any one of the above-mentioned interaction processing methods.
[0054] It can be seen from the above technical solution that in an interactive processing method and device disclosed in the present application, after obtaining the first user intention by identifying the obtained target input data, the response operation matching the first user intention is not directly executed, but the first user intention is first processed based on the target reference information to obtain the second user intention, so that the response operation matching the second user intention can be executed. It can be seen that in the present application, the directly obtained user intention is first adjusted according to the target reference information, so that a user intention that better meets the user's needs can be obtained, and then the response operation is executed accordingly, which can better meet the user's needs, thereby improving the user's interactive experience. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] In order to more clearly illustrate the technical solutions of 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 only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0056] Figure 1 A flowchart of an interactive processing method provided in an embodiment of the present application;
[0057] Figure 2 : The figure in the figure is an example of an agent deployed locally in an electronic device;
[0058] Figure 3 A schematic diagram of the structure of an interactive processing device provided in an embodiment of the present application;
[0059] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;
[0060] Figure 5 This is a schematic diagram of the module design applicable to mobile phones in this application;
[0061] Figure 6 This is a flowchart of the implementation of this application applicable to mobile phones;
[0062] Figure 7 This is an example diagram of user interaction when this application is applied to a mobile phone;
[0063] Figure 8 This is an example diagram of identifying user intent in the prior art;
[0064] Fig. 9 This is an example diagram of adjusting the user's intention after recognizing the user's intention when the application is applied to a mobile phone;
[0065] Fig.10 A schematic diagram of identifying user intentions and performing operations through an intelligent agent in the prior art;
[0066] Fig.11 This is an example diagram of adjusting user intent through local portraits when this application is applied to mobile phones;
[0067] Fig.12 This is a flowchart of the interaction between the modules when this application is applied to a mobile phone;
[0068] Fig.13 A schematic diagram of a response operation of executing a user's intention through an intelligent agent when the present application is applied to a mobile phone;
[0069] Fig.14 A schematic diagram of identifying user intentions and performing operations through an intelligent agent when this application is applied to a mobile phone;
[0070] Fig.15 This is another schematic diagram of identifying user intentions and performing operations through an intelligent agent when the present application is applied to a mobile phone. DETAILED DESCRIPTION
[0071] 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.
[0072] refer to Figure 1 As shown, it is a flowchart of an interactive processing method provided in an embodiment of the present application, which can be applied to electronic devices capable of data processing, such as mobile phones, tablet computers, computers or servers, etc. The technical solution in this embodiment is mainly used to improve the user's interactive experience.
[0073] Specifically, the method in this embodiment may include the following process:
[0074] Step 101: Identify the obtained target input data to obtain a first user intention.
[0075] The target input data may be input data in any form, such as text input data, handwriting input data, gesture input data, voice input data, image input data, etc.
[0076] In one implementation, the target input data in this embodiment can be obtained through an interactive interface. For example, the user performs input operations in the form of text, handwriting, gestures, voice, images, etc. on the interactive interface to obtain the target input data. For example, the interactive interface can be output to the user through a mobile phone, and the user enters the text "Tianjin" on the interactive interface, and the text "Tianjin" is the target input data.
[0077] In one implementation, in step 101, an intent understanding model may be used to perform intent recognition to obtain the first user intent.
[0078] Among them, the intent understanding model can be a trained machine learning model, and the intent understanding model can process the target input data to output the identified first user intent.
[0079] Step 102: Based on the target reference information, the first user intention is processed to obtain a second user intention.
[0080] Among them, the target reference information is related to the target input data.
[0081] For example, the target reference information may be user portrait information, other input data, prompt words, current environment information or context information, etc.
[0082] Specifically, in this embodiment, the first user intention may be adjusted based on the target reference information to obtain the second user intention.
[0083] The second user intent can better represent the user needs of the user to whom the target input data belongs than the first user intent. For example, the number of words expressing the intent in the second user intent is greater than the number of words expressing the intent in the first user intent, and accordingly, the second user intent is more accurate than the first user intent in expressing the intent.
[0084] For example, the first user intention is "introducing Tianjin", and the first user intention is adjusted using the context information "study tour" to obtain the second user intention "introducing university information in Tianjin", which can better reflect the user's operational intention.
[0085] Step 103: Execute a response operation matching the second user intention.
[0086] For example, according to the second user intention of “introducing university information in Tianjin”, information about universities related to Tianjin is searched in a database or the Internet.
[0087] Specifically, in step 103, an execution strategy may be first determined based on the second user intention, and a corresponding response operation may be performed according to the determined execution strategy.
[0088] There may be a variety of different execution strategies, for example, the execution strategy may be: a strategy for executing a response operation immediately; or the execution strategy may be: a strategy for executing a response operation according to a pre-set execution plan.
[0089] It should be noted that the target reference information corresponding to different target input data is different or the same. In this embodiment, the corresponding target reference information can be obtained for the target input data. For example, the corresponding target reference information can be obtained for different types of target input data, or target input data in different scenarios. For another example, in this embodiment, for the target input data, corresponding reference information can be selected from multiple preset reference information as the target reference information. For another example, in this embodiment, for the target input data, the corresponding scene information or environmental information can be used as the target reference information.
[0090] It can be seen from the above technical solution that in an interactive processing method provided by an embodiment of the present application, after obtaining the first user intention by identifying the obtained target input data, the response operation matching the first user intention is not directly executed, but the first user intention is first processed based on the target reference information to obtain the second user intention, so that the response operation matching the second user intention can be executed. It can be seen that in this embodiment, the directly obtained user intention is first adjusted according to the target reference information, so that a user intention that better meets the user's needs can be obtained, and then the response operation is executed accordingly, which can better meet the user's needs, thereby improving the user's interactive experience.
[0091] In one implementation, in this embodiment, the target reference information can be obtained in the following manner:
[0092] First, content information and / or attribute information of target input data is obtained; then, based on the content information and / or attribute information, target reference information is acquired.
[0093] It should be noted that the target reference information is related to or not related to the target input data. Specifically, it can be understood that: the target reference information can be the default reference information, or the target reference information can be the reference information adapted to the current environment or interaction scenario, in which case the target reference information is not related to the target input data. Alternatively, the target reference information can be determined based on the target input data, in which case the target reference information is related to the target input data.
[0094] Among them, content information refers to the input content in the target input information. Based on this, in this embodiment, the input content can be extracted from the target input data, and then the current interaction scene can be determined based on the input content, and the target reference information can be obtained according to the interaction scene. For example, the current interaction scene is a scene for discussing study tours, and the target reference information such as "university" or "school" is determined based on this, or the method of obtaining the target reference information is determined based on this, such as obtaining the target reference information from the preset reference information or obtaining the target reference information according to the scene type.
[0095] The attribute information refers to the type information, source information, input time, input location, data volume of the input data, etc., which characterize the target input data. The type information can be text type, voice type, image type, URL type, or code type, etc.; the source information refers to the source identifier of the person or device to which the target input data belongs. Based on this, in this embodiment, the target input data can be parsed to obtain at least one attribute information, and then the target reference information is obtained according to the attribute information.
[0096] Based on the above implementation, in this embodiment, when acquiring target reference information based on content information and / or attribute information, it can be implemented in any one or more of the following ways:
[0097] In one implementation, in this embodiment, the interaction scene represented by the content information can be determined, and target reference information matching the interaction scene can be acquired.
[0098] For example, the content information may be the interactive content in a social media application. Correspondingly, the interactive scenario may be a social scenario between friends or family members through a social media application. Based on this, the target reference information matched according to the social scenario may be the context information of the interaction or the user portrait information of both parties.
[0099] For example, the content information may be the meeting content in a conference application. Correspondingly, the interaction scenario may be: a work scenario between the conference application and the participants. Based on this, the target reference information matched according to the work scenario may also be work-related meeting minutes, related meeting plans, or meeting-related content, etc.
[0100] For another example, the content information may be the transaction communication content in a shopping application. Accordingly, the interaction scenario may be a transaction scenario for communicating with a merchant. Based on this, the target reference information matched according to the transaction scenario may be information related to the transaction object or user portrait information of the buyer.
[0101] For example, content information can control the control content in an application or control interface. Correspondingly, the interaction scenario can be: a device control scenario used by users to implement device settings or controls. Based on this, the target reference information matched according to the device control scenario can be the user's setting preferences or the device's historical usage data, etc.
[0102] For example, content information can call the calling content or functional content of the function provided by the device. Correspondingly, the interaction scenario can be: a scenario in which the user uses to call the functional service that the device can provide. Based on this, the target reference information matched according to the scenario can be: user historical usage data or device configuration update data, etc.
[0103] In one implementation, in this embodiment, the interactive object in the content information may be determined, and then the target reference information matching the interactive object may be acquired.
[0104] Among them, the interaction object can be: people in social interaction, participants in conference interaction, transaction objects in transaction communication, controlled objects (such as devices, applications, etc.), required functional service objects, etc. Then the target reference information matched with the interaction object can be the user portrait information of the user or the user portrait information of the two interaction parties, or the group portrait information of the group to which one or both interaction parties belong; it can be the meeting minutes related to the participants, the related meeting plan, or the content related to the meeting, etc.; it can be the information related to the transaction object or the user portrait information of the buyer; or it can be the controlled historical data of the controlled object, etc.
[0105] In one implementation, in this embodiment, the data type of the target input data represented by the attribute information can be determined, and then the target reference information matching the data type can be obtained.
[0106] The data types of the target input data may include: text type, voice type, image type, website link type, code type, etc.
[0107] It should be noted that different data types can match different target reference information. For example, the text type can match the target reference information of the processing options used to process the text; for another example, the text type can match the target reference information related to the style or color that makes the text more in line with the user's habits or preferences, or the modified tone; for another example, the image type can match the additional attribute information that matches the theme or style of the image as the target reference information, etc.
[0108] In one implementation, in this embodiment, source information of the target input data represented by the attribute information may be determined, and then target reference information matching the source information may be determined.
[0109] The source information can be understood as: the user (person), device, etc. from which the target input data comes. Based on this, if the source information indicates that the target input data comes from a person, then the target reference information of the user portrait information or historical usage data related to the person can be obtained according to the source information; or, if the source information indicates that the target input data comes from a device, then the configuration data or log information of the device can be obtained according to the source information as the target reference information.
[0110] In one implementation, in this embodiment, the input position of the target input data represented by the attribute information can be determined, and then the target reference information matching the input position can be obtained.
[0111] Among them, the input location here can be: the spatial environment where the electronic device is located when the target input data is input, such as indoors or outdoors, home or work area, city, country, etc. Correspondingly, the target reference information matching the input location can be: historical related data matching the current spatial environment or data that can reflect the local characteristics and style.
[0112] Alternatively, the input position here may also be: an input box for the application being operated on the electronic device, that is, the input position may be the position of the input box, and accordingly, the target reference information matching the input position may be: information related to the application, for example, specific information related to the application that matches the target input data.
[0113] In one implementation, in this embodiment, the data volume of the target input data represented by the attribute information can be determined, and then the target reference information matching the data volume can be obtained.
[0114] The data volume here may be the data volume of any parameter related to the target input data, for example, the data volume may be the number of keywords in the target input data, or the data volume may be the resolution of an image in the target input data.
[0115] In one implementation, in this embodiment, user portrait information associated with content information and / or attribute information may be obtained, and then the user portrait information may be used as target reference information.
[0116] Specifically, in this embodiment, keyword extraction or parsing can be performed on content information and / or attribute information to obtain user portrait information related thereto, such as portrait information of "user preference for gourmet food or preference for dieting", and then the user portrait information is used as target reference information.
[0117] In one implementation, in this embodiment, long-term memory associated with content information and / or attribute information may be acquired and used as target reference information.
[0118] Among them, the long-term memory can be read from the historical information of the user or device from which the target input data comes. For example, the information most frequently processed by the user through the device in the past year is food, and accordingly, food is the user's long-term memory, and the corresponding target reference information can be "food".
[0119] In one implementation, in this embodiment, short-term memory associated with content information and / or attribute information may be acquired and used as target reference information.
[0120] Among them, short-term memory can be read from the historical information of the user or device from which the target input data comes. For example, the user is currently chatting with friends about "school" through a social application on the device. Accordingly, the school category is the user's short-term memory, and the corresponding target reference information can be "university".
[0121] In one implementation, in this embodiment, the target reference information can be obtained in the following manner:
[0122] First, evaluation information for the first user intention is obtained, and then corresponding target reference information is acquired based on the evaluation information.
[0123] The evaluation information for the first user's intention may be: feedback information from the user on the first user's intention, for example, the feedback information is: feedback information that approves the first user's intention or feedback information that disapproves the first user's intention. For another example, the feedback information is: feedback information of further suggestions from the user on the first user's intention. For another example, the feedback information is: feedback information of adjustments or corrections from the user on the first user's intention. Specifically, in this embodiment, the user's feedback operation may be obtained through the interactive interface and the evaluation information may be obtained accordingly.
[0124] For example, the first user intention may be: "Introduce Tianjin", and the user evaluates the first user intention through the interactive interface, "This is not right" or "I want to see what delicious food there is". Based on this, for the evaluation information of "This is not right", the target reference information of changing to other regions such as "Beijing" is obtained, and the first user intention can be adjusted accordingly to obtain the second user intention such as "Introduce Beijing"; and for the evaluation information of "I want to see what delicious food there is", the target reference information such as "food" is obtained, and the first user intention can be adjusted accordingly to obtain the second user intention such as "Introduce Tianjin's food".
[0125] Based on the above implementation, in this embodiment, when obtaining corresponding target reference information based on evaluation information, it can be implemented in any one or more of the following ways:
[0126] In one implementation, in this embodiment, when the evaluation information includes first feedback information of the target user on the first user intention, the preset reference information matching the first user intention may be used as the target reference information.
[0127] The first feedback information may be: feedback information of a positive evaluation made by the target user in response to the first user's intention. For example, feedback information of approving the recommended content represented by the first user's intention. Based on this, in this embodiment, the preset reference information matching the first user's intention is used as the target reference information.
[0128] In one implementation, in this embodiment, when the evaluation information includes second feedback information of the target user on the intention of the first user, the second feedback information can be directly used as the target reference information, or the target reference information can be generated based on the second feedback information.
[0129] Among them, the second feedback information can be: feedback information of positive evaluation made by the target user in response to the first user intention. For example, feedback information that does not approve the recommended content represented by the first user intention. Based on this, in this embodiment, the second feedback information can be used as target reference information, for example, "wrong" can be used as target reference information; or, new target reference information can be generated based on the second feedback information, for example, according to the feedback information of "wrong", the target reference information of "change city" is generated. Accordingly, the first user intention can be adjusted according to the target reference information to obtain the second user intention. For example, the first user intention of "introducing Tianjin" is adjusted to the second user intention of "introducing Beijing" according to "changing cities".
[0130] In one implementation, in this embodiment, when the evaluation information includes the supplementary information of the target user to the first user's intention, the target reference information may be generated based on the supplementary information.
[0131] Among them, the supplementary information may be: the target user's supplementary limitation or supplementary content for the first user's intention. For example, the supplementary information may be: similar content "also Beijing" added to "Introducing Tianjin" or the added information "Nankai District" with a limited scene or scope. Accordingly, the target reference information "Add Beijing's city" or "Further introduce Nankai District" is generated accordingly. Accordingly, the first user intention can be adjusted according to the target reference information to obtain the second user intention. For example, the first user intention of "Introducing Tianjin" is adjusted to the second user intention of "Introducing Nankai District of Tianjin" according to "Further introducing Nankai District".
[0132] Based on the above implementation scheme, in step 102 of this embodiment, when processing the first user intention based on the target reference information, it can be implemented by at least one implementation method:
[0133] In one implementation, in this embodiment, an attribute tag corresponding to the target reference information or a prompt word corresponding to the target reference information may be added to the first user intention to obtain the second user intention.
[0134] Among them, the target reference information can be obtained by the implementation method in the foregoing text. The attribute label here can be: a behavioral attribute or a descriptive attribute generated based on the target reference information. The behavioral attribute can be an action attribute of a control type or an execution type, for example, action attributes such as opening, closing, uninstalling, installing, increasing, decreasing, etc. The descriptive attribute can be a descriptive attribute that makes it easier for the model to understand or execute, for example, descriptive attributes such as polishing, extension, limitation, and translation. The prompt word can be a prompt content that further expands the first user intention, for example, expanding a word into a sentence, or adding an execution instruction for a keyword. Specifically, the prompt word can be a descriptive word extracted from the target reference information. Based on this, in this embodiment, the corresponding attribute label or prompt word can be added to the first user intention according to the attribute label or prompt word corresponding to the target reference information, so that a new user intention, that is, the second user intention, can be obtained.
[0135] In one implementation, in this embodiment, the target reference information may be used to generate or replace the first user intent to obtain the second user intent.
[0136] Among them, in this embodiment, the target reference information and the first user intention are input into the model of artificial intelligence generated content AIGC (Artificial Intelligence Generated Content), and the AIGC model performs generation processing, so that the second user intention output by the AIGC model can be obtained.
[0137] For example, the target reference information "add cities of Beijing" and the first user intention "introduce Tianjin" are input into the AIGC model, so that the AIGC model can output the second user intention "introduce Tianjin and Beijing".
[0138] Alternatively, the corresponding data in the first user intention can be replaced with the relevant data in the target reference information to obtain the second user intention. For example, the target user is accustomed to turning on the eye protection mode when using the electronic device. When the electronic device recognizes that the first user intention represented by the "display mode" input by the user is "changing the current display mode to the high brightness mode", the "high brightness mode" can be replaced based on the user habits or preferences provided by the user portrait information to obtain the second user intention of "changing the current display mode to the eye protection mode".
[0139] In one implementation, in this embodiment, the first user intention may be processed based on a processing strategy corresponding to the type and / or source of the target reference information to obtain the second user intention.
[0140] The type of the target reference information may be the type of user portrait, the type of long-term memory and short-term memory, the type of evaluation or feedback information on the first user's intention, the type of text data, the type of image data, etc. The source of the target reference information may be understood as: the method of obtaining the target reference information, such as the multiple methods of obtaining the target reference information in the previous text.
[0141] Based on this, the processing strategy corresponding to the type and / or source of the target reference information may include: an identification processing strategy for generating or understanding the target reference information and the first user intention using different models.
[0142] For example, the processing strategy may be: a processing strategy for editing the first user intent using the target reference information, and the processing strategy may be implemented based on an operation on an interactive interface. For example, the target reference information and the first user intent are output on the interactive interface and provided to the target user, and the target user edits the first user intent on the interactive interface according to the target reference information to obtain the second user intent.
[0143] For another example, the processing strategy may be: using the target reference information to adjust the weight, priority, or order of the intent keywords in the first user's intent. The processing strategy may be executed in the background or implemented based on the configuration operation on the interactive interface. For example, in this embodiment, the intent keywords in the first user's intent are directly replaced according to the keywords in the target reference information to obtain the second user's intent; or, in this embodiment, the target reference information and the first user's intent are output on the interactive interface to the target user, and the target user adjusts the order of the intent keywords in the first user's intent on the interactive interface according to the keywords in the target reference information to obtain the second user's intent.
[0144] For another example, the processing strategy may be: using the target reference information to replace the first user intention. For example, in this embodiment, in a scenario where the first user intention represents the target user performing security or privacy processing, the target reference information replaces the first user intention to obtain the second user intention.
[0145] Specifically, in this embodiment, when the first user intention is processed based on the processing strategy corresponding to the type and / or source of the target reference information, it can be implemented by at least one of the following implementation methods:
[0146] In one implementation, when the target reference information is user portrait information of the target user, the first user intent may be generated or updated based on the user portrait information to obtain the second user intent.
[0147] For example, the target reference information is user portrait information associated with the content information and / or attribute information of the target input data, such as the portrait information of "user preference for food". In this way, based on the user portrait information, the intent keywords in the first user intent can be regenerated or replaced to obtain the second user intent. For example, for the first user intent of "introducing Tianjin", "Tianjin" is replaced to obtain the second user intent of "introducing food".
[0148] In one implementation, when the target reference information only includes long-term memory, the first user intention can be adjusted or edited based on the long-term memory to obtain the second user intention.
[0149] For example, the target reference information only includes the long-term memory "food category", based on which, the first user intention can be edited based on the long-term memory to obtain the second user intention "introducing Tianjin's food".
[0150] In one implementation, when the target reference information includes short-term memory and long-term memory, the first user intention is updated based on the short-term memory, or, when the first user intention cannot be updated using the short-term memory, the first user intention is updated using the long-term memory to obtain the second user intention.
[0151] For example, the target reference information includes the long-term memory "food category" and the short-term memory "university". Based on this, "university" can be used first to update the first user intention "introducing Tianjin" to obtain the second user intention "introducing universities in Tianjin"; if the short-term memory cannot be used to update the first user intention, such as if the short-term memory is not recognized, the long-term memory "food category" can be used to update the first user intention "introducing Tianjin" to obtain the second user intention "introducing Tianjin's food".
[0152] In one implementation, when the target reference information is evaluation information for the first user intention, the first user intention can be replaced based on the evaluation information or the evaluation information and the first user intention can be processed again to understand the intention so as to obtain the second user intention.
[0153] For example, the evaluation information may be: "Not good", based on which, the evaluation information is used to replace the intent keyword of the first user intent "Introduce Tianjin" to obtain the second user intent "Introduce Beijing". For another example, the evaluation information may be "I want to see what delicious food there is", based on which, the evaluation information and the first user intent "Introduce Tianjin" are processed again using the model to obtain the new second user intent "Introduce Tianjin's food".
[0154] In one implementation, when the target reference information comes from a target application, the first user intent may be updated based on the application information of the target application to obtain the second user intent.
[0155] The update process here may be at least one of replacement, production, adjustment, editing, etc. For example, the first application intent of "uninstalling video A" is updated based on the target application "video B" to obtain the second user intent of "uninstalling video B".
[0156] In one implementation, when the target reference information comes from the target space environment, the first user intention is modeled based on the environmental parameters of the target space environment to obtain the second user intention.
[0157] For example, the target space environment refers to: stadium, indoor, city, country and other space environments. Based on this, in this embodiment, the first user intention "introduce this place" can be modeled based on the environmental parameters of the target space environment, such as "C mall in Beijing", to obtain the second user intention "introduce the food in C mall in Beijing".
[0158] Based on the above implementation scheme, step 103 can be implemented by at least one of the following implementation methods when executing the response operation matching the second user intention:
[0159] In one implementation, in this embodiment, a corresponding execution strategy may be determined based on the second user intention, so as to call a corresponding application tool to perform a target response operation based on the execution strategy.
[0160] The execution strategy may include: execution or execution after a certain delay, etc. The application tool may include a calendar, email, large model, computer manager or other device assistants, etc.
[0161] For example, in this embodiment, the computer manager can be used to immediately execute the target response operation of "uninstalling video B" intended by the second user.
[0162] In one implementation, in this embodiment, a corresponding execution strategy can be determined based on the second user intention, so as to call the corresponding agent to perform the target response operation based on the execution strategy.
[0163] Among them, the intelligent agent deployed locally on the intelligent electronic device, such as Figure 2As shown in , the laptop can be locally deployed with agents: SmartEngine, PCM agent, Windows agent, model or others. The SmartEngine can implement scene scheduling, performance switching, charging settings, graphics card overclocking, intelligent noise reduction, color management and other functions; PCM can implement virus detection, space cleaning, one-click acceleration, CPU usage, power consumption ranking and other functions; Windows agent can implement resolution settings, add Bluetooth devices, wireless LAN WLAN (Wireless Local Area Network) switch, background image settings, screen related settings and other functions; the model can realize functions such as text generation, image generation and video generation.
[0164] For example, in this embodiment, according to the immediate execution strategy of the second user intention of "uninstalling video B", the PCM (Phase Change Memory) agent can be called to perform the target response operation, that is, uninstalling video B.
[0165] In one implementation, in this embodiment, a corresponding execution strategy can be determined based on the second user intention and the operating status information of the electronic device, so as to call the corresponding application tool or agent to perform the target response operation based on the execution strategy.
[0166] The operation status information of the electronic device may include: the usage status of the application tool or the intelligent agent and the performance status of the device. For example, the usage status of the application tool or the intelligent agent may be: whether it is occupied. In this embodiment, the corresponding execution strategy is determined accordingly. For example, if the application tool or the intelligent agent is currently occupied or the performance status of the electronic device indicates that the electronic device is busy, then the execution strategy may be: delay execution until the application tool or the intelligent agent enters an idle state or the electronic device is idle.
[0167] Based on the above implementation, this embodiment may also include at least one of the following processing:
[0168] In an implementation manner, in this embodiment, the second user intention may also be updated based on the third feedback information obtained regarding the execution result of the response operation.
[0169] The third feedback information may be: feedback information of the target user on the execution result after the response operation matching the second user's intention is executed. For example, feedback information of approval or disapproval of the execution result, etc. Based on this, in this embodiment, the second user's intention may be updated according to the third feedback information, and then the response operation matching the newly obtained second user's intention may be re-executed.
[0170] For example, after executing the response operation that matches "Introducing Tianjin's cuisine", that is, outputting information about Tianjin's cuisine, the target user's feedback on the information about Tianjin's cuisine is obtained, such as feedback information that agrees with the execution result, and "Introducing Tianjin's cuisine" is updated according to the feedback information, and the new second user intention of "outputting more cuisine information" is obtained, and then the response operation of "outputting more cuisine information" can be re-executed.
[0171] In one implementation, in this embodiment, the second user intention may be updated based on an input operation acting on the interactive interface of the target agent.
[0172] The target agent may be an AI Agent integrated in the electronic device. The target agent may provide an interactive interface, and the target interface prompts the target user to perform input operations on the second user's intention, so that the target user can perform input operations on the interactive interface of the target agent, and the second user's intention can be updated on the electronic device based on the input operations.
[0173] The input operation here can be: the configuration operation performed by the target user on the interactive interface for the second user intent. For example, the target user configures "Introduce Tianjin's delicacies" on the interactive interface of the AI Agent, and a new second user intent "Introduce Tianjin's fried dough twist and steamed buns" can be obtained, and the response operation of "Introduce Tianjin's fried dough twist and steamed buns" can be re-executed accordingly.
[0174] refer to Figure 3 , is a schematic diagram of the structure of an interactive processing device provided in an embodiment of the present application, which can be configured in an electronic device capable of data processing, such as a mobile phone, a tablet computer, a computer or a server, etc. The technical solution in this embodiment is mainly used to improve the user's interactive experience.
[0175] Specifically, the device in this embodiment may include the following units:
[0176] A first intention obtaining unit 301 is used to identify the obtained target input data to obtain a first user intention;
[0177] A second intention obtaining unit 302, configured to process the first user intention based on target reference information to obtain a second user intention;
[0178] An operation execution unit 303, configured to execute a response operation matching the second user intention;
[0179] The target reference information corresponding to different target input data is different or the same.
[0180] It can be seen from the above technical solution that in an interactive processing device provided by an embodiment of the present application, after obtaining the first user intention by identifying the obtained target input data, the response operation matching the first user intention is not directly executed, but the first user intention is first processed based on the target reference information to obtain the second user intention, so that the response operation matching the second user intention can be executed. It can be seen that in this embodiment, the directly obtained user intention is first adjusted according to the target reference information, so that a user intention that better meets the user's needs can be obtained, and then the response operation is executed accordingly, which can better meet the user's needs, thereby improving the user's interactive experience.
[0181] In one implementation, the second intention obtaining unit 302 is further configured to implement at least one of the following:
[0182] Obtaining content information and / or attribute information of the target input data, and acquiring the target reference information based on the content information and / or the attribute information, wherein the target reference information is related to or unrelated to the target input data;
[0183] Evaluation information for the first user intention is obtained, and corresponding target reference information is acquired based on the evaluation information.
[0184] In one implementation, the second intention obtaining unit 302 obtains the target reference information based on the content information and / or the attribute information by at least one of the following:
[0185] Determine the interaction scenario represented by the content information, and obtain target reference information matching the interaction scenario;
[0186] Determine an interactive object in the content information, and obtain target reference information matching the interactive object;
[0187] Determine a data type of the target input data represented by the attribute information, and obtain target reference information matching the data type;
[0188] Determine source information of the target input data represented by the attribute information, and obtain target reference information matching the source information;
[0189] Determine an input position of the target input data represented by the attribute information, and obtain target reference information matching the input position;
[0190] Determine the data volume of the target input data represented by the attribute information, and obtain target reference information matching the data volume.
[0191] In one implementation, the second intention obtaining unit 302 obtains the target reference information based on the content information and / or the attribute information by at least one of the following:
[0192] Acquire user portrait information associated with the content information and / or the attribute information, and use the user portrait information as the target reference information;
[0193] Acquire long-term memory associated with the content information and / or the attribute information, and use the long-term memory as the target reference information;
[0194] Acquire short-term memory associated with the content information and / or the attribute information, and use the short-term memory as the target reference information.
[0195] In one implementation, the second intention obtaining unit 302 obtains the corresponding target reference information based on the evaluation information by at least one of the following:
[0196] In a case where the evaluation information includes first feedback information of a target user on the first user intention, using preset reference information matching the first user intention as the target reference information;
[0197] In a case where the evaluation information includes second feedback information of the target user on the first user's intention, using the second feedback information directly as the target reference information, or generating the target reference information based on the second feedback information;
[0198] In a case where the evaluation information includes supplementary information of the target user's intention to the first user, the target reference information is generated based on the supplementary information.
[0199] In one implementation, the second intention obtaining unit 302 processes the first user intention based on the target reference information by at least one of the following:
[0200] adding an attribute tag corresponding to the target reference information or a prompt word corresponding to the target reference information to the first user intention, so as to obtain the second user intention;
[0201] Using the target reference information, generating or replacing the first user intention to obtain the second user intention;
[0202] The first user intention is processed based on a processing strategy corresponding to the type and / or source of the target reference information to obtain the second user intention.
[0203] In one implementation, the second intention obtaining unit 302 processes the first user intention with a corresponding processing strategy based on the type and / or source of the target reference information by at least one of the following:
[0204] In a case where the target reference information is user portrait information of a target user, generating or updating the first user intent based on the user portrait information to obtain the second user intent;
[0205] In a case where the target reference information only includes long-term memory, adjusting or editing the first user intention based on the long-term memory to obtain the second user intention;
[0206] In the case where the target reference information includes short-term memory and long-term memory, the first user intention is updated based on the short-term memory, or, in the case where the first user intention cannot be updated using the short-term memory, the first user intention is updated using the long-term memory to obtain the second user intention;
[0207] In a case where the target reference information is evaluation information for the first user intention, replacing the first user intention based on the evaluation information or performing intent understanding processing on the evaluation information and the first user intention again to obtain the second user intention;
[0208] In a case where the target reference information comes from a target application, updating the first user intent based on application information of the target application to obtain the second user intent;
[0209] In a case where the target reference information comes from a target space environment, a model generation process is performed on the first user intention based on environmental parameters of the target space environment to obtain the second user intention.
[0210] In one implementation, the operation execution unit 303 executes the response operation matching the second user intention by at least one of the following:
[0211] Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding application tool to perform a target response operation based on the execution strategy;
[0212] Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding agent to perform a target response operation based on the execution strategy;
[0213] A corresponding execution strategy is determined based on the second user intention and the operating status information of the electronic device, so as to call a corresponding application tool or agent to perform a target response operation based on the execution strategy.
[0214] In one implementation, the second intention obtaining unit 302 further includes at least one of the following processes:
[0215] updating the second user intention based on the obtained third feedback information of the execution result of the response operation;
[0216] The second user intention is updated based on the input operation acting on the interactive interface of the target agent.
[0217] It should be noted that the specific implementation method of each unit in this embodiment can refer to the corresponding content in the previous text and will not be described in detail here.
[0218] refer to Figure 4 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, and the electronic device may be a mobile phone, a tablet computer, or a computer. The electronic device in this embodiment may include the following structure:
[0219] Memory 401, used to store computer programs and data generated by the execution of the computer programs;
[0220] The processor 402 is used to execute a computer program to implement the interactive processing method described in any of the above embodiments.
[0221] An embodiment of the present application further provides a computer storage medium on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, the interactive processing method described in any of the above embodiments is implemented.
[0222] An embodiment of the present application also provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the interactive processing method described in any of the above embodiments.
[0223] Taking a mobile phone as an example, the technical solution of the present application is described below:
[0224] After the user inputs a prompt such as text or an image, the system can understand the intention and proactively push relevant service content based on the user's portrait or habits, making an artificial intelligence personal computer (AIPC) that understands you better.
[0225] The current implementation scheme usually binds the training corpus to the intent through pre-training, which is relatively fixed and not flexible enough. For example, the intent mapped to "Video A" is "Open Video A". After the user inputs, the intent understanding returns the intent of "Open Video A", and its work is completed. The subsequent additional operation of "If not installed, prompt the user to install or pop up the installation interface" is only the behavior of the local software control logic based on the result of the intent feedback.
[0226] In view of this, the technical solution proposed in this application requires the support of the following contents:
[0227] 1. Data collection and analysis: First, we need to collect user data, including user behavior data, preference data, historical data, and the types and contents of files that are usually processed, etc. By analyzing and mining this data, we can understand the user's interests and needs and form a personal local portrait of the user.
[0228] 2. Recommendation algorithm: Determines the execution of user commands based on the long and short memory information of the user portrait and the intent category.
[0229] like Figure 5 As shown in , a schematic diagram of a module design for implementing the present application is provided. In this embodiment, after the user instruction is input, the user portrait information constructed by short-term memory and long-term memory is used to obtain the corresponding task list of the user intention through the intention understanding module. The execution strategy of each user intention in the task list can be immediate execution or delayed execution. Based on this, according to the execution strategy, corresponding tools such as calendars, emails, large models, etc. are used to execute response operations that match the user intention.
[0230] Combination Figure 6 The implementation process shown is as follows:
[0231] First, after waiting for the user's instruction, the intention understanding module is used to identify the intention of the input data corresponding to the user instruction to obtain the first user intention, and then the short-term memory module is used to find out whether there is relevant short-term memory (i.e. short-term memory). If there is relevant short-term memory, then the auxiliary attribute value can be added to the first user intention according to the short-term memory, so that the obtained second user intention is added to the task list; if there is no relevant short-term memory, then the long-term memory module can be used to find out whether there is relevant long-term memory (long-term memory). If there is relevant long-term memory, then the auxiliary attribute value can be added to the first user intention according to the long-term memory, so that the obtained second user intention is added to the task list. If there is no relevant long-term memory, then the task can be executed according to the task list.
[0232] It can be seen that this embodiment has the following advantages:
[0233] 1. Make the purpose of the user prompt more personal to the user;
[0234] 2. Combining user portraits, multi-intent or user context data can make the software smarter.
[0235] For example, based on short-term memory (conversation context), the intent understanding is as follows: Figure 7 As shown, the user’s first question is “How to uninstall software” and the user’s second question is “A video”.
[0236] If we follow the current implementation plan, Figure 8 As shown in , after waiting for the instruction of "Video A", the user intention of "open Video A" can be understood according to the intention of the predefined method, and then a response operation is performed, namely, Video A is opened.
[0237] According to the implementation scheme in this embodiment, Fig. 9 As shown in, after waiting for the instruction of "Video A", you can follow the target reference information "Uninstall software" corresponding to "How to uninstall software" in the short-term memory module, use the intention understanding module to obtain the user's intention "Uninstall Video A" based on short-term memory, and then execute the response operation, that is, open Video A.
[0238] For example, when the user's prompt is "Tianjin", the original implementation is as follows Fig.10 As shown in , the corresponding user intention is identified through intention understanding, and then the response operation matching the user intention is performed by the model and other intelligent agents. The implementation scheme in this embodiment is as follows Fig.11 As shown in the figure, the corresponding user intent is identified through intent understanding, and then the target reference information, that is, the prompt after completion, is obtained in combination with the user's local portrait. The user intent is then adjusted accordingly, and the response operation that matches the user intent is executed through the model and other intelligent agents.
[0239] For example, the local user portrait may be: portrait information such as tickets and hotel reservation software obtained through transportation invoice information; portrait information such as attraction introductions and itinerary planning obtained through travel photos from various places; portrait information of food software obtained through food summaries from various places; portrait information such as university information obtained through university information collection.
[0240] Taking the user prompt as "Tianjin" as an example, the interaction process between modules is as follows Fig.12 As shown in:
[0241] First, the intention understanding module identifies the place name type of "Tianjin"; then, the intention understanding module requests the long / short memory module to obtain memory information;
[0242] In one case, namely case 1, if the short-term memory module feeds back short-term memory, the content is university information, which is relevant to the place name, then the intention understanding module polishes the user intention corresponding to the prompt, obtains the new user intention "introducing university information in Tianjin", and then passes the user intention corresponding to the new prompt to the tool module, which calls the corresponding application tool or agent to perform the response operation;
[0243] In another case, namely case 2, if the short-term memory module does not feedback short-term memory, but the long-term memory module feedbacks long-term memory, and the content is food, then a food tool is defined in the intention understanding module, and the input parameter is "Tianjin". In this way, the new user intention "Tianjin" and "food tool" can be obtained, and then the user intention corresponding to the new prompt is passed to the tool module, and the tool module calls the corresponding application tool or intelligent agent to perform the response operation, such as starting the food software and inputting "Tianjin" into the food software, so that the food key outputs food recommendation information related to "Tianjin".
[0244] In addition, in this embodiment, when executing the response operation matching the user's intention, the local agent of the electronic device is implemented, as described in detail below:
[0245] The technical solution of this embodiment uses the intention understanding module to identify the user's intention, and combines the local agent set of the electronic device to provide the user with one or more agent combination calls to provide a more convenient and comprehensive solution for the user. The core of the technical solution of this embodiment mainly includes the training of the intention understanding model in the intention understanding module, the relationship mapping between the agent and the intention understanding model, and the registration and call of the agent's capabilities, as follows:
[0246] 1. Model training of intent understanding model:
[0247] Model training can provide corresponding words and response slots according to the capabilities of the intelligent agent. Take the smart engine's scene scheduling as an example:
[0248] Table 1 Slots and Scripts
[0249]
[0250] Through Table 1 and the generalized content of the model, an intent understanding model is formed. When the user's prompt is similar to the above-mentioned speech content, the intent understanding model can lock the slot information of the graph.
[0251] 2. Relationship mapping between intelligent agents and intention understanding models:
[0252] Still taking the above case as an example, after the intention understanding training is completed, it is necessary to establish a mapping relationship between it and the agent. Here, the design is to mark it by adding cmd and param, as shown in Table 2:
[0253] Table 2 cmd plus param
[0254]
[0255] At this time, when the user prompt hits the above intent, the specific cmd and param can be located. For example, if the user enters "I'm in a meeting", se.scene.status.set combined with meeting will be returned.
[0256] 3. Agent capability registration:
[0257] Due to the differences in the capabilities of each device and the installation of related functional software, each device behaves differently. When the technical solution in this embodiment is started, the support status of each agent on the device will be detected. This is also used for subsequent agent calls and user information feedback logic. The mapping relationship is shown in Table 3:
[0258] Table 3 Mapping relationship
[0259]
[0260] The precodition field is used to define whether the capability is supported. If the current device only supports the meeting scene but not the game scene, the technical solution of this embodiment will only return se.scene.ability_meeting when detecting the SE agent, indicating that it has the capability of the meeting scene.
[0261] 4. Agent call:
[0262] After receiving se.scene.status.set combined with meeting, the technical solution in this embodiment needs to pass this information to the smart engine SDK and call the function of the software itself to perform settings.
[0263] 5. Call feedback:
[0264] After the call, the corresponding words will be displayed to the user according to the setting results of SE.
[0265] 6. Schematic diagram of the agent range, such as Fig.13 as shown in .
[0266] The specific process is as follows Fig.14 As shown in:
[0267] First, obtain the information of each agent, and then register the capabilities of each agent into the system; after waiting for the user to input the prompt, use the intent understanding model to analyze the prompt input by the user. If the capability of the intent mapping corresponding to the prompt is satisfied by this solution, then call the corresponding agent to perform the response operation that matches the user's intent, and obtain the agent's execution result and feedback; if it is not satisfied, then feedback is given upward that the capability is not supported.
[0268] The technical solution in this embodiment has the following advantages:
[0269] (1) Combined with the intent understanding model, a new interaction method is defined for users to call the local agent of the device;
[0270] (2) It can help users save tedious operation steps and make the operation more convenient for users;
[0271] (3) This solution can facilitate the access of new agents and the removal of existing agents through cmd mapping relationships, which is convenient and flexible and increases the adaptability of the software.
[0272] Take the prompt "I'm going to a meeting" as an example. Fig.15 As shown in , it is an interactive flow chart of performing operations based on user intention in this embodiment;
[0273] First, the user starts the application software implementing the present embodiment. The module implemented in the present embodiment loads the SDK corresponding to the intelligent agent such as SmartEngine and PCM and requires capability registration. The intelligent agent provides capability feedback to the module implemented in the present embodiment.
[0274] Afterwards, the user inputs the prompt "Help me set the meeting scene" on the interactive interface provided by the application software implementing this embodiment. After the prompt is given to the intention understanding module, the intention understanding module uses the module implemented in this embodiment to set the meeting scene in SmartEngine based on cmd plus param (such as se.scene.status.set+meeting). The setting result, such as scussed, is fed back to the module implemented in this embodiment, and then "the meeting scene is set up" is fed back to the application software;
[0275] Afterwards, the user inputs the prompt "Help me clean up the memory" on the interactive interface provided by the application software that implements this embodiment. After the prompt is given to the intention understanding module, the intention understanding module uses the module implemented in this embodiment to call PCM based on cmd plus param (such as pcm.device.memory). PCM pops up a memory cleaning page for user operation, and then returns a command response to the module implemented in this embodiment, and then returns a command response to the application software.
[0276] 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.
[0277] 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.
[0278] 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.
[0279] 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. An interactive processing method, include: Identify the obtained target input data to obtain a first user intention; Based on the target reference information, the first user intention is processed to obtain a second user intention; executing a response operation matching the second user intention; The target reference information corresponding to different target input data is different or the same.
2. The method according to claim 1, further comprising at least one of the following: obtaining content information and / or attribute information of the target input data, and acquiring the target reference information based on the content information and / or the attribute information, in, The target reference information is related to or unrelated to the target input data; Evaluation information for the first user intention is obtained, and corresponding target reference information is acquired based on the evaluation information.
3. The method according to claim 2, in, Acquiring the target reference information based on the content information and / or the attribute information includes at least one of the following: Determine the interaction scenario represented by the content information, and obtain target reference information matching the interaction scenario; Determine an interactive object in the content information, and obtain target reference information matching the interactive object; Determine a data type of the target input data represented by the attribute information, and obtain target reference information matching the data type; Determine source information of the target input data represented by the attribute information, and obtain target reference information matching the source information; Determine an input position of the target input data represented by the attribute information, and obtain target reference information matching the input position; Determine the data volume of the target input data represented by the attribute information, and obtain target reference information matching the data volume.
4. The method according to claim 2, in, Acquiring the target reference information based on the content information and / or the attribute information includes at least one of the following: Acquire user portrait information associated with the content information and / or the attribute information, and use the user portrait information as the target reference information; Acquire long-term memory associated with the content information and / or the attribute information, and use the long-term memory as the target reference information; Acquire short-term memory associated with the content information and / or the attribute information, and use the short-term memory as the target reference information.
5. The method according to claim 2, in, Acquiring corresponding target reference information based on the evaluation information includes at least one of the following: In a case where the evaluation information includes first feedback information of a target user on the first user intention, using preset reference information matching the first user intention as the target reference information; In a case where the evaluation information includes second feedback information of the target user on the first user's intention, using the second feedback information directly as the target reference information, or generating the target reference information based on the second feedback information; In a case where the evaluation information includes supplementary information of the target user's intention to the first user, the target reference information is generated based on the supplementary information.
6. The method according to any one of claims 1 to 5, in, Processing the first user intention based on the target reference information to obtain a second user intention includes at least one of the following: adding an attribute tag corresponding to the target reference information or a prompt word corresponding to the target reference information to the first user intention, so as to obtain the second user intention; Using the target reference information, generating or replacing the first user intention to obtain the second user intention; The first user intention is processed based on a processing strategy corresponding to the type and / or source of the target reference information to obtain the second user intention.
7. The method according to claim 6, in, Processing the first user intention with a corresponding processing strategy based on the type and / or source of the target reference information to obtain the second user intention includes at least one of the following: In a case where the target reference information is user portrait information of a target user, generating or updating the first user intent based on the user portrait information to obtain the second user intent; In a case where the target reference information only includes long-term memory, adjusting or editing the first user intention based on the long-term memory to obtain the second user intention; In the case where the target reference information includes short-term memory and long-term memory, the first user intention is updated based on the short-term memory, or, in the case where the first user intention cannot be updated using the short-term memory, the first user intention is updated using the long-term memory to obtain the second user intention; In a case where the target reference information is evaluation information for the first user intention, replacing the first user intention based on the evaluation information or performing intent understanding processing on the evaluation information and the first user intention again to obtain the second user intention; In a case where the target reference information comes from a target application, updating the first user intent based on application information of the target application to obtain the second user intent; In a case where the target reference information comes from a target space environment, a model generation process is performed on the first user intention based on environmental parameters of the target space environment to obtain the second user intention.
8. The method according to claim 1, in, Performing a response operation matching the second user intention includes at least one of the following: Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding application tool to perform a target response operation based on the execution strategy; Determine a corresponding execution strategy based on the second user intention, so as to call a corresponding agent to perform a target response operation based on the execution strategy; A corresponding execution strategy is determined based on the second user intention and the operating status information of the electronic device, so as to call a corresponding application tool or agent to perform a target response operation based on the execution strategy.
9. The method according to claim 1, further comprising at least one of the following: updating the second user intention based on the obtained third feedback information of the execution result of the response operation; The second user intention is updated based on the input operation acting on the interactive interface of the target agent.
10. An interactive processing device, include: A first intention obtaining unit, used for identifying the obtained target input data to obtain a first user intention; a second intention obtaining unit, configured to process the first user intention based on target reference information to obtain a second user intention; an operation execution unit, configured to execute a response operation matching the second user's intention; The target reference information corresponding to different target input data is different or the same.
