Interaction method and apparatus, electronic device, and storage medium

Abstract

The application provides an interaction method, device, electronic equipment and storage medium, candidate interaction content is selected from an interaction content set based on current time and triggerable time of each interaction content in the interaction content set; human-computer interaction is performed based on the candidate interaction content; the each interaction content is obtained by content classification on historical interaction data, the interaction content set is maintained based on a memory period of the each interaction content, and the memory period and the triggerable time of the each interaction content are determined based on an interaction category to which the each interaction content belongs. The method, device, electronic equipment and storage medium provided by the application reduce terminal data storage, reduce terminal calculation, and thus improve the response speed of the terminal during interaction. Meanwhile, the candidate interaction content is selected from the interaction content set, human-computer interaction is performed, appropriate and close-to-user historical interaction data human-computer interaction is provided, and the experience of the user is improved.

Description

Technical Field

[0001] This invention relates to the field of human-computer interaction technology, and in particular to an interaction method, device, electronic device, and storage medium. Background Technology

[0002] Current human-computer interaction technology is no longer limited to user-initiated interactions; it can also involve proactive interactions via the terminal. In these technologies, the terminal typically utilizes a server to store a large amount of historical interaction data. The interaction content is then retrieved from this vast amount of historical data, and the interaction can proceed accordingly.

[0003] However, by searching for and interacting with content from stored historical interaction data, it is difficult to select appropriate content for interaction, resulting in low utilization of the information contained in historical interaction data and a poor user experience in human-computer interaction. Summary of the Invention

[0004] This invention provides an interaction method, device, electronic device, and storage medium to address the shortcomings of poor human-computer interaction experience in the prior art.

[0005] This invention provides an interaction method, comprising:

[0006] Based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the interactive content set;

[0007] Based on the candidate interaction content, human-computer interaction is performed;

[0008] Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs.

[0009] According to the interaction method provided by the present invention, the step of determining the triggerable time includes:

[0010] If the interaction content belongs to the first interaction category, the triggerable time of the interaction content is determined based on the generation time of the interaction content and the preset delay time of the interaction category to which the interaction content belongs.

[0011] If the interaction content belongs to the second category, a preference analysis is performed on the interaction content, and the triggerable time of the interaction content is determined based on the preference analysis results.

[0012] If the interaction content belongs to the third category, the event temporality of the interaction content is identified, and the triggerable time of the interaction content is determined based on the identified event temporality.

[0013] According to the interaction method provided by the present invention, determining the triggerable time of the interaction content based on the identified event temporality includes:

[0014] If the event is in the future tense, the triggerable time of the interactive content is determined based on the occurrence time and / or the generation time of the interactive content contained in the interactive content.

[0015] If the event is in the past tense, the interactive content is classified into events, and the triggerable time of the interactive content is determined based on the event type obtained from the classification.

[0016] According to the interaction method provided by the present invention, the step of determining the memory cycle includes:

[0017] When the interaction category to which the interactive content belongs is the first category or the second category, the preset memory period corresponding to the interaction category to which the interactive content belongs is determined as the memory period of the interactive content.

[0018] If the interaction content belongs to the third category, the memory period of the interaction content is determined based on the event temporality of the interaction content.

[0019] According to the interaction method provided by the present invention, determining the memory period of the interaction content based on the event temporality of the interaction content includes:

[0020] If the event is in the future tense, the memory period of the interaction content is determined based on whether the interaction content includes the occurrence time.

[0021] If the event is in the past tense, the memory period of the interactive content is determined based on the event type of the interactive content.

[0022] According to the interaction method provided by the present invention, the human-computer interaction based on the candidate interaction content includes:

[0023] Get the current interaction data;

[0024] From the candidate interactive content, select interactive content that is related to the current interactive data for human-computer interaction.

[0025] According to the interaction method provided by the present invention, the interaction category includes at least one of the following: event temporal category, event state category, psychological category, physiological category, phenomenon category, entity category, user-associated entity category, and person category.

[0026] The present invention also provides an interactive device, comprising:

[0027] The candidate interaction unit selects candidate interaction content from the interaction content set based on the current time and the triggerable time of each interaction content in the interaction content set;

[0028] The interaction unit performs human-computer interaction based on the candidate interaction content;

[0029] Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs.

[0030] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the interaction method as described above.

[0031] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the interaction method as described above.

[0032] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the interaction method as described above.

[0033] The interactive method, apparatus, electronic device, and storage medium provided by this invention obtain interactive content by classifying historical interactive data, maintain an interactive content set based on the memory period of each interactive content, and determine the memory period and triggerable time of each interactive content based on the interactive category to which each interactive content belongs. This reduces the amount of data stored on the terminal, thereby reducing the terminal's computational load and improving the terminal's response speed during interaction. Next, based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the interactive content set; based on the candidate interactive content, human-computer interaction is performed, providing appropriate human-computer interaction that closely resembles the user's historical interactive data, thus enhancing the user experience. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0035] Figure 1 This is one of the flowcharts illustrating the interaction method provided by the present invention;

[0036] Figure 2 This is the second flowchart illustrating the interaction method provided by the present invention;

[0037] Figure 3 This is a schematic diagram of the structure of the interactive device provided by the present invention;

[0038] Figure 4 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0040] To address the problem that related technologies in human-computer interaction often fail to select appropriate interactive content, resulting in low utilization of information contained in historical interaction data and a poor user experience, this invention provides an interaction method to achieve appropriate human-computer interaction that closely reflects the user's historical interaction data. Figure 1 One of the flowcharts of the interaction method provided by the present invention is shown below. Figure 1 As shown, the method includes:

[0041] Step 110: Based on the current time and the triggerable time of each interactive content in the interactive content set, select candidate interactive content from the interactive content set;

[0042] Step 120: Perform human-computer interaction based on the candidate interaction content;

[0043] Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs.

[0044] Specifically, historical interaction data refers to the interaction data generated between the terminal and the user before the current time. Historical interaction data can include voice data and image data. To fully utilize the information contained within the pre-stored historical interaction data during human-computer interaction, the data can be pre-classified to obtain different categories of interaction content, and the historical interaction data can be stored and maintained in the form of interaction content.

[0045] Here, historical interaction data can be categorized according to a preset period, such as classifying historical interaction data added every few hours. Content classification here refers to categorizing based on the content reflected in the historical interaction data. For example, voice data collected from the terminal can be used as historical interaction data, transcribed into historical interaction text, and its semantic features extracted for content classification. Alternatively, image data collected from the terminal can be used as historical interaction data, and the targets contained in the image data can be identified through target recognition, followed by content classification. Another option is to use image data collected from the terminal as historical interaction data, obtain the corresponding descriptive text based on a trained image description model, and classify the content based on the semantic features of the descriptive text, thereby obtaining the interaction content.

[0046] It's understandable that interactive content refers to the specific content within the content category of historical interactive data. Interactive content can be text reflecting information such as time, place, people, and emotions; specifically, it can be in the form of word segments, phrases, or structured text, such as triples (Xiaoming, likes, puppy). The specific information corresponding to the content category obtained by classifying historical interactive data can be used as interactive content.

[0047] Here, the content category to which the interactive content belongs must be a content category that can be used to indicate subsequent human-computer interaction. That is, the content category obtained by classifying content through historical interaction data can indicate that the historical interaction data does not contain information on content categories that can be used for human-computer interaction. If so, that part of the historical interaction data can be deleted, and only the information corresponding to the content category to which the historical interaction data belongs can be retained. The interactive content obtained in this way can be understood as data after content classification and filtering of historical interaction data. Compared with storing redundant historical interaction data, storing interactive content that is classified and refined, which has a smaller data volume and can reflect interactive information, can reduce the storage burden on the terminal storage space and facilitate data retrieval by the terminal during human-computer interaction, thereby improving the efficiency of human-computer interaction.

[0048] Considering that the amount of interactive content that needs to be stored will increase as the amount of historical interactive data on the terminal increases, the interactive content in the interactive content set can be maintained through a memory cycle for different categories of interactive content.

[0049] Here, the set of interactive content built upon the interactive content is not simply added to and never reduced; rather, it is maintained in real-time according to the memory period corresponding to its respective interactive category. That is, the memory periods for each interactive category are different. The interactive content of each category is combined into an interactive content set, and it can be managed according to the interactive period corresponding to its category. Interactive content that reaches its memory period can be deleted from the interactive content set. Generally, when the interactive content belongs to an interactive category that reflects user preferences, surrounding entity information, or other information closely related to the user's life (e.g., event status, entity, user-related entity), the memory period for this type of interactive category can be a long-term memory period, such as six months or a year. When the interactive content belongs to an interactive category that reflects the user's life phenomena, psychological or physiological information (e.g., phenomenon, psychological, physiological), the memory period for this type of interactive category can be short-term memory, such as one week or one month. For example, if the interactive content is "I have a toothache," and the corresponding interactive category is physiological, then the memory period for this interactive category is a short-term memory period, specifically one week. Furthermore, the memory period here can be preset, such as to one week, one month, or one year. Understandably, managing the set of interactive content according to the memory period, compared to storing all historical interaction data in existing technologies, saves more storage resources on the terminal, thereby improving the terminal's response speed during interaction.

[0050] Furthermore, trigger times can be set for each interactive content in the interactive content set, facilitating the selection of the most suitable interactive content for application at the current time during human-computer interaction. The trigger time here can be determined by the interactive category to which each interactive content belongs; for example, a specific trigger time can be pre-set for each interactive category. Generally speaking, different interactive categories correspond to different trigger times. When the interactive content belongs to an interactive category that reflects the user's emotions, physiological information, weather conditions, and other information commonly shared daily (e.g., psychological, physiological, phenomenon, related entity, or person categories), the trigger time for this interactive category can be a time after the corresponding interactive content's generation time, such as one day or one week. When the interactive content belongs to an interactive category that reflects the user's psychological state during an event or their feelings towards surrounding entities (e.g., event status or entity categories), the trigger time for this interactive category can be... The trigger time can be determined by the user's preference for the event or entity. For example, the higher the user's preference, the closer the trigger time is to the current time, and the lower the user's preference, the further the trigger time is from the current time. When the interaction content belongs to an interaction category that reflects an event that the user will perform at a certain time, such as an event tense category, the trigger time of this interaction category can be determined by the tense of the event. For example, if the event tense is past tense, the trigger time can be a time after the generation time of the interaction content. Or, if the event tense is future tense, the trigger time can be the time when the event occurs in the future.

[0051] Specifically, by using the terminal's current time and the triggerable time of each interactive content in the interactive content set, for example, the interactive content can be sorted according to the order of its triggerable time relative to the current time, and the interactive content with the highest triggerable time can be selected as candidate interactive content. Alternatively, interactive content with a triggerable time less than a certain threshold can be pre-set as candidate interactive content. It is understood that there can be one or more candidate interactive contents here.

[0052] Furthermore, after obtaining candidate interaction content, interaction can be performed based on the actual interaction status of the current terminal. For example, if the terminal is currently interacting with the user, it can select the most relevant interaction content from the candidate content based on the current interaction data for human-computer interaction. Alternatively, if the terminal is woken up by the user but does not receive any user-initiated interaction content within a certain time, it can directly select the most recently triggered interaction content from the candidate content for human-computer interaction. Here, human-computer interaction can involve inputting the interaction content into a pre-set question database, retrieving relevant question text from the database, and converting the question text into voice data, image data, or a combination of voice and image.

[0053] The method provided in this invention categorizes historical interaction data to obtain interactive content, maintains an interactive content set based on the memory period of each interactive content, and determines the memory period and triggerable time of each interactive content based on its interaction category. This reduces the amount of data stored on the terminal, thereby reducing the terminal's computational load and improving the terminal's response speed during interaction. Next, based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the set. Based on these candidate interactive content, human-computer interaction is performed, providing appropriate human-computer interaction that closely reflects the user's historical interaction data, thus enhancing the user experience.

[0054] Based on any of the above embodiments, the method for determining the triggerable time of interactive content includes:

[0055] If the interaction content belongs to the first category, the triggerable time of the interaction content is determined based on the generation time of the interaction content and the preset delay time of the interaction category to which the interaction content belongs.

[0056] The first category here can include psychological, physiological, phenomenon, related entity, and person categories.

[0057] Specifically, the interactive content corresponding to the psychological category can reflect the user's emotional state, such as happiness or sadness; the interactive content corresponding to the physiological category can reflect the user's physical condition, such as a headache; the interactive content corresponding to the phenomenon category can reflect weather phenomena or other natural phenomena, such as sunny weather or blooming flowers; the interactive content corresponding to the associated entity category can reflect entity information related to the user, such as the user's toys or pets; and the interactive content corresponding to the person category can reflect the user's relationships with people around them, such as the user's parents, friends, or teachers. In other words, the first category reflects a topic related to the user or themselves, including states, things, phenomena, or people related to the user. Interactions on a specific topic typically follow a predictable cycle.

[0058] Therefore, for each interaction category under the first category, a preset delay time can be set based on the generation time of the interaction content corresponding to that category to determine the triggerable time of the interaction content. The generation time of the interaction content here can be the generation time of the historical interaction data corresponding to that content. Furthermore, the preset delay time here refers to the time period between the generation time of the interaction content and the triggerable time. It is understandable that different interaction categories have different preset delay times, which can be determined by the topic's timeliness and repeatability. For example, when the interaction content includes a user's toothache, toothache can belong to the physiological interaction category. Topics related to physiological information generally have short timeliness and are generally not repeatable; the preset delay time could be set to one day or one week. As another example, when the interaction content includes a user's toy, toy can belong to the user-associated entity interaction category. Topics related to user-associated entities generally have longer timeliness and are generally repeatable; the preset delay time could be set to weekly or monthly. That is, the preset delay time can be a fixed time period or a certain time frequency. When the topic of the interaction category is non-repeatable, its corresponding preset delay time can be a fixed time period. When the topic of the interaction category is repeatable, its corresponding preset delay time can be a certain time frequency.

[0059] Furthermore, after obtaining the generation time of the interactive content and the preset delay time of the interaction category to which the interactive content belongs, the triggerable time of the interactive content can be calculated by counting the number of days backward from the generation time when the preset delay time is a fixed time period; when the preset delay time is a certain time frequency, the triggerable time of the interactive content can be calculated by counting the preset delay time from the generation time. It is understandable that when the preset delay time is a certain frequency, the triggerable time is usually triggered two to three times at the frequency of the preset delay time based on the generation time. Similarly, it is understandable that the triggerable time, specifically a certain time period within a day, can be a period of habitual user interaction obtained from historical interaction records, or it can be when the user's interaction with the terminal ends on the day of the trigger time.

[0060] If the interaction content belongs to the second category, a preference analysis is performed on the interaction content, and the triggerable time of the interaction content is determined based on the preference analysis results.

[0061] Here, the second category can be either event state or entity. Specifically, the interactive content corresponding to the event state category can reflect the user's attitude towards the event, such as desire or dislike; the interactive content corresponding to the entity category can reflect the user's preferences for various entities, such as fear of snakes or liking airplanes. That is, the second category reflects the user's preferences for events or entities, and usually, for interactions based on preferences, there is a predictable pattern in the trigger time corresponding to different preference results.

[0062] Therefore, for each interaction category under the second category, preference analysis can be performed on the interaction content corresponding to the interaction category. The trigger time for the interaction content can be determined based on the preference analysis results and the pre-set trigger time corresponding to the preference analysis results. This preference analysis of the interaction content can be performed using a preference analysis model. The preference analysis results can be represented by three levels: high, medium, and low, or in the form of a rating.

[0063] Furthermore, after obtaining the preference analysis results of the interactive content, different preset delay times can be set for different levels or ratings. Starting from the generation time of the interactive content, the time elapsed after the preset delay time is the triggerable time for that interactive content. This preset delay time can be a fixed time period, meaning the interactive content cannot be repeated; or it can be a time frequency, meaning the interactive content can be repeated. For example, if the preference analysis result is "highly liked," the preset delay time can be zero, and the triggerable time for the corresponding interactive content is the day the interactive content was generated. If the preference analysis result is "highly desired," the preset delay time can be weekly, and the triggerable time for the corresponding interactive content is every week after the generation time, typically two or three times. That is, when the preference analysis result of the interactive content is positive, the higher the level or rating of the preference analysis result, the shorter the preset delay time and the higher the repetition frequency can be. When the preference analysis result of the interactive content is negative, the higher the level or rating of the preference analysis result, the longer the interval between triggerable times and the lower the repetition frequency can be. If the level or rating exceeds a certain threshold, the interactive content can no longer be triggered.

[0064] If the interaction content belongs to the third category, the event temporality of the interaction content is identified, and the triggerable time of the interaction content is determined based on the identified event temporality.

[0065] Here, the third category could be the event tense category. Specifically, the interactive content corresponding to the event tense category can reflect the tense of the event described by the user, which can be past tense, present tense, or future tense. And usually, for interactions that target the event tense, there is a pattern to the trigger time corresponding to different event tenses.

[0066] Therefore, for each interaction category under the third category, the event tense can be obtained by identifying the event tense of the interaction content corresponding to the interaction category. Then, the triggerable time of the interaction content can be determined by pre-setting different triggerable times corresponding to different event tenses. This event tense identification of the interaction content can be achieved by comparing the interaction content with a pre-built interaction tense lexicon, and the comparison result determines the event tense contained in the interaction content. Generally, when the event tense in the interaction content is past tense, the triggerable time of the interaction content can be determined based on other attributes of the event, such as the event type; when the event tense in the interaction content is future tense, the triggerable time of the interaction content can be determined by the occurrence time of the event reflected by the event tense in the interaction content, or by the generation time of the interaction content; when the event tense in the interaction content is currently occurring, the interaction content may be the topic the user is currently interacting with, and active interaction can be avoided without triggering an ongoing event.

[0067] The method provided in this embodiment of the invention, compared to directly selecting interactive content from all interactive data for interaction, further categorizes the interactive content into a first category, a second category, and a third category. Based on the characteristics of different categories, the triggerable time of the interactive content corresponding to each interactive content under the corresponding category is determined, making the determined triggerable time more consistent with the triggerable time in the actual interactive scenario, and improving the accuracy of the interaction timing when the terminal initiates an active interaction.

[0068] Based on any of the above embodiments, determining the triggerable time of the interactive content based on the identified event temporality includes:

[0069] If the event is in the future tense, the triggerable time of the interactive content is determined based on the occurrence time and / or the generation time of the interactive content contained in the interactive content.

[0070] Specifically, when the event tense is future tense, the event tense included in the interactive content can be the specific time the event occurs, or a rough estimate of the future time. When the event tense reflects the specific time the event occurs, the triggerable time for the corresponding interactive content can be on the day of the event, or after a preset delay time, or after a preset delay time after the time the interactive content is generated. This preset delay time can be several days, a week, or a month. It's understandable that when the event in the interactive content is a meaningful event, such as an exam, a competition, or a trip to the zoo, the event's occurrence time will be prioritized as the triggerable time when the interactive content includes the event's occurrence time.

[0071] If the event is in the past tense, the interactive content is classified into events, and the triggerable time of the interactive content is determined based on the event type obtained from the classification.

[0072] Specifically, if the event is in the past tense, then the event has already occurred. The interactive content can be categorized using an event classification model based on graph attention networks to determine the event type. This event type can include common-sense events, meaningful events, and other events. Meaningful events can include things like a person's experiences.

[0073] Furthermore, after determining the event type, the triggerable time for the corresponding interactive content can be determined based on the pre-set triggerable times for each event type. It's understandable that common-sense events, such as brushing teeth or getting dressed, are frequently occurring in daily life and have low interactive value; therefore, the triggerable time for the corresponding interactive content can be left unset, meaning no human-computer interaction related to the interactive content will be initiated in subsequent interactions. When the event type is a meaningful event, such as an exam, competition, or outdoor activity, its interactive value is relatively high; therefore, the triggerable time for the corresponding interactive content can be a preset delay time after the interactive content's generation time. This preset delay time can be several days, a week, or a month. When the event type is another type, its interactive value may be moderate; therefore, the triggerable time for the corresponding interactive content can be the generation time of the interactive content, or a preset delay time after the generation time of the interactive content. This preset delay time can be a relatively short time period, such as one day or two days.

[0074] Understandably, regardless of whether the event tense of the interactive content is past, future, or ongoing, once the interactive content corresponding to the event triggers human-computer interaction, the terminal will not trigger human-computer interaction again for that interactive content unless the user initiates further interaction with that interactive content. This avoids multiple interactions with the same event in the interactive content, thereby improving the user's interactive experience.

[0075] The method provided in this invention determines the triggerable time of interactive content based on the occurrence time and / or generation time of the interactive content when the event tense is future; and classifies the interactive content into events when the event tense is past, and determines the triggerable time of the interactive content based on the event types obtained from the classification. This makes the triggerable time of the interactive content consistent with the event attributes in the interactive content, improving the accuracy and authenticity of the triggerable time of the interactive content, and thus enhancing the user's interactive experience.

[0076] Based on any of the above embodiments, the step of determining the memory period includes:

[0077] When the interaction category to which the interactive content belongs is the first category or the second category, the preset memory period corresponding to the interaction category to which the interactive content belongs is determined as the memory period of the interactive content.

[0078] Here, when the interaction category is Category 1 or Category 2, the corresponding interaction content does not contain event temporal information, making it difficult to directly determine its memory period through the interaction content. To facilitate obtaining its memory period, it can be determined by preset memory period. The memory period here can reflect the storage period of the interaction content in the interaction content set on the terminal or its corresponding storage device. When the memory period of the interaction content is reached, the interaction content can be deleted from the interaction content set to reduce the amount of interaction content stored on the terminal and improve the speed of the terminal's response to interactions.

[0079] Furthermore, the preset memory period here can be divided into long-term and short-term preset memory periods. Long-term preset memory periods can be three months, six months, or one year, while short-term preset memory periods can be one week or one month. It's understandable that the preset memory period corresponding to each interaction category is different. This can be achieved by pre-setting a specific preset memory period for each interaction category. For example, the preset memory period for the first category of interactions could be a long-term preset memory period, such as six months. More importantly, the memory period of interactive content is related to its triggerable time. Generally, interactive content stored in the interactive content set serves as candidate interactive content for the terminal to actively interact with. Once the interactive content reaches its triggerable time, it loses its value as candidate interactive content and can be removed from the interactive content set. Therefore, the preset memory period of interactive content can also be determined based on its triggerable time. For example, the triggerable time of the interactive content can be directly used as its memory period; that is, the time period between the triggerable time and the generation time of the interactive content is the memory period of that interactive content.

[0080] If the interaction content belongs to the third category, the memory period of the interaction content is determined based on the event temporality of the interaction content.

[0081] Here, when the interaction category is the third category, the interaction content contains information that reflects the event's temporal state. Therefore, the memory period of the interaction content can be determined by the event's temporal state. For example, if the event's temporal state is future, the memory period can be a preset delay time starting from the event's occurrence time, or a preset delay time starting from the interaction content's generation time. This preset delay time can be one week, one month, six months, etc. If the event's temporal state is past, the memory period can be determined by other event attributes. For example, different memory periods can be preset for different event types. For instance, if the event type is a meaningful event, the memory period of the interaction content could be six months.

[0082] The method provided in this embodiment of the invention improves the utilization rate of terminal storage by determining the preset memory period corresponding to the interaction category to which the interaction content belongs as the memory period of the interaction content when the interaction category to which the interaction content belongs is a first category or a second category; and by determining the memory period of the interaction content based on the event temporality of the interaction content when the interaction category to which the interaction content belongs is a third category. At the same time, it ensures the accuracy of the interaction content in the interaction content set, thereby improving the appropriateness of the interaction content when the terminal initiates human-computer interaction.

[0083] Based on any of the above embodiments, determining the memory period of the interactive content based on the event temporality of the interactive content includes:

[0084] If the event is in the future tense, the memory period of the interaction content is determined based on whether the interaction content includes the occurrence time.

[0085] Specifically, when the event is in the future tense, if the interaction content does not include the event's occurrence time, the memory period of the interaction content can be determined by its generation time. For example, a preset delay time after the generation time of the interaction content can be used as the memory period. If the interaction content includes the event's occurrence time, the memory period can be determined by the event's occurrence time. For example, the time period between the event's occurrence time and the generation time of the corresponding interaction content can be used as the memory period, or a preset delay time after the event's occurrence time can be used as the memory period. This preset delay time can be one week, one month, or the time period between the triggerable time of the interaction content and the generation time of the corresponding interaction content.

[0086] If the event is in the past tense, the memory period of the interactive content is determined based on the event type of the interactive content.

[0087] Specifically, when the event tense is past tense, different event types reflect different interaction values, and the memory period for interactive content with different interaction values ​​also varies. Generally, event types can include common-sense events, meaningful events, and other events. Common-sense events are frequently occurring daily activities such as brushing teeth and getting dressed; their interaction value is low, so the memory period for the corresponding interactive content can be relatively short, such as the same day or one or two days after the interaction content's generation time. When the event type is meaningful, such as exams, competitions, or outdoor activities, its interaction value is relatively high; the memory period for the corresponding interactive content can be a preset delay time after the interaction content's generation time. This preset delay time can be one week, one month, or several days after its triggerable time. When the event type is other events, its interaction value may be medium, and the memory period for the corresponding interactive content can also be a preset delay time after the interaction content's generation time; this preset delay time can be relatively short, such as two or three days. It is understandable that, regardless of the event type of the interactive content, the memory period of the interactive content must be after its triggerable time. That is, the interactive content cannot be deleted or removed from the interactive content set before its triggerable time has arrived.

[0088] Based on any of the above embodiments, Figure 2 The second flowchart of the interaction method provided by the present invention is as follows: Figure 2 As shown, in step 120, the human-computer interaction based on the candidate interaction content includes:

[0089] Step 210: Obtain the current interaction data;

[0090] Step 220: Select interactive content related to the current interactive data from the candidate interactive content for human-computer interaction.

[0091] Specifically, firstly, the current interaction data collected by the terminal is transcribed into the current interaction text via speech. This current interaction text reflects the main content of the current human-computer interaction. It is understandable that the closer the interaction content initiated by the terminal is to the current interaction text, the higher the accuracy of its proactive interaction content.

[0092] Next, the semantic similarity between the current interactive text and candidate interactive content can be calculated, and the candidate interactive content with the highest semantic similarity can be used as the interactive content for this human-computer interaction. Alternatively, the current interactive text can be categorized, and based on the interaction category of the interactive text and the interaction categories of each candidate interactive content, interactive content with the same interaction category as the current interactive text can be selected as the interactive content for this human-computer interaction. Here, candidate interactive content can be interactive content in the interactive content set whose triggerable time is relatively close to the current time.

[0093] Finally, select interactive content related to the current interactive data for human-computer interaction, which can be done through voice and / or images.

[0094] The method provided in this embodiment of the invention obtains current interaction data and selects interaction content related to the current interaction data from candidate interaction content for human-computer interaction, so that the active interaction initiated by the terminal is more in line with the interaction content of the current interaction, and further improves the accuracy of the interaction content when the terminal initiates active interaction.

[0095] Based on any of the above embodiments, the interaction category includes at least one of the following: event temporal category, event state category, psychological category, physiological category, phenomenon category, entity category, user-associated entity category, and person category.

[0096] Specifically, when the interaction category is event tense, the interaction content can include information reflecting the future tense, past tense, and ongoing tense of the event; when the interaction category is event state, the interaction content can include different emotional information of the user in daily events, such as liking, disliking, fearing, and not being afraid; when the interaction category is psychological, the interaction content can be sadness, grief, or anger; when the interaction category is physiological, the interaction content can be toothache or headache; when the interaction category is phenomenon, the interaction content can include weather phenomena or other natural phenomena; when the interaction category is entity, the interaction content can include various entities that exist in the user's life, such as refrigerators, televisions, and cars; when the interaction category is user-associated entity, the interaction content can include entities belonging to the user, such as backpacks, toys, and pets; when the interaction category is person, the interaction content can include information about the user's parents, friends, teachers, and other people.

[0097] Understandably, meticulously categorizing interactive content by interaction type is more conducive to improving the accuracy of determining the corresponding triggerable time through different interaction types. Similarly, determining the memory cycle of interactive content through different interaction types is also conducive to more meticulous maintenance of the terminal's interactive content set, improving the effective utilization of storage space, reducing the terminal's computing load, and thus improving the terminal's response speed.

[0098] Based on any of the above embodiments, the present invention also provides an interaction method, the method comprising:

[0099] First, based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the interactive content set;

[0100] Next, based on the candidate interaction content, human-computer interaction is performed;

[0101] Specifically, each interactive content is obtained by classifying historical interactive data. That is, the classification results of historical interactive data are directly used as the interactive content of historical interactive data, while other interactive data are discarded and not stored as interactive content.

[0102] Furthermore, the set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs.

[0103] Here, the memory period for each interactive content can be divided into long-term memory and short-term memory. The interactive content corresponding to the long-term memory period can be used to create user profiles. These user profiles are more conducive to developing an interactive style that resonates with the user and initiating human-computer interactions that interest them. Furthermore, when the interactive content involves significant user experiences, such as a user's first competition or first tooth extraction, this content can be included in the user's growth record. After obtaining the user's growth record, when selecting candidate interactive content on the terminal, content related to the user's growth record can be prioritized. Additionally, the user profiles and growth records recorded here can be recorded separately from the interactive content set, or they can be stored within the interactive content set with corresponding tags.

[0104] Based on any of the above embodiments Figure 3 This is a schematic diagram of the structure of the interactive device provided by the present invention, such as... Figure 3 As shown, the device includes:

[0105] The candidate interaction unit 310 selects candidate interaction content from the interaction content set based on the current time and the triggerable time of each interaction content in the interaction content set.

[0106] The interaction unit 320 performs human-computer interaction based on the candidate interaction content;

[0107] Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs.

[0108] The apparatus provided in this invention categorizes historical interaction data to obtain interactive content, maintains an interactive content set based on the memory period of each interactive content, and determines the memory period and triggerable time of each interactive content based on its interaction category. This reduces the amount of data stored on the terminal, thereby reducing the terminal's computational load and improving the terminal's response speed during interaction. Next, based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the set. Based on these candidate interactive content, human-computer interaction is performed, providing appropriate human-computer interaction that closely reflects the user's historical interaction data, thus enhancing the user experience.

[0109] Based on any of the above embodiments, the candidate interaction unit is specifically used for:

[0110] If the interaction content belongs to the first interaction category, the triggerable time of the interaction content is determined based on the generation time of the interaction content and the preset delay time of the interaction category to which the interaction content belongs.

[0111] If the interaction content belongs to the second category, a preference analysis is performed on the interaction content, and the triggerable time of the interaction content is determined based on the preference analysis results.

[0112] If the interaction content belongs to the third category, the event temporality of the interaction content is identified, and the triggerable time of the interaction content is determined based on the identified event temporality.

[0113] Based on any of the above embodiments, the candidate interaction unit is specifically used for:

[0114] If the event is in the future tense, the triggerable time of the interactive content is determined based on the occurrence time and / or the generation time of the interactive content contained in the interactive content.

[0115] If the event is in the past tense, the interactive content is classified into events, and the triggerable time of the interactive content is determined based on the event type obtained from the classification.

[0116] Based on any of the above embodiments, the candidate interaction unit is specifically used for:

[0117] When the interaction category to which the interactive content belongs is the first category or the second category, the preset memory period corresponding to the interaction category to which the interactive content belongs is determined as the memory period of the interactive content.

[0118] If the interaction content belongs to the third category, the memory period of the interaction content is determined based on the event temporality of the interaction content.

[0119] Based on any of the above embodiments, the candidate interaction unit is specifically used for:

[0120] If the event is in the future tense, the memory period of the interaction content is determined based on whether the interaction content includes the occurrence time.

[0121] If the event is in the past tense, the memory period of the interactive content is determined based on the event type of the interactive content.

[0122] Based on any of the above embodiments, the interaction unit is specifically used for:

[0123] Get the current interaction data;

[0124] From the candidate interactive content, select interactive content that is related to the current interactive data for human-computer interaction.

[0125] Based on any of the above embodiments, the interaction category includes at least one of the following: event temporal category, event state category, psychological category, physiological category, phenomenon category, entity category, user-associated entity category, and person category.

[0126] Figure 4 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440, wherein the processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute an interaction method, which includes: selecting candidate interaction content from the interaction content set based on the current time and the triggerable time of each interaction content in the interaction content set; performing human-computer interaction based on the candidate interaction content; wherein each interaction content is obtained by classifying historical interaction data, the interaction content set is maintained based on the memory period of each interaction content, and the memory period and triggerable time of each interaction content are determined based on the interaction category to which each interaction content belongs.

[0127] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0128] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the interaction methods provided by the above methods. The method includes: selecting candidate interaction content from the interaction content set based on the current time and the triggerable time of each interaction content in the interaction content set; performing human-computer interaction based on the candidate interaction content; wherein each interaction content is obtained by classifying historical interaction data, the interaction content set is maintained based on the memory period of each interaction content, and the memory period and triggerable time of each interaction content are determined based on the interaction category to which each interaction content belongs.

[0129] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the interaction method provided by the above methods. The method includes: selecting candidate interaction content from the set of interaction content based on the current time and the triggerable time of each interaction content in the set of interaction content; performing human-computer interaction based on the candidate interaction content; wherein each interaction content is obtained by classifying historical interaction data, the set of interaction content is maintained based on the memory period of each interaction content, and the memory period and triggerable time of each interaction content are determined based on the interaction category to which each interaction content belongs.

[0130] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0131] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An interaction method, characterized in that, include: Based on the current time and the triggerable time of each interactive content in the interactive content set, candidate interactive content is selected from the interactive content set; Based on the candidate interaction content, human-computer interaction is performed; Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs. The interaction categories include at least one of the following: event temporal type, event state type, psychological type, physiological type, phenomenon type, entity type, user-associated entity type, and person type; The steps for determining the triggerable time include: When the interaction content belongs to the first interaction category, the triggerable time of the interaction content is determined based on the generation time of the interaction content and the preset delay time of the interaction category to which the interaction content belongs; the first category includes the psychological category, the physiological category, the phenomenon category, the user-associated entity category, and the person category; If the interaction content belongs to the second category, a preference analysis is performed on the interaction content, and the triggerable time of the interaction content is determined based on the preference analysis results; the second category includes the event state class and the entity class; When the interaction content belongs to the third category, the event temporality of the interaction content is identified, and the triggerable time of the interaction content is determined based on the identified event temporality; the third category includes event temporality classes; The step of determining the triggerable time of the interactive content based on the identified event temporality includes: If the event is in the future tense, the triggerable time of the interactive content is determined based on the occurrence time and / or the generation time of the interactive content contained in the interactive content. If the event is in the past tense, the interactive content is classified into events, and the triggerable time of the interactive content is determined based on the event type obtained from the classification.

2. The interaction method according to claim 1, characterized in that, The steps for determining the memory cycle include: When the interaction category to which the interactive content belongs is the first category or the second category, the preset memory period corresponding to the interaction category to which the interactive content belongs is determined as the memory period of the interactive content. If the interaction content belongs to the third category, the memory period of the interaction content is determined based on the event temporality of the interaction content.

3. The interaction method according to claim 2, characterized in that, Determining the memory period of the interactive content based on the event temporality of the interactive content includes: If the event is in the future tense, the memory period of the interaction content is determined based on whether the interaction content includes the occurrence time. If the event is in the past tense, the memory period of the interactive content is determined based on the event type of the interactive content.

4. The interaction method according to claim 1, characterized in that, The human-computer interaction based on the candidate interaction content includes: Get the current interaction data; From the candidate interactive content, select interactive content that is related to the current interactive data for human-computer interaction.

5. An interactive device, characterized in that, include: The candidate interaction unit selects candidate interaction content from the interaction content set based on the current time and the triggerable time of each interaction content in the interaction content set; The interaction unit performs human-computer interaction based on the candidate interaction content; Each interactive content is obtained by classifying historical interactive data. The set of interactive content is maintained based on the memory period of each interactive content, and the memory period and triggerable time of each interactive content are determined based on the interactive category to which each interactive content belongs. The interaction categories include at least one of the following: event temporal type, event state type, psychological type, physiological type, phenomenon type, entity type, user-associated entity type, and person type; The candidate interaction unit is specifically used for: When the interaction content belongs to the first interaction category, the triggerable time of the interaction content is determined based on the generation time of the interaction content and the preset delay time of the interaction category to which the interaction content belongs; the first category includes the psychological category, the physiological category, the phenomenon category, the user-associated entity category, and the person category; If the interaction content belongs to the second category, a preference analysis is performed on the interaction content, and the triggerable time of the interaction content is determined based on the preference analysis results. The second category includes the event state class and the entity class; When the interaction content belongs to the third category, the event temporality of the interaction content is identified, and the triggerable time of the interaction content is determined based on the identified event temporality; the third category includes event temporality classes; The candidate interaction unit is also specifically used for: If the event is in the future tense, the triggerable time of the interactive content is determined based on the occurrence time and / or the generation time of the interactive content contained in the interactive content. If the event is in the past tense, the interactive content is classified into events, and the triggerable time of the interactive content is determined based on the event type obtained from the classification.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the interaction method as described in any one of claims 1 to 4.

7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the interaction method as described in any one of claims 1 to 4.

Images