Interactive assistant training method, terminal and computer-readable storage medium

Abstract

The present invention belongs to the field of artificial intelligence technology and relates to a training method, terminal and computer-readable storage medium for an interactive assistant. The training method for an interactive assistant includes the following steps: in response to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information and personalized information, performing an interactive response according to the user information, and / or responding according to the response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control. Therefore, the present invention can improve the interactive assistant's level of awareness of users, better meet the needs of users with a personalized response method, and make users more willing to provide feedback to the interactive assistant to improve the intelligence of the interactive assistant, thereby improving the user's usage experience and user stickiness in a two-way manner.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology, and in particular relates to a training method, a terminal and a computer-readable storage medium for an interactive assistant. Background Art

[0002] With the development and popularization of artificial intelligence and smart phones, more and more interactive assistants represented by intelligent voice assistants are involved in users' daily lives. For example, Xiaodu, Xiaoai, Siri or Tmall Genie on the market can achieve simple conversations or complete user-specified skills through human-computer interaction methods such as voice interaction.

[0003] However, the various interactive assistants on the market mentioned above are actually very mechanical in nature, though they appear to be intelligent on the surface. This is mainly due to the following two problems: (1) The interactive assistant does not know the specific identity of the user who triggered the operation instruction and cannot determine the user's identity; (2) The user guidance function of the interactive assistant is too mechanical (for example, it cannot provide different intelligent guidance based on different users) and cannot accurately understand the user. Therefore, how to make the interactive assistant improve user information so that it can better recognize and understand users, and perform different intelligent interactions based on different users to improve the user's experience in using the interactive assistant is something that technicians in this field need to consider.

[0004] In view of the above problems, those skilled in the art have been seeking solutions.

[0005] The preceding description is intended to provide general background information and does not necessarily constitute prior art. Summary of the Invention

[0006] The technical problem solved by the present invention is to provide a training method, terminal and computer-readable storage medium for an interactive assistant, so that the interactive assistant can more comprehensively recognize the user, identify the user's identity, and understand the user's habits. By gradually guiding the user and unlocking permissions, the user's usage stickiness is increased. At the same time, the assistant is trained and the user is understood, so that each user can obtain personalized services provided by the interactive assistant, thereby improving the user's usage experience.

[0007] It can improve the interactive assistant's understanding of users, better meet users' needs with personalized responses, and make users more willing to provide feedback to the interactive assistant to improve the intelligence of the interactive assistant, thereby improving the user experience and user stickiness in a two-way manner.

[0008] The present invention solves the technical problem by adopting the following technical solutions:

[0009] The present invention provides a training method for an interactive assistant, which includes the following steps: in response to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information; performing an interactive response based on the user information, and / or responding based on response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control.

[0010] The present invention also provides a terminal, comprising a processor and a memory: the processor is configured to execute a computer program stored in the memory to implement any one of the above steps of the method for seeking help.

[0011] The present invention also provides a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any of the above-mentioned rescue methods are implemented.

[0012] The present invention also provides a training method, terminal, and computer-readable storage medium for an interactive assistant. The interactive assistant training method includes the following steps: in response to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information; performing an interactive response based on the user information; and / or responding based on response mode information corresponding to the content to be executed and the user information. The interactive response includes at least one of information completion guidance and preference recommendation control. Therefore, the present invention can improve the interactive assistant's understanding of the user, better meet user needs with personalized responses, and increase user willingness to provide feedback to the interactive assistant, thereby improving the assistant's intelligence and improving the user experience and user stickiness in a two-way manner. Furthermore, the interactive assistant can automatically infer user preferences based on the user's location and search queries, and automatically display recommendations. By recording the user's personalized settings and user habits, each interactive assistant is unique to the user, providing unique responses that meet the user's needs. By assigning user-level operations, the interactive assistant's operational permissions and complaint credibility are tied together, encouraging users to use the interactive assistant more frequently. This improves the functionality and user experience, and encourages them to provide feedback. This allows developers to accurately understand customer needs and develop targeted products, enhancing the user experience and the intelligence of the interactive assistant. Furthermore, user information can be stored and synchronized in the cloud, increasing user stickiness.

[0013] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the following specifically cites preferred embodiments and describes them in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0015] Figure 1 A schematic flow chart of a training method for an interactive assistant provided in the first embodiment of the present invention;

[0016] Figure 2 A first structural diagram of a terminal provided in a second embodiment of the present invention;

[0017] Figure 3 A second structural diagram of a terminal provided in the second embodiment of the present invention;

[0018] Figure 4 This is an application environment diagram provided for the second embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the embodiments described are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0020] The embodiments of the present invention are further described below in conjunction with the accompanying drawings.

[0021] First embodiment

[0022] Figure 1 This is a flow chart of the training method of the interactive assistant according to the first embodiment of the present invention. Figure 1 .

[0023] Step S1: in response to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information.

[0024] In one embodiment, the interactive assistant described in this embodiment includes but is not limited to a related intelligent assistant that can realize human-computer interaction with the user, preferably a voice assistant such as Xiaodu, Xiaoai, Siri or Tmall Genie on the market, etc., and can also be other types of interactive assistants such as intelligent assistants that are inconvenient for other disabled people such as the deaf and dumb to use, which complete human-computer interaction by outputting text and pictures. Furthermore, for specific applicable terminals, it can include but is not limited to interactive assistants installed in mobile terminals or vehicle-mounted terminals, and can also be interactive assistants in other intelligent terminals, such as voice assistants in smart homes, intelligent robots that provide services to tourists in public places such as train stations and airports, etc. It can be understood that the above specific forms of interactive assistants and application platforms are only descriptive descriptions of the technology, not limitations of the technology, and specific implementation methods may not be limited to the above examples.

[0025] In one embodiment, methods for implementing responses to trigger instructions include, but are not limited to, "active" and "automatic" triggering methods. Specifically, "active" triggering, i.e., triggering by user operation, may include, but are not limited to, the user waking up the interactive assistant through voice calls, or activating / waking up by clicking / touching relevant buttons to respond to the trigger instruction. Furthermore, "automatic" triggering may include, but are not limited to, the interactive assistant having a certain level of intelligence, being able to intelligently identify whether the user currently needs the interactive assistant's related services. For example, after the user arrives home, the interactive assistant placed in the house actively wakes up to inquire about the user's needs, or provides corresponding suggestions or opinions to the user when the user searches for relevant content, or automatically responds to the trigger instruction to provide services to the user when the user arrives at a certain location, arrives at a certain time, or stays for a certain period of time, such as when the user is leaving home for dinner. In addition, "automatic" triggering may also include automatic triggering, waking up the interactive assistant to remind the user to use the interactive assistant after the user has used the interactive assistant for more than a certain period of time or a certain period of time has passed since the last use, such as 15 days. Because in the process of the growth of interactive assistants, users have very high requirements for experience. Many people will give up using interactive assistants after trying them several times because they are dissatisfied or not used to interacting with them. This has an impact on the user experience. For developers, it is a serious loss if they cannot clearly obtain the user's specific needs and opinions and suggestions. Therefore, for those who have not used the interactive assistant for a certain period of time, the interactive assistant can be automatically triggered to remind users to use the interactive assistant, cultivate the user's habit of using the interactive assistant, improve the intelligence of the interactive assistant, make users willing to use the interactive assistant, and developers can make timely adjustments to the user's needs, thereby achieving a win-win technical effect. It is understandable that since there are many situations in real life where the interactive assistant can participate in and meet user needs, it is difficult to enumerate them all, so this is only for illustrative purposes, that is, when any one or more of the above information related to user needs can be used as a condition for automatically responding to the trigger instruction, then the trigger instruction can be responded to.

[0026] In one embodiment, the content to be executed acquired in step S1 is information generated by the user's actual interaction with the interactive assistant. Taking a voice assistant as an example, the acquired content to be executed includes, but is not limited to, voice information, behavior information, and demand information containing user needs. Specifically, for example, if a user says something like "buy me a train ticket" to the interactive assistant, the voice information containing the desired purpose is the corresponding content to be executed. For another example, an interactive assistant that can recognize user behavior, such as a smart home interactive assistant configured to turn on or off the lights by snapping fingers, can acquire user behavior information to execute subsequent control or response operations. For another example, an interactive assistant that exchanges information via text, such as "buy me a train ticket" typed into the interactive assistant with a desired purpose, is the corresponding content to be executed. It is understood that there are many types of interactive assistants and many ways to express needs. The explanation of the content to be executed here is merely an example and does not constitute a technical limitation. Furthermore, user information primarily refers to a collection of information related to the user's personal needs or personalization, including at least one of identity information, user level, user habit information, and personalized information. Furthermore, identity information is used by the interactive assistant to identify the user. The corresponding user identity and / or nickname information allows the interactive assistant to know who the "car owner / me" is, allowing it to respond or perform targeted actions. User level can be understood as a numerical representation of the user's level of experience with the interactive assistant. For example, it can be set as: Apprentice Assistant Level 1 -> Basic Assistant Level 2 -> Intermediate Assistant Level 3 -> Advanced Assistant Level 4, etc. User level is tied to the interactive assistant, and as the user level increases, the interactive assistant's features and other services are gradually unlocked. This will be described in detail later; this is just an illustration. User habit information is the sum of user habits collected by the interactive assistant during its interactions with the user, enabling it to intelligently trigger the interactive assistant or provide targeted services. Personalized information is information set by the user or adjusted by the interactive assistant based on user habits, tailored to the user's preferences and needs, enabling the interactive assistant to respond uniquely to the user. Specifically, for a voice assistant, for example, personalized information can include a user-specific voice package, conversation style, and so on. It is understandable that the above disclosures and examples of various information are all for explanation of the technology, rather than for limitation of the technology. The specific situation is not limited to the above examples, and corresponding expansions or changes will be performed according to the specific situation.

[0027] Step S2: respond interactively according to the user information, and / or respond according to the response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control.

[0028] In one embodiment, an interactive response is performed based on user information, meaning that the interactive assistant can be controlled solely based on user information. The interactive response includes at least one of information improvement guidance and preference recommendation control. Regarding information improvement guidance, it is understood that improving various types of user information enables the interactive assistant to better understand the user, enabling responses or controls tailored to the user's habits. Secondly, because user information includes user levels, gradually increasing user levels can gradually expand the interactive assistant's permissions. Specifically, improving user information can also increase user levels, thereby expanding user permissions and enhancing the user experience. In another embodiment, the interactive response also includes preference recommendation control. It is understood that user information contains comprehensive information about user habits, preferences, etc., and when certain conditions are met, the interactive assistant can be automatically triggered to provide suggestions or recommendations to the user, thereby enhancing the interactive assistant's intelligence and the user experience. Specifically, information improvement guidance and preference recommendation control will be described in detail in subsequent steps and are only briefly explained here.

[0029] In one embodiment, step S2 includes: providing information improvement guidance according to the degree of improvement corresponding to the user information, and after the information improvement guidance is completed, improving the user level in the user information.

[0030] In one embodiment, the step of providing information completion guidance based on the degree of completeness corresponding to the user information includes: when the user information corresponds to a first degree of completeness, providing first information completion guidance, the first information completion guidance including entry guidance for identity information and / or a first function introduction, the identity information including biometric information; and / or, when the user information corresponds to a second degree of completeness, providing second information completion guidance, the second information completion guidance including user authentication guidance and / or a second function introduction; and / or, when the user information corresponds to a third degree of completeness, providing third information completion guidance, the third information completion guidance including payment binding guidance and / or a third function introduction.

[0031] In one embodiment, as explained above, user levels are related to the functions and permissions of the interactive assistant. More complete user information allows the interactive assistant to better understand the user and thus better meet their needs. By increasing the number of user levels, users can be encouraged to use the interactive assistant more frequently, and their levels can be improved during use. For example, when user information reaches a first level of completeness, a first information completion guide is provided. The first information completion guide includes guidance on entering identity information and / or a first function introduction. Identity information includes biometric information. The first level of completeness is preferably for new users who have just started using the interactive assistant. The first completion guide includes guidance on entering identity information and / or a first function introduction. For example, the first function introduction includes a self-introduction by the interactive assistant. The user is guided through the process of entering biometric information via voice, text, or other interactive methods. Biometric information may include, but is not limited to, facial recognition, voiceprint recording, or fingerprint input, as well as corresponding user identity and / or nickname information to let the interactive assistant know who the "car owner / me" is. Because user levels are tied to interactive assistant functionality, the interactive assistant corresponding to the first level of user information, preferably a new user, only has basic permissions and functionality, corresponding to the aforementioned apprentice assistant level 1. An interactive assistant that only knows the user's identity can only perform certain functions specific to that identity, or functions that can be performed independently of other identities. For example, the aforementioned recommendation control for favorites: for example, if a user searches for a cafe and passes by a cafe they've searched for, the user will be prompted with a "nearby cafe" and proactively recommend it to them.

[0032] In another embodiment, when the user information corresponds to the second level of completion, a second information completion guidance is performed, and the second information completion guidance includes user authentication guidance and / or a second function introduction. The function corresponding to the second level of completion requires further improvement of identity information, such as user identity guidance, specifically including guiding the user to enter identity card information for real-name authentication. Specifically, the functions that can be used for the second level of completion may include but are not limited to: when the user says "buy me a train ticket" or "I want to reserve a meeting room at 9 o'clock", such information that requires external output of "who I am", the user will be automatically helped to enter identity information. Furthermore, the implementation method of the second function introduction can refer to the description in the aforementioned first function introduction. The distinction between the first and the second is for guidance and introduction of different functions, that is, the distinction is made based on the difference in content. The specific implementation methods may be the same or similar, and will not be described here or later.

[0033] In other embodiments, when user information corresponds to the third level of completeness, guidance for completing the third information is provided. This guidance includes payment binding guidance and / or introduction to third functions. Functions corresponding to the third level of completeness are primarily used when the interactive assistant is involved in payment processes. For example, when a user needs to reserve parking, purchase train tickets, or movie tickets, and needs to make a payment, but lacks sufficient payment information, the user is guided to complete their payment information, i.e., perform payment binding, linking their Alipay, WeChat, bank card, or other payment account to the interactive assistant, enabling automatic payment upon purchase. Similarly, after information completion, if expanded functions or permissions are available, these can be introduced to the user to facilitate subsequent use. It is worth noting that, except for the first level of completeness, which is preferred for new users, there is no specific order of precedence between the second and third levels of completeness. They simply share the same format, distinguished by the terms second and third. In actual implementation, there is no restriction that the third level of completeness must be reached after the second level of completeness; the third level of completeness can also be reached before the second level of completeness.

[0034] In one embodiment, the above information enhancement primarily targets the identity information within the user information. However, the user information used by the interactive assistant to identify the user includes, in addition to identity information, user habit information and personalized information. The meaning of each piece of information has been previously explained and will not be elaborated upon here. It is understood that user habit information and personalized information are relatively broad concepts, unlike the specific biometric information used to identify a user or payment information used for payment. In layman's terms, the interactive assistant is not a simple fill-in-the-blank question, but rather a "reading question" that requires gradual analysis and acquisition through extensive use. Therefore, the acquisition method cannot be simply guided but requires extensive user experience. Therefore, the enhancement of user habit information and personalized information can be achieved by informing the user, for example, "After prolonged interaction with the interactive assistant, the interactive assistant can analyze your behavior, infer your preferences, and make targeted adjustments. We sincerely recommend that you use the interactive assistant frequently to improve your user experience!" For example, user habit information can include information about a user's frequent exercise routines, indicating that the user enjoys exercise. Therefore, if the user has no weekend plans, recommendations for outdoor fitness activities such as mountain climbing can be made. Alternatively, when ordering takeout, buying groceries, or dining, the user can be provided with a rough analysis of the calorie count and health status of the food, providing dining recommendations. Furthermore, personalized information can be provided through guidance similar to the aforementioned first and second information improvement guidance, or by the user themselves. Similarly, personalized information can be provided through "reading exercises" similar to mobile phone user habit information, analyzing the user's preferences through in-depth understanding and making targeted adjustments. Furthermore, the specific content of personalized information can include, but is not limited to, settings for the voice assistant's timbre, tone, and response method, and for other interactive assistants, settings for the user interface, response method, and so on.

[0035] In one embodiment, after the information completion guidance is completed, the user level in the user information is upgraded. Specifically, the user level can be stepped, such as the aforementioned trainee assistant level 1 -> junior assistant level 2 -> intermediate assistant level 3 -> senior assistant level 4. Alternatively, a mapping approach can be used, such as gradually unlocking permissions based on completed information. Specifically, each completed piece of information corresponds to a corresponding permission level for the operation, rather than necessarily a numerical expression. The user level is gradually upgraded based on the level of information completion.

[0036] In one embodiment, step S2 includes performing an interactive response based on user information and / or responding based on response mode information corresponding to the content to be executed and the user information. The interactive response includes at least one of information completion guidance and preference recommendation control, and further includes: obtaining preference recommendation triggers, which include at least one of location information and user search information; obtaining point of interest information based on the preference recommendation triggers and user habit information in the user information; and controlling display based on the point of interest information. Specifically, in this embodiment, for example, when a user arrives at a certain location, previously searched content may be recommended; when a user is searching for related content, targeted recommendations may be made based on user habits; when a user has a gap in their itinerary or plans, or is in the process of making arrangements, targeted recommendations may be made based on user habits, and so on. The recommendation method may be adaptively adapted to the interactive assistant's interaction method, such as voice announcements by a voice assistant or, when using a terminal with a screen, presentation of recommendations via bubbles, pop-ups, or other methods. It is understood that in real life, the number of situations in which preference recommendations can be made by intelligent assistants is difficult to exhaust, and the above examples are intended to illustrate the technology rather than to limit it.

[0037] In one embodiment, in step S2: interactive response is performed based on user information, and / or response is performed based on response mode information corresponding to the content to be executed and the user information, wherein, after the interactive response includes at least one of information completion guidance and preference recommendation control, it includes: obtaining account information corresponding to the user information; storing the user information in the cloud space corresponding to the account information. In this embodiment, that is, the user information is bound to the account information, and the account information is stored in the cloud space, so that the interactive assistant can keep the latest user data synchronized even when switching between terminals, so that even if other devices are changed, as long as the account information is kept synchronized, the user's user habit information and the user's assistant level will be saved and switched at any time. There is no need to worry that the new interactive assistant will no longer be the previous assistant after the device is switched, thereby improving the user experience, allowing users to be happy to explore the pleasure of switching and using the interactive assistant between multiple terminals, and increasing user stickiness.

[0038] In another embodiment, the step of synchronizing account information may also include synchronization of training data. It is understandable that the interactive assistant requires a lot of training to improve its intelligence, and a large amount of training data can be synchronized at the same time as the account information is synchronized. Therefore, when synchronizing account information, the terminal can simultaneously extract effective training data from the big data of training obtained from the cloud, and then iteratively update the interactive assistant. More specifically, the update method can be an update iteration based on the current user level, thereby reducing the amount of training data transmitted. As for the time of data synchronization, synchronization can be performed after each update of user information, or after each terminal switch and before triggering the interactive assistant.

[0039] In one embodiment, in step S2, after interactively responding based on user information and / or responding based on response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control, the following steps are performed: obtaining feedback information; matching the corresponding complaint credibility based on the user level in the user information; and performing feedback control on the developer side based on the feedback information and the complaint credibility. Furthermore, in the step of obtaining feedback information, the following steps are performed: triggering a feedback information collection operation based on the content to be executed to obtain feedback information; and / or executing a feedback information collection operation to obtain feedback information when the abnormal response information meets the feedback information acquisition conditions.

[0040] In one embodiment, when a user experiences failures using the interactive assistant, they can submit complaints and requests. Complaints can increase the assistant's user level, encouraging users to provide more feedback. For example, the initial level can be set to level 1, and the level increases with each complaint. During the feedback process, the progress of complaints can be monitored at all times. Specifically, higher-level users receive faster responses and updates to their complaints and requests, encouraging users to proactively use the voice function. Even if the function fails, as long as the user has a request, developers can more quickly identify the problem and understand the user's needs. Simply put, the feedback process involves obtaining feedback information, determining the user level, and matching it with the corresponding complaint credibility. The feedback information is then sent to the developer. Complaint credibility, in other words, determines the user's credibility based on their level. It is understood that higher-level users, due to their frequency and duration of use, have more experience and understanding of whether feedback is necessary and worthwhile. Therefore, feedback from higher-level users is more valuable to developers, and therefore, feedback from high-level users can be prioritized in the processing process. Furthermore, feedback from low-ranking users isn't necessarily without value. Therefore, the credibility of complaints from low-ranking users can be matched to their value. For example, each piece of feedback can be posted to a unified community platform, where community members can simply communicate and respond to small, resolvable issues, thus reducing the time required to process the feedback. Furthermore, the value of feedback can be determined based on information such as the number of likes on the platform. For example, even if a low-ranking user posts feedback on the platform, a large number of likes or comments from a diverse group of users can increase the number of views. This is especially true if the comments all express the same or similar unanswered questions. Even if the posting user has a lower rank, their complaint credibility can be increased, leading to a priority resolution. Feedback control on the developer side means that upon receiving user feedback, developers can promptly analyze and process the feedback to answer questions and resolve issues. Furthermore, if issues arise, updated versions can be synchronized to the training big data in the cloud, ensuring that every user is updated with the latest data.And because of the intelligent nature of the interactive assistant, interactive notifications can also be made after feedback information is collected or when the feedback information is resolved. For example, taking the voice assistant as an example, when the collection is completed, the user will be notified, "Okay, we have collected your feedback and will immediately perform follow-up processing on your feedback content. We will notify you in time after the processing is completed. Thank you for your use and trust!"; if the feedback information is resolved, the user can be notified, "Hello! The content of your feedback has now been resolved. We will first give you the following feedback..."

[0041] In one embodiment, for the step of obtaining feedback information by triggering the feedback information collection operation according to the content to be executed. The specific implementation method is mainly based on the feedback information collection triggered by the content to be executed. The specific implementation method can be based on the user's active call out. For example, for the interactive assistant of the voice assistant, it can be when the user says "I want to complain" or "I want feedback" to directly express the desire to interact with feedback, the feedback information collection is turned on; it can also be an indirect complaint by the user, such as when there are expressions like "Why is it like this?" or "Is it okay?", the information feedback collection is turned on. Furthermore, during the collection, the user can be prompted that feedback is being collected, such as "I'm sorry, is there anything I did wrong? I beg you to tell me, I will definitely correct it in time and get better!" It can be understood that the above are all illustrative treatments of the technology, not limitations of the technology.

[0042] In another embodiment, when an abnormal response meets the feedback information acquisition conditions, a feedback information collection operation is performed to obtain feedback information. Specifically, for example, during the interaction process, a user repeatedly makes the same request, each time without performing the subsequent steps. In this case, it can be roughly inferred that the user is dissatisfied with the existing response provided by the interactive assistant and deserves or requires feedback. Therefore, in this case, a feedback information collection operation can also be performed to obtain feedback information. Specifically, when a user makes the same or similar request three times in a row through the interactive assistant but does not respond to the interactive assistant's request, a prompt can be output to the user to collect feedback. For example, for a voice user, a voice prompt can be used to ask, "Hello, you didn't select the request I just made. Is there something you are dissatisfied with? Please let me know immediately so that I can accurately understand your request!" Therefore, in this embodiment, abnormal response information can automatically trigger feedback information collection, thereby improving the intelligence of the interactive assistant. Furthermore, for triggering based on content to be executed or triggering based on abnormal response information, there may even be overlap in the specific implementation methods. Therefore, the specific implementation methods are not limited to the above examples. The above examples are all explanations of the technology, not limitations on the technology. In the specific implementation, you can use the above as a reference and make adaptive adjustments and changes according to the actual situation, and they are all within the scope of protection of this application.

[0043] In one embodiment, in all of the above steps, the response mode information includes at least one of response authority information and response method information. In step S2: interactive response is performed based on the user information, and / or based on the response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control, and then includes: matching the corresponding response authority information based on the user level; and / or matching the corresponding response method information based on the personalized information; and responding based on the content to be executed, the response authority information, and / or the response method information.

[0044] In one embodiment, the response permission information is information corresponding to the user level, which is information about the functions of the interactive assistant that can be performed according to the degree of perfection of the user information, etc. Specifically, as described above, when corresponding to the first degree of perfection, the first function is unlocked, that is, the response permission corresponding to the first function; when corresponding to the second and third degrees of perfection, the second and third functions are unlocked, that is, the response permissions of the second and third corresponding interactive assistants. Furthermore, since the user level has been introduced in the previous text as having two forms, either numerical or tabular, and the corresponding permissions have a mutual mapping relationship, for example, for the numerical user level form, lv1 corresponds to the first level response permission, lv2 corresponds to the second level response permission... and so on; for the tabular user level, the response permission is a corresponding relationship with it. For example, if a certain user information is improved, the degree of perfection of the user information is improved, and the user level is also improved. Then, the improved user information can correspond to a response permission. As for the response method information, it mainly includes two types of information: response assistant information and / or response method information. For the former, since interactive assistants come in different types, such as voice assistants, text assistants, and smart homes, the corresponding assistant information determines which assistant or terminal the interactive assistant uses for interaction. For example, when a user is driving, a car's smart assistant is prioritized, when using a mobile terminal, a voice assistant is preferred, and when resting at home, a smart home is the best choice, and so on. The response method information mainly controls the interactive assistant to respond in a personalized mode according to the personalized information. Personalized information has been described in detail above. The main purpose is to determine whether the interactive assistant responds in a user-defined or adaptive manner. For example, different voices, timbres, and volumes can be used for voice assistants, while smart homes can be tailored to the user's different lifestyles, such as work and rest, diet, and other habits. Therefore, after the response permission information matches the user level and / or the personalized information matches the corresponding response method information, when executing a response based on the content to be executed, the response permission information and / or response method information are used to respond, so that each interactive assistant used by each person has a different response method to meet the user's personalized needs.

[0045] In one embodiment, the step of responding based on the response mode information corresponding to the content to be executed and the user information includes: when it is determined that the interactive authority is exceeded based on the content to be executed and the response authority information, information completion guidance is performed to expand the response authority information. In this embodiment, which is an extension of the information completion guidance step, the information completion guidance is preferably performed before the user uses the assistant. For example, the first level of completion is preferably used to guide new users. The information completion guidance in this embodiment is preferably mainly performed during use. For example, the user is familiar with the use of the interactive assistant, but discovers that the interactive assistant has insufficient authority during use. In this case, the interactive assistant prompts the user to complete the user information to improve the user level and expand the user authority, so that the interactive assistant can better serve the user, thereby achieving the technical effect of improving user stickiness and enhancing the intelligence of the interactive assistant.

[0046] The present invention first provides a training method for an interactive assistant, comprising the following steps: Step S1: Responding to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information; Step S2: Response based on the user information, and / or based on the response mode information corresponding to the content to be executed and the user information, wherein the response includes at least one of information completion guidance and preference recommendation control. Therefore, the present invention can improve the interactive assistant's understanding of the user, better meet the user's needs with personalized responses, and make the user more willing to provide feedback to the interactive assistant, thereby improving the interactive assistant's intelligence and improving the user experience and user stickiness in a two-way manner. Furthermore, the method can automatically infer the user's preferences based on the user's location and search, and automatically display recommendations. By recording the user's personalized settings and user habits, each interactive assistant is unique to the user, providing different responses that meet the user's needs. And by assigning user levels, the interactive assistant's operating authority and complaint credibility are bound together, which encourages users to use the interactive assistant more. The functions and experience available to users are also improved accordingly, and they are also willing to provide feedback, so that developers can accurately understand customer needs and carry out purposeful development, improving user experience and the intelligence of the interactive assistant. And user information can be stored in the cloud, kept synchronized with the cloud, and improve user stickiness. Secondly, it deepens the learning of the interactive assistant, enhances the interaction and communication between the interactive assistant and the user, and improves the interactive assistant's learning ability to understand user preferences. In addition, the level of the interactive assistant is linked to the services enjoyed by the user. Higher levels correspond to better services, and good feedback improves the interactive assistant's improvement capabilities.

[0047] Second embodiment

[0048] Figure 2 1 is a first structural diagram of a terminal provided by the second embodiment of the present invention. In order to clearly describe the terminal 110 provided by the second embodiment of the present invention, please refer to Figure 1 and Figure 2 .

[0049] The terminal 110 provided in the second embodiment of the present invention includes: a processor A101 and a memory A201, wherein the processor A101 is used to execute the computer program A6 stored in the memory A201 to implement the steps of the training method of the interactive assistant as described in the first embodiment.

[0050] In one embodiment, the terminal 110 provided in this embodiment may include at least one processor A101 and at least one memory A201. The at least one processor A101 may be referred to as a processing unit A1, and the at least one memory A201 may be referred to as a storage unit A2. Specifically, the storage unit A2 stores a computer program A6. When the computer program A6 is executed by the processing unit A1, the terminal 110 provided in this embodiment implements the steps of the interactive assistant training method described in the first embodiment. For example, Figure 1 Step S1 shown in: in response to a trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information; step S2: interactively responding according to the user information, and / or responding according to the response mode information corresponding to the content to be executed and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control.

[0051] In one implementation, the terminal 110 provided in this embodiment may include multiple memories A201 (referred to as storage units A2 for short).

[0052] The storage unit A2 may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a ferromagnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); the magnetic surface memory may be a magnetic disk or a magnetic tape. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus random access memory (DRRAM).The storage unit A2 described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memories.

[0053] In one embodiment, the terminal 110 further includes a bus connecting various components (e.g., the processor A101, the memory A201, the touch screen A3, etc.). For example, for an interactive assistant included in an intelligent robot that provides guidance or advice to tourists in public places such as train stations and airports, voice interaction is not effective due to the noisy environment of public places. However, the touch screen A3 can provide a good human-computer interaction effect.

[0054] In one implementation, the terminal 110 in this embodiment may further include a communication interface (eg, I / O interface A4 ), which may be used to communicate with an external device.

[0055] In one embodiment, the terminal 110 provided in this embodiment may further include a communication device A5. For example, when user information and account information need to be bound and synchronized, the I / O interface A4 and / or the communication device A5 are required.

[0056] The terminal 110 provided in the second embodiment of the present invention includes a memory A101 and a processor A201, and the processor A101 is used to execute the computer program A6 stored in the memory A201 to implement the steps of the interactive assistant training method described in the first embodiment. Therefore, the terminal 110 provided in this embodiment can improve the interactive assistant's understanding of the user, better meet the user's needs with a personalized response method, and make the user more willing to provide feedback to the interactive assistant to improve the intelligence of the interactive assistant, thereby improving the user's usage experience and user stickiness in a two-way manner.

[0057] Figure 3 1 is a second structural diagram of the terminal 110 provided in the second embodiment of the present invention.

[0058] See also Figure 3 In other implementations, the terminal 110 provided in the second embodiment of the present invention includes: a central controller 801 (CPU for short), a read-only memory 802 (ROM for short), a random access memory 803 (RAM for short), a communication bus 804, an I / O interface 805, an input unit 806, an output unit 807, a storage unit 808, and a communication unit 809.

[0059] In other implementations, the central controller 801 (CPU for short), the read-only memory 802 (ROM for short), and the random access memory 803 (RAM for short) may constitute a processor.

[0060] In other embodiments, the terminal 110 may implement the training method of the interactive assistant as provided in the first embodiment, thereby improving the interactive assistant's understanding of the user, better meeting the user's needs with a personalized response method, and making the user more willing to provide feedback to the interactive assistant to improve the intelligence of the interactive assistant, thereby improving the user's usage experience and user stickiness in a two-way manner.

[0061] The second embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program A6. When the computer program A6 is executed by the processor A101, it implements the steps of the training method of the interactive assistant described in the first embodiment.

[0062] In one implementation, the computer-readable storage medium provided by this embodiment may include any entity or device, recording medium capable of carrying computer program code, such as ROM, RAM, magnetic disk, optical disk, flash memory, etc.

[0063] Figure 4 This is an application environment diagram provided by the second embodiment of the present invention.

[0064] join Figure 4 The application environment provided by the second embodiment of the present invention includes a terminal 110 and a user 120.

[0065] In one embodiment, the specific application environment can be an interaction between the user 120 and the terminal 110, and the terminal 110 executes a specific training method of the interactive assistant, such as payment behavior, reservation behavior, etc. In other embodiments, if there is a need for external hardware to participate, the terminal 110 can control the external hardware to implement it, and the needs raised by the user 120, such as smart furniture, can be controlled by the terminal 110 to control the start and stop of home appliances, etc. In another embodiment, the specific application environment can also include a server, such as cloud synchronization of data and cloud storage of user information, which requires the participation of the server. It can be understood that the above is only a simple description of the application environment, in which the terminal 110 and the user 120 are the minimum configuration of the application environment. In addition, other external devices can be added according to the specific application scenario, such as the aforementioned server.

[0066] The computer program A6 stored in a computer-readable storage medium provided in the second embodiment of the present invention, when executed by processor A101, can improve the interactive assistant's understanding of the user, better meet the user's needs with personalized responses, and make the user more willing to provide feedback to the interactive assistant, thereby improving the interactive assistant's intelligence and enhancing the user experience and user stickiness. Furthermore, it can automatically infer the user's preferences based on the user's location and search results, and automatically display recommendations. By recording the user's personalized settings and user habits, each interactive assistant is unique to the user, providing different responses that meet the user's needs. Furthermore, by assigning user levels, the interactive assistant's operating permissions and complaint credibility are tied together, encouraging users to use the interactive assistant more frequently, improving the user's available functions and experience, and increasing their willingness to provide feedback. This allows developers to accurately understand customer needs and develop targeted products, improving the user experience and the interactive assistant's intelligence. Furthermore, user information can be stored in the cloud and synchronized with the cloud, improving user stickiness. Secondly, it deepens the interactive assistant's learning, enhances interaction and communication between the interactive assistant and the user, and improves the interactive assistant's ability to understand user preferences. In addition, the level of the interactive assistant is linked to the services enjoyed by the user. Higher levels correspond to better services, and good feedback enhances the interactive assistant's improvement capabilities.

[0067] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0068] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, parts, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further in combination with the context in the specific embodiment. In this article, unless otherwise specified, the meaning of "multiple" and "several" is two or more.

[0069] It should be understood that, although the various steps in the flowchart in the embodiment of the present application are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence in the order indicated by the arrows. Unless clearly stated herein, the execution of these steps is not strictly limited in order, and they can be performed in other orders. Moreover, at least a portion of the steps in the figure may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and their execution order is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0070] Those skilled in the art will appreciate that all or part of the steps of the above-described method embodiments can be implemented by hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium, which, when executed, performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0071] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A training method for an interactive assistant, characterized in that: The steps include: In response to the trigger instruction, obtaining content to be executed and / or user information, wherein the user information includes at least one of identity information, user level, user habit information, and personalized information; Performing an interactive response based on the user information, and / or responding based on the response mode information corresponding to the to-be-executed content and the user information, wherein the interactive response includes at least one of information improvement guidance and preference recommendation control; After the step of interactively responding according to the user information and / or responding according to the response mode information corresponding to the to-be-executed content and the user information, the method further includes: Get feedback; Matching the corresponding complaint credibility according to the user level in the user information; wherein the value of the feedback content of low-level users is matched with the corresponding complaint credibility; Feedback control is performed on the developer side based on the feedback information and complaint credibility. Feedback content from high-level users is processed with higher priority. For low-level users, each piece of feedback is posted to a unified community platform, and the value of the feedback content is determined based on the number of likes on the platform. If a large number of different users like the feedback content, which increases the number of views, even if the user who posted it has a lower user level, the credibility of the complaint is also improved, thereby giving priority to resolving the feedback issue. The steps of obtaining feedback information include: triggering a feedback information collection operation according to the content to be executed to obtain the feedback information; the content to be executed includes interactive content in which the user directly expresses his / her desire to provide feedback or indirect complaint content of the user; and / or When the abnormal response information meets the feedback information acquisition condition, performing the feedback information collection operation to obtain the feedback information; The abnormal response information meets the feedback information acquisition condition when the user makes the same request multiple times during the interaction process, but does not perform subsequent steps each time.

2. The interactive assistant training method according to claim 1, wherein the step of interactively responding based on the user information comprises: Provide guidance on information completion according to the degree of completion of the user information; After the information improvement guidance is completed, the user level in the user information is improved.

3. The interactive assistant training method according to claim 2, wherein the step of guiding the user to complete the information according to the degree of completeness of the user information comprises: When the user information corresponds to a first level of completion, providing first information completion guidance, the first information completion guidance including guidance for entering identity information and / or a first function introduction, the identity information including biometric information; and / or, When the user information corresponds to a second level of completion, providing second information completion guidance, wherein the second information completion guidance includes user authentication guidance and / or second function introduction; and / or, When the user information corresponds to the third level of completion, a third information completion guidance is performed, where the third information completion guidance includes a payment binding guidance and / or a third function introduction.

4. The interactive assistant training method according to claim 1, wherein the step of interactively responding based on the user information and / or responding based on the response mode information corresponding to the to-be-executed content and the user information comprises: Obtaining a preference recommendation trigger factor, wherein the recommendation trigger factor includes at least one of location information and user search information; Acquiring point of interest information based on the preference recommendation trigger factor and the user habit information in the user information; Display control is performed according to the point of interest information.

5. The interactive assistant training method according to claim 1, wherein after the step of interactively responding according to the user information and / or responding according to the response mode information corresponding to the to-be-executed content and the user information, the method further comprises: Obtaining the account information corresponding to the user information; The user information is stored in the cloud space corresponding to the account information.

6. The interactive assistant training method according to any one of claims 1 to 5, wherein the response mode information includes at least one of response authority information and response method information; The step of interactively responding according to the user information, and / or responding according to the response mode information corresponding to the to-be-executed content and the user information, includes: Matching the corresponding response authority information according to the user level, and / or matching the corresponding response mode information according to the personalized information; Respond according to the to-be-executed content, the response authority information and / or the response method information.

7. The interactive assistant training method according to claim 6, wherein the step of responding according to the response mode information corresponding to the to-be-executed content and the user information comprises: When it is determined according to the to-be-executed content and the response authority information that the interaction authority is exceeded, the information improvement guidance is performed to expand the response authority information.

8. A terminal, characterized in that: Including processor and memory: The processor is configured to execute the computer program stored in the memory to implement the steps of the training method for an interactive assistant according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the training method for an interactive assistant according to any one of claims 1 to 7.

Images