Method for learning user preferences, device for learning user preferences, and intelligent device

Abstract

The application provides a method and device for learning user preferences and a smart device, and relates to the field of device control. The method comprises the following steps: receiving a voice instruction of a user, wherein the voice instruction is used for controlling the smart device; detecting whether a preset field is included in the voice instruction, wherein the preset field is used for representing preference information of the user; and in the case where the preset field is included in the voice instruction, obtaining the user preference information of the user for the smart device based on the voice instruction. The method can improve the learning efficiency of the user's use preference for the smart device.

Description

Technical Field

[0001] This application relates to the field of device control, and more specifically, to a method for learning user preferences, a device for learning user preferences, and a smart device in the field of device control. Background Technology

[0002] Currently, most home systems learn preferences and recommend patterns by counting frequency after users perform the same operations repeatedly. However, this learning strategy relies on long-term and large-scale data statistics, resulting in low efficiency in learning user preferences for smart devices.

[0003] Therefore, improving the learning efficiency of users' preferences for using smart devices is an urgent problem to be solved. Summary of the Invention

[0004] This application provides a method for learning user preferences, an apparatus for learning user preferences, and a smart device. The method can improve the efficiency of learning user preferences for using smart devices.

[0005] Firstly, a method for learning user preferences is provided, which includes: It receives voice commands from users, which are used to control smart devices. Detect whether the voice command includes a preset field, which is used to represent the user's preference information; When the voice command includes preset fields, the user's preference information for the smart device can be obtained based on the voice command.

[0006] In the above technical solution, after receiving a user's voice command to control the smart device, it detects whether the voice command includes a preset field; and if the preset field is detected, it determines the user's preference information based on the voice command. By receiving the user's voice command to control the smart device and detecting the preset field representing preference information, the user's preference information for the smart device can be accurately extracted directly from the voice command, achieving rapid identification and acquisition of user preferences, improving the efficiency and accuracy of the smart device's understanding of user intentions, and enhancing the user's learning efficiency of using the smart device.

[0007] In conjunction with the first aspect, in some possible implementations, user preference information for smart devices is obtained based on voice commands, including: Based on voice commands, determine tuple data, which includes: user identification information, target control action, and target control conditions required to execute the target control action; Based on tuple data, user preference information is obtained.

[0008] In the above technical solution, based on voice commands, a multivariate dataset including user identification information, target control actions, and target control conditions is obtained; then, user preference information is obtained based on the multivariate dataset. By structuring the user preference information into multivariate datasets containing user identification, target control actions, and control conditions, the expression of user preferences becomes more standardized and the logic clearer, facilitating subsequent storage, matching, and retrieval, and improving the manageability and reusability of preference information.

[0009] In combination with the first aspect and the above implementation methods, in some possible implementations, the method further includes: Acquire target information, which includes the current user's voice commands and / or current environment information; Determine whether the target information meets the target control conditions; When the target information meets the target control conditions, the target control action corresponding to the target control conditions is obtained based on the target information; Based on the target control action, generate target control instructions; and based on the target control instructions, control the target device corresponding to the target control instructions to execute the target control action.

[0010] In the above technical solution, after obtaining user preferences, the system further combines the current voice command or environmental information to determine whether the preset control conditions are met, and automatically generates control commands to drive the target device to perform corresponding actions, thereby realizing automated and intelligent control based on user preferences without requiring the user to repeatedly issue commands. When the target information includes the current environmental information and matches the user preferences, the system can meet the user's needs in advance without the user inputting commands, providing the user with proactive and accurate personalized services, and improving ease of use and intelligent experience.

[0011] Combining the first aspect and the above implementation methods, in some possible implementation methods, the target control conditions include voice triggering conditions and / or environmental triggering conditions. Determining whether the target information meets the target control conditions includes: If the target information is a voice command, and the voice command meets the voice triggering conditions, then the target information is determined to meet the target control conditions. If the target information is the current environment information, and the environment information meets the environment triggering conditions, then the target information is determined to meet the target control conditions. If the target information includes both voice commands and current environment information, then when the voice command meets the voice triggering condition and the environment information meets the environment triggering condition, the target information is determined to meet the target control condition.

[0012] In the above technical solution, the control conditions are refined into voice trigger conditions and / or environment trigger conditions, and single condition or combination condition judgment is supported, making the preference trigger mechanism more flexible and more in line with actual use scenarios. It can adapt to diverse interaction methods and environmental changes, and enhance the robustness and applicability of the control logic.

[0013] In combination with the first aspect and the above implementation methods, in some possible implementations, the method further includes: If the voice command does not include a preset field, the user's historical voice commands are obtained based on the user's voice command. The user intent corresponding to the historical voice command is the same as the user intent corresponding to the voice command. Based on historical voice commands, obtain the historical control actions corresponding to the historical voice commands; Based on historical control actions and voice commands, user preference information is obtained.

[0014] In the above technical solution, when the voice command does not include a preset field, the historical voice commands and corresponding historical control actions of the consensus map are retrieved and analyzed to obtain user preference information. This makes up for the lack of explicit preference information, realizes the mining and learning of users' potential and implicit preferences, expands the scope of user preference acquisition, and improves the completeness and continuity of preference learning.

[0015] Combining the first aspect and the above implementation methods, in some possible implementation methods, user preference information is obtained based on historical control actions and voice commands, including: Based on the execution frequency of various control actions in the historical control actions, the target historical control action is determined. The target historical control action is the historical control action whose execution frequency reaches the preset frequency. Based on the target's historical control actions, obtain historical environment information when the target's historical control actions are executed; Based on voice commands, historical environmental information, and target historical control actions, user preference information is obtained.

[0016] In the above technical solution, the high-frequency historical control actions are filtered by statistically analyzing their execution frequency, and user preference information is generated by combining the historical environment information corresponding to the historical control actions. This makes the obtained user preference information more in line with the user's actual usage habits, more representative and reliable, and avoids interference from accidental operations on preference learning.

[0017] In combination with the first aspect and the above implementation methods, in some possible implementations, the method further includes: Store user preferences for smart devices in a pre-defined preference memory.

[0018] In the above technical solution, the learned user preference information is uniformly stored in a preset preference memory, realizing centralized and persistent storage of preference information, which facilitates subsequent multiple calls, comparisons and updates, and provides data support for continuous learning of user habits and the realization of personalized intelligent control.

[0019] In combination with the first aspect and the above implementation methods, in some possible implementations, the method further includes: Real-time monitoring of new voice commands and corresponding control actions from users to smart devices, and analysis of whether the new voice commands and control actions are consistent with the user preference information already stored in the preference memory; If there is a discrepancy, the user preference information in the preference memory will be updated based on the new voice command and the corresponding control action; and the update time and update content will be recorded to facilitate the tracking of the user's preference change trajectory.

[0020] In the above technical solution, by monitoring the user's new voice commands and control actions in real time, inconsistent information in the preference memory bank is promptly detected and updated. At the same time, the update time and content are recorded to ensure that the user's preference information is always up-to-date and most accurate, and the trajectory of preference changes can be traced, thereby improving the system's adaptability and traceability.

[0021] Combining the first aspect and the above implementation methods, in some possible implementation methods, the preset fields include preference keywords, preference degree descriptors, and target device association words. Among them, preference keywords are used to characterize the user's tendency to control smart devices, preference degree descriptors are used to characterize the user's preference intensity for control effect, and target device association words are used to clarify the specific smart device corresponding to the user's preference.

[0022] In the above technical solution, the preset fields are subdivided into preference keywords, preference degree descriptions, and target device related words. From the three dimensions of demand tendency, preference intensity, and target device, the voice commands containing user preferences are accurately determined, which improves the parsing accuracy and distinguishability of preference information in voice commands, reduces misidentification, and makes the learned user preferences more accurate.

[0023] Secondly, an apparatus for learning user preferences is provided, the apparatus comprising: An interaction module is used to receive voice commands from users, which are used to control smart devices. The recognition module is used to detect whether the voice command includes a preset field, the preset field being used to represent the user's preference information; The processing module is configured to, when the preset field is included in the voice command, obtain the user's user preference information for the smart device based on the voice command.

[0024] In conjunction with the second aspect, in some possible implementations, the processing module is also used to determine tuple data based on voice commands. The tuple data includes: user identification information, target control action, and target control conditions required to execute the target control action; and to obtain user preference information based on the tuple data.

[0025] In conjunction with the second aspect and the above implementation methods, in some possible implementation methods, the interaction module is further used to acquire target information, which includes the current user's voice command and / or the current environment information; the processing module is further used to determine whether the target information meets the target control conditions; when the target information meets the target control conditions, based on the target information, acquire the target control action corresponding to the target control conditions; based on the target control action, generate a target control instruction; and based on the target control instruction, control the target device corresponding to the target control instruction to execute the target control action.

[0026] In conjunction with the second aspect and the above implementation methods, in some possible implementation methods, the target control conditions include voice triggering conditions and / or environment triggering conditions. The processing module is further configured to determine that the target information meets the target control conditions if the target information is a voice command and the voice command meets the voice triggering conditions; determine that the target information meets the target control conditions if the target information is current environment information and the environment information meets the environment triggering conditions; and determine that the target information meets the target control conditions if the target information includes both voice commands and current environment information.

[0027] In conjunction with the second aspect and the above implementation methods, in some possible implementation methods, the interaction module is further used to obtain the user's historical voice commands based on the user's voice commands when the voice commands do not include preset fields, wherein the user intent corresponding to the historical voice commands is the same as the user intent corresponding to the voice commands; and to obtain the historical control actions corresponding to the historical voice commands based on the historical voice commands; the processing module is further used to obtain the user's user preference information based on the historical control actions and voice commands.

[0028] In conjunction with the second aspect and the above implementation methods, in some possible implementation methods, the processing module is also used to determine the target historical control action based on the execution frequency of various control actions in the historical control actions. The target historical control action refers to the historical control action whose execution frequency reaches a preset frequency. Based on the target historical control action, the module obtains the historical environment information when the target historical control action is executed. Based on the voice command, the historical environment information, and the target historical control action, the module obtains the user's user preference information.

[0029] In conjunction with the second aspect and the above-described implementation, in some possible implementations, the device further includes a memory module for storing user preference information for the smart device in a preset preference memory.

[0030] In conjunction with the second aspect and the above implementation methods, in some possible implementation methods, the processing module is also used to monitor in real time the user's new voice commands and corresponding control actions for the smart device, analyze whether the new voice commands and control actions are consistent with the user preference information already stored in the preference memory; if they are inconsistent, update the corresponding user preference information in the preference memory based on the new voice commands and corresponding control actions; and record the update time and update content in order to trace the user's preference change trajectory.

[0031] Combining the second aspect and the above implementation methods, in some possible implementation methods, the preset fields include preference keywords, preference degree descriptors, and target device association words. Among them, preference keywords are used to characterize the user's tendency to control smart devices, preference degree descriptors are used to characterize the user's preference intensity for control effect, and target device association words are used to clarify the specific smart device corresponding to the user's preference.

[0032] Thirdly, a smart device is provided, including a memory and a processor. The memory is used to store executable program code, and the processor is used to call and run the executable program code from the memory, causing the smart device to perform the methods of the first aspect or any possible implementation thereof.

[0033] Fourthly, a computer program product is provided, comprising: computer program code, which, when run on a computer, causes the computer to perform the methods described in the first aspect or any possible implementation thereof.

[0034] Fifthly, a computer-readable storage medium is provided that stores computer program code, which, when executed on a computer, causes the computer to perform the methods described in the first aspect or any possible implementation thereof. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the system architecture of a voice interaction system provided in an embodiment of this application; Figure 2 This is a schematic flowchart illustrating a method for learning user preferences provided in an embodiment of this application; Figure 3 This is a schematic flowchart illustrating another method for learning user preferences provided in an embodiment of this application; Figure 4 This is a schematic flowchart illustrating a control method for an intelligent device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of a device for learning user preferences provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of a smart device provided in an embodiment of this application. Detailed Implementation

[0036] The technical solutions in this application will be clearly and thoroughly described below with reference to the accompanying drawings. In the description of the embodiments of this application, unless otherwise stated, " / " means "or," for example, A / B can mean A or B. "And / or" in the text is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Furthermore, in the description of the embodiments of this application, "multiple" refers to two or more than two.

[0037] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0038] In existing smart home systems, most systems only support users manually setting preferred scene modes, such as "coming home" mode; or they rely on users repeatedly performing the same actions and then using simple frequency statistics to recommend fixed options. This existing "passive" preference learning method has serious limitations: the system cannot proactively capture and solidify users' long-term preferences from single, complex natural language expressions; and the learned preferences are usually isolated data points, failing to form a structured user profile and thus unable to be flexibly invoked across different scenarios and devices; furthermore, it cannot deeply correlate user preferences with their application scenarios and objects, resulting in rigid and inaccurate personalized services.

[0039] To address the technical problems existing in the prior art, embodiments of this application provide a method, apparatus, and smart device for learning user preferences. The method, upon receiving a voice command from a user to control the smart device, detects whether the voice command includes a preset field representing the user's preference information. If the voice command includes the preset field, the user's preference information for the smart device is obtained based on the voice command. This method provides a convenient way for users to proactively learn their preferences, improving the efficiency of learning user preferences for smart devices.

[0040] The following is combined Figures 1 to 3The method for learning user preferences provided in the embodiments of this application will be described in detail.

[0041] First, this application provides a voice interaction system. Figure 1 This is a schematic diagram of the system architecture of a voice interaction system provided in an embodiment of this application. Figure 1 As shown, the voice interaction system 100 includes an interaction module 101, a speech recognition module 102, a natural language understanding module 103, an intent recognition module 104, a preference declaration recognizer 105, a preference parsing engine 106, a preference memory 107, a preference trigger 108, and a device control module 109.

[0042] In existing technologies, after receiving a user's voice command, the voice recognition module converts the user's voice command into text; then, the natural language understanding module and intent recognition module perform intent recognition on the text to determine the control command; finally, the device is controlled by the device control module. Compared to existing voice interaction systems, the voice interaction system 100 of this application adds a preference declaration recognizer 105, a preference parsing engine 106, a preference memory 107, and a preference trigger 108.

[0043] The functions of each module in the voice interaction system 100 will be introduced below.

[0044] The interaction module 101 serves as the direct interaction entry point between the voice interaction system 100 and the user and the external environment. Through various channels such as voice, software programs, and sensors, it enables the input of user commands, real-time information collection, status perception, and feedback output, facilitating two-way information transmission between humans and the system, and between the system and the environment. This provides fundamental interactive support for subsequent processing, decision-making, and execution. In this embodiment, the interaction module 101 may also be referred to as the user interaction layer, etc.

[0045] After receiving a speech signal, the speech recognition module 102 (Automatic Speech Recognition, ASR) converts the speech signal into text content, completing the accurate conversion from sound to text, and providing processable text data for subsequent natural language understanding, instruction execution, information retrieval and other processes.

[0046] The Natural Language Understanding (NLU) module 103 is used to perform deep analysis on the text output by the speech recognition module 102, completing word segmentation, syntactic analysis, entity extraction, and converting human natural language into machine-executable structured semantic information.

[0047] The intent recognition module 104 is used to analyze the user's intent based on the semantic information obtained from NLU parsing, clarify the user's direct needs, and match the corresponding control operations.

[0048] The preference declaration recognizer 105 is used to identify preference declarations expressed by the user and extract information related to personalization settings from the speech. In this embodiment, the preference declaration recognizer 105 can also be referred to as a recognition module.

[0049] The preference parsing engine 106 is used to parse the identified preference information into a tuple structure to form structured user preference information. In this embodiment, the preference parsing engine 106 can also be referred to as a processing module.

[0050] The preference memory 107 is used to store and remember the user preference information generated by the preference parsing engine 106.

[0051] The preference trigger 108 is associated with the preference memory 107 and is used to actively call the corresponding user preference information to determine the control command when the trigger condition is detected.

[0052] The device control module 109 is used to precisely control and execute related devices based on the user intent obtained by the intent recognition module 104 or the user preference information retrieved by the preference trigger 108.

[0053] Optionally, the voice interaction system 100 provided in this application embodiment can be configured in the smart device to be controlled, or configured in the processor of the smart device; or configured in the chip of the processor of the smart device; or the voice interaction system 100 can be independently configured outside the smart device to be controlled, and establish a connection between the central control management device and the smart device to be controlled.

[0054] Figure 2 This is a schematic flowchart illustrating a method for learning user preferences provided in an embodiment of this application. It should be understood that this method can be applied to... Figure 1 The voice interaction system shown.

[0055] For example, such as Figure 2 As shown, the method 200 includes: S201 receives voice commands from the user, which are used to control smart devices.

[0056] For example, the user's voice commands are received through the interaction module.

[0057] Optionally, the process of receiving user voice commands can be triggered by wake word detection. For example, after detecting a preset wake word, the system immediately receives and collects subsequent voice commands issued by the user, which include control operations such as turning the smart device on or off, and adjusting it.

[0058] S202, detect whether the voice command includes a preset field, which is used to represent the user's preference information.

[0059] Preset fields, also known as preference declaration fields, refer to information related to personalized settings in voice commands. Examples include phrases like "I like…" or "Remember, when I… I always…". Preset fields can also be called preference declarations or explicit preference information.

[0060] For example, the preset fields may include preference keywords, preference degree descriptions, and target device association words. The preference keywords are used to characterize the user's tendency to control the smart device, the preference degree descriptions are used to characterize the user's preference intensity for the control effect, and the target device association words are used to specify the specific smart device corresponding to the user's preference.

[0061] For example, preference keywords can include like, preference, habit, want, more inclined, etc.; preference degree description words can include very, especially, quite, average, slightly, must, never, etc. If there are no obvious degree words in the voice command, it can be understood as the default strength; if a field including especially like is detected, the degree word can be determined to be especially; target device association words are used to specify the corresponding device, which can be master bedroom air conditioner, living room TV, kitchen range hood, bedroom humidifier, etc.

[0062] For example, after receiving a voice command, the voice command is converted into text information by the voice recognition module in the voice interaction system. Then, the natural speech understanding module performs word segmentation and syntactic analysis on the text information to obtain structured semantic information. Then, the preference declaration recognizer analyzes the structured semantic information to determine whether the structured semantic information includes any of the preset fields mentioned above.

[0063] For example, if a user's voice command is received as "I like the air conditioner set to 26 degrees Celsius and the lowest fan speed when I sleep," after analysis, it can be determined that the voice command includes a preset field indicating preference information: "I like… when I…". If the received voice command is "Air conditioner set to 26 degrees Celsius and the lowest fan speed," which does not contain the preset field, it can be identified as a regular control command; therefore, the air conditioner can be directly controlled to perform the action of setting it to 26 degrees Celsius and the lowest fan speed.

[0064] It should be noted that the preset fields can be configured by the user or the manufacturer. The above is only an example description, and the embodiments of this application do not specifically limit the preset fields. When the preset fields include multiple fields, it can be determined that the voice command includes any one or more preset fields when the voice command is detected to include any one or more preset fields.

[0065] In this embodiment, the preset fields are subdivided into preference keywords, preference degree descriptions, and target device related words. Voice commands containing user preferences are accurately determined from three dimensions: demand tendency, preference intensity, and target device. This improves the parsing accuracy and distinguishability of preference information in voice commands, reduces misidentification, and makes the learned user preferences more accurate.

[0066] S203, when the voice command includes preset fields, obtains the user's preference information for the smart device based on the voice command.

[0067] For example, if it is determined that the user's voice command includes a preset field, it indicates that the user has a need for active preference learning. Then, preference learning can be performed on the user based on the voice command to obtain the user's preference information for the smart device, that is, the user's usage preference information when using the smart device.

[0068] In one implementation, the process of obtaining user preference information for smart devices based on voice commands may specifically include: Based on voice commands, determine tuple data, which includes: user identification information, target control action, and target control conditions required to execute the target control action; Based on tuple data, user preference information is obtained.

[0069] The user identification information may include the user's voiceprint feature or user image, etc., to distinguish different users; the target control action may specifically include the target control device and the target operating parameters.

[0070] For example, if it is determined that the voice command includes preset fields, the voice command is analyzed by a preference parsing engine to obtain the tuple data indicated by the voice command; then, user preference information corresponding to the user can be obtained based on the tuple data.

[0071] For example, if a user inputs the voice command "I prefer the master bedroom air conditioner to be set to 26 degrees Celsius when I sleep," firstly, the voiceprint features of the voice command or the user's image can be analyzed to obtain the user's identification information as User 1 in the multi-byte dataset. Then, parsing the voice command reveals that the target control device is the master bedroom air conditioner, the target operating parameter is 26 degrees Celsius, and the target control condition for executing the target control action is "sleeping." Finally, the user's preference information can be obtained as "Target control condition = sleeping, Target control action = master bedroom air conditioner turned on at 26 degrees Celsius."

[0072] Optionally, the aforementioned tuple data may be triple data including user identification information, target control action, and target control conditions required to execute the target control action, or it may include user preference-related data other than the above three types of data, such as device execution priority, user historical habits, etc.

[0073] In this embodiment, based on voice commands, a tuple of data including user identification information, target control actions, and target control conditions is obtained; then, user preference information is obtained based on the tuple data. By structuring the user preference information into tuple data containing user identification, target control actions, and control conditions, the expression of user preferences becomes more standardized and the logic clearer, facilitating subsequent storage, matching, and retrieval, thereby improving the manageability and reusability of preference information.

[0074] In one implementation, user preference information for smart devices is stored in a preset preference memory.

[0075] For example, during use, the voice interaction system analyzes the user preference information of different users, and then stores each user's user preference information for smart devices in a preset preference memory.

[0076] For example, user preference information can be stored in the preference memory in the form of "User ID = User N; Target control condition = X; Target control device = Y; Target operating parameter = Z".

[0077] In this embodiment, the learned user preference information is uniformly stored in a preset preference memory, realizing centralized and persistent storage of preference information, which facilitates subsequent multiple calls, comparisons and updates, and provides data support for continuous learning of user habits and the realization of personalized intelligent control.

[0078] Based on learning user preferences, this application also proposes a process for applying user preference information. In one implementation, user preference information can be used to control smart devices such as home appliances. The process of controlling the devices may include: Acquire target information, which includes the current user's voice commands and / or current environment information; Determine whether the target information meets the target control conditions; When the target information meets the target control conditions, the target control action corresponding to the target control conditions is obtained based on the target information; Based on the target control action, generate target control instructions; and based on the target control instructions, control the target device corresponding to the target control instructions to execute the target control action.

[0079] For example, the current user's voice commands and current environmental information are monitored. The current environmental information may include time, season, weather, temperature, and humidity. After obtaining the target information, it can be compared with target control conditions in the preference memory to determine whether the current target information matches the target control conditions. That is, it determines whether the current user's voice commands and / or current environmental information conform to the target control conditions stored in the preference memory. If the target information conforms to the target control conditions, the target control action corresponding to the target control conditions can be retrieved from the preference memory based on the target information. Then, a corresponding target control command can be generated based on the extracted target control action, and the target device can be controlled according to the target control command to execute the target control action and meet the user's preference requirements.

[0080] Optionally, while generating the target control command, a target prompt message can also be generated. This prompt message asks the user whether they need to control the target device to perform the target control action. If confirmation is received from the user, the target device can be controlled according to the target control command. If negative feedback is received, or if no feedback is received within a preset time (e.g., 5 seconds), the target device will not be controlled. This scheme ensures that device control fully meets the user's current actual needs, avoiding ineffective operations caused by direct control.

[0081] In this embodiment, after obtaining user preferences, the system further combines the current voice command or environmental information to determine whether the preset control conditions are met, and automatically generates control commands to drive the target device to perform corresponding actions, thereby realizing automated and intelligent control based on user preferences without requiring the user to repeatedly issue commands. When the target information includes the current environmental information and matches the user preferences, the system can meet the user's needs in advance without the user inputting commands, providing the user with proactive and accurate personalized services, and improving ease of use and intelligent experience.

[0082] In one implementation, the target control conditions include voice triggering conditions and / or environmental triggering conditions, and the process of determining whether the target information meets the target control conditions may specifically include: If the target information is a voice command, and the voice command meets the voice triggering conditions, then the target information is determined to meet the target control conditions. If the target information is the current environment information, and the environment information meets the environment triggering conditions, then the target information is determined to meet the target control conditions. If the target information includes both voice commands and current environment information, then when the voice command meets the voice triggering condition and the environment information meets the environment triggering condition, the target information is determined to meet the target control condition.

[0083] For example, target control conditions can be divided into voice trigger conditions and environmental trigger conditions. When determining whether target information meets the target control conditions, the system will adopt corresponding judgment logic according to the type of target information. When the target information is only a voice command, the system parses and matches the content, intent, keywords, etc. of the voice command. If the voice command meets the preset voice trigger conditions, the target information is determined to meet the target control conditions. When the target information is only current environmental information, the system collects environmental data such as temperature, humidity, light, human presence, and time through sensors. If the environmental information meets the preset environmental trigger conditions, the target information is determined to meet the target control conditions. When the target information contains both voice commands and current environmental information, the system needs to verify both the voice commands and environmental information simultaneously. Only when the voice command meets the voice trigger conditions and the current environmental information meets the environmental trigger conditions will the target information be finally determined to meet the target control conditions, thereby ensuring the accuracy and rationality of the control commands.

[0084] For example, after obtaining the target information, if the target information is the current user's voice command, the current user's identification information can be analyzed; then, the user preference information corresponding to the current user's identification information can be retrieved from the preference memory, and the target control conditions can be extracted from the user preference information; then, it can be determined whether the current user's voice command meets the target control conditions, and the judgment result can be obtained.

[0085] For example, the system detects that user 1 inputs the voice command "I want to sleep"; based on user 1's voiceprint information, it extracts the user preference information corresponding to user 1 from the preference memory bank as "target control condition = sleep, target control action = master bedroom air conditioner turned on at 26℃". Then, it detects that the voice command "I want to sleep" matches the voice trigger condition "target control condition = sleep", and then obtains the corresponding target control action as "master bedroom air conditioner turned on at 26℃". In this way, it can generate a target control command to control the master bedroom air conditioner to turn on at 26℃, so as to control the device to perform an action that conforms to the user's preference.

[0086] For example, after obtaining the target information, if the target information is the current environment information, it is detected whether there is a user in the current scene; if there is a user, the user's identification information is obtained, and the user preference information corresponding to the current user's identification information is retrieved from the preference memory, and the target control conditions are extracted from the user preference information; then it is determined whether the current environment information meets the target control conditions, and the judgment result is obtained.

[0087] For example, the system detects that the current time is 11 PM and that user 2 is in the current scene; based on user 2's identification information, it extracts the user preference information corresponding to user 2 from the preference memory as "target control condition = 11 PM / sleep, target control action = turn off living room light". Then, it detects that the time 11 PM matches the environmental trigger condition "target control condition = 11 PM", so the corresponding target control action is "turn off living room light". Then, it can generate a target control command to control the living room light to turn off, so as to control the device to perform an action that conforms to the user's preference.

[0088] For example, when the user's voice command and current environmental parameters are obtained, the user's voice command and current environmental parameters are matched with the target control conditions stored in the preference memory bank to obtain a judgment result.

[0089] For example, user 3's user preference information is "target control condition = daytime, sleep, target control action = close the master bedroom curtains"; when receiving user 3's voice command "I want to sleep" and detecting that the current environment is daytime, the corresponding target control action can be determined as "close the master bedroom curtains", and then a target control command can be generated to instruct the master bedroom curtains to be closed, so as to control the device to perform an action that conforms to the user's preference.

[0090] In this embodiment, the control conditions are refined into voice trigger conditions and / or environmental trigger conditions, and single or combined condition judgments are supported, making the preference trigger mechanism more flexible and more in line with actual use scenarios. It can adapt to diverse interaction methods and environmental changes, and enhance the robustness and applicability of the control logic.

[0091] In one implementation, if the voice command does not include a preset field, the user's historical voice commands are obtained based on the user's voice command, and the user intent corresponding to the historical voice commands is the same as the user intent corresponding to the voice command. Based on historical voice commands, obtain the historical control actions corresponding to the historical voice commands; Based on historical control actions and voice commands, user preference information is obtained.

[0092] For example, if the semantic command input by the user does not include a preset field for indicating preference information, in order to continuously learn the user's usage preferences, the historical voice command that is the same as the user's intention indicated by the current voice command can be obtained from the historical interaction information based on the user's current voice command input; then, the corresponding historical control action can be matched based on the historical voice command; and the user's user preference information can be obtained by combining the historical control action and the currently received voice command.

[0093] For example, when a user inputs the voice command "I feel a little hot," the user's intention is analyzed as feeling hot. Based on this user intention, historical voice commands such as "I'm too hot, set the air conditioner to 25°C" and "The temperature is too high" can be detected from historical interaction records. Then, the historical control action of setting the air conditioner temperature to 25°C after the user inputs the historical voice command can be obtained. Furthermore, by combining the user's current voice command "I feel a little hot" with the historical control action of setting the air conditioner temperature to 25°C, the user's usage preferences can be analyzed and learned to obtain the user's user preference information.

[0094] In this embodiment of the application, when the voice command does not include a preset field, the user preference information is obtained by retrieving the historical voice commands of the agreement map and the historical control actions corresponding to the historical voice commands. This makes up for the problem of missing explicit preference information, realizes the mining and learning of the user's potential and implicit preferences, expands the scope of user preference acquisition, and improves the completeness and continuity of preference learning.

[0095] In one implementation, the process of obtaining user preference information based on historical control actions and voice commands may specifically include: Based on the execution frequency of various control actions in the historical control actions, the target historical control action is determined. The target historical control action is the historical control action whose execution frequency reaches the preset frequency. Based on the target's historical control actions, obtain historical environment information when the target's historical control actions are executed; Based on voice commands, historical environmental information, and target historical control actions, user preference information is obtained.

[0096] For example, after determining the historical control actions corresponding to historical voice commands, the execution frequency of various historical control actions can be statistically obtained, target historical control actions that reach the preset frequency can be filtered out, and various historical environmental information such as temperature, humidity, time, and personnel status can be obtained simultaneously when these target actions are executed. Finally, combined with the needs clearly stated in the user's voice commands, the voice commands, target control actions and historical environmental information can be associated to extract contextualized and personalized user preferences. At the same time, the preference information can be continuously optimized through the user's subsequent adjustment behavior.

[0097] For example, if a user frequently executes the command "turn on the air conditioner at 7:30 PM and set the temperature to 26°C" over the past 30 days, and considering the high temperature and humidity of the evening, the presence of people in the living room, and related voice commands, the corresponding air conditioner preferences can be extracted. Another example is the use of frequently turning on the living room lights in the morning and evening on weekdays, along with the corresponding lighting, time environment, and voice commands, to extract lighting preferences for different time periods. Finally, based on frequent on / off actions, corresponding lighting and time environment, and voice commands, the curtain control preferences for different morning and evening scenarios can be identified.

[0098] In this embodiment, the high-frequency historical control actions are filtered by statistically analyzing their execution frequency, and user preference information is generated by combining the historical environment information corresponding to the historical control actions. This makes the obtained user preference information more in line with the user's actual usage habits, more representative and reliable, and avoids interference from accidental operations on preference learning.

[0099] In one implementation, the system monitors new voice commands and corresponding control actions of the user on the smart device in real time, and analyzes whether the new voice commands and control actions are consistent with the user preference information already stored in the preference memory. If there is a discrepancy, the user preference information in the preference memory will be updated based on the new voice command and the corresponding control action; and the update time and update content will be recorded to facilitate the tracking of the user's preference change trajectory.

[0100] For example, to ensure that the user preference information stored in the preference memory always matches the user's latest needs, it is necessary to monitor in real time any new voice commands issued by the user to the smart device, as well as the corresponding control actions triggered by the new voice commands. The monitoring process must ensure real-time performance and accuracy to avoid missing valid commands and actions or misjudging invalid operations. After monitoring is completed, the semantics of the new voice commands are parsed to clarify the user's core needs, and the corresponding control actions are associated with them. Then, a precise comparison is made with the user preference information in the same scenario already stored in the preference memory to determine whether the two are consistent. Specifically, it can be determined whether the new voice commands and control actions are consistent with the command intent, control parameters, applicable scenarios, and other dimensions indicated in the preference memory.

[0101] For example, if the parsed new voice commands and control actions are inconsistent with the existing user preference information, it indicates that the user preference has changed. Based on the new commands and control actions, the corresponding original preference information in the preference memory bank needs to be updated. During the update, the core reasonable content should be retained and the deviation parts should be corrected to ensure that the updated preference matches the user's current needs. At the same time, the specific time of this update, the original preference content, the new commands and actions, and the updated preference content should be recorded in detail to form a complete update record, which is convenient for tracing the change trajectory of user preferences and analyzing the change pattern of user preferences.

[0102] For example, suppose the preference memory stores user preference information as "target control condition = weekday evening 19:30, target control action = air conditioner turned on cooling 26℃". Subsequently, if the user adds a new voice command "turn on air conditioner at 8 pm, cooling 25℃" and the corresponding control action, after parsing, it is found that the command intent and control parameters are inconsistent with the existing preferences. Therefore, the air conditioner preference in the memory is updated to "target control condition = weekday evening 20:00, target control action = air conditioner turned on cooling 25℃". At the same time, the update time, the original preference, the new command and action, and the updated preference are recorded.

[0103] In this embodiment, by monitoring the user's new voice commands and control actions in real time, inconsistent information in the preference memory is promptly detected and updated. At the same time, the update time and content are recorded to ensure that the user's preference information is always up-to-date and most accurate, and the trajectory of preference changes can be traced, thereby improving the system's adaptability and traceability.

[0104] In another implementation, after obtaining the user's preference information and generating a comprehensive preference memory, more advanced home services can be provided to the user by combining the user preference information, such as providing health advice or device energy management, etc.

[0105] In summary, in this embodiment, after receiving a user's voice command to control a smart device, it detects whether the voice command includes a preset field; and if the preset field is detected, it determines the user's preference information based on the voice command. By receiving the user's voice command to control the smart device and detecting the preset field representing preference information, it is possible to directly and accurately extract the user's preference information for the smart device from the voice command, achieving rapid identification and acquisition of user preferences, improving the efficiency and accuracy of the smart device's understanding of user intentions, and enhancing the user's learning efficiency regarding the smart device's usage preferences.

[0106] Figure 3 This is a schematic flowchart illustrating another method for learning user preferences provided in an embodiment of this application. It should be understood that this method can be applied to... Figure 1 The voice interaction system shown.

[0107] For example, such as Figure 3 As shown, the method 300 includes: S301 receives user voice commands, which are used to control smart devices.

[0108] For example, the user's voice commands are received through the interaction module of the voice interaction system.

[0109] S302 converts voice commands into text information.

[0110] For example, after receiving a user's voice command, the voice command can be processed by a voice recognition module to convert the voice signal of the voice command into text information.

[0111] In one implementation, during the conversion process, unclear pronunciation, connected speech, and colloquial expressions in the voice commands can be corrected to ensure that the converted text information accurately reflects the user's voice commands. Figure 1 This is to avoid misjudgments in subsequent steps due to conversion errors, while retaining key information in the voice commands to provide accurate textual basis for subsequent preset field judgments.

[0112] S303, determine whether the text information corresponding to the voice command contains a preset field. If yes, execute S304; otherwise, execute S306.

[0113] The preset fields are used to represent user preference information. Specifically, these fields may include preference keywords, preference intensity descriptions, and target device association terms. Preference keywords characterize the user's tendency to control smart devices, preference intensity descriptions characterize the strength of the user's preference for control effects, and target device association terms specify the specific smart device corresponding to the user's preference.

[0114] For example, keyword retrieval and matching are performed on the converted text information to determine whether it contains preset fields pre-defined by the system. When preset fields are detected in the text information, it indicates that there is a preference expression in the user instruction, and the process proceeds to the preference information extraction process in S304. When preset fields are not detected, it indicates that the user instruction is only a routine control requirement and there is no need to extract preference information, and the process proceeds directly to the user intent parsing process in S306.

[0115] S304 determines tuple data based on voice commands.

[0116] For example, after confirming that the text information corresponding to the voice command contains preset fields, the voice command and the converted text information are deeply analyzed to extract key core data and construct and determine tuple data. This tuple data is structured data used to represent user preference-related information, specifically including three core dimensions: user identification information, used to distinguish the preferences of different users and ensure the relevance of preference information; target control action, indicating the specific operation that the user wants the smart device to perform; and target control conditions required to perform the target control action, such as environmental trigger conditions and voice trigger conditions.

[0117] S305, based on tuple data, obtains user preference information and stores the user preference information in the preference memory.

[0118] For example, the determined plural data is further analyzed and refined. Combined with preset preference parsing rules, user preference information for smart devices is mined and generated from three dimensions: user identifier, target control action, and target control conditions. This preference information can accurately reflect the user's control needs, preference intensity, and applicable scenarios for smart devices. At the same time, the generated user preference information is stored in the system's built-in preference memory according to a preset format, realizing centralized and persistent storage of preference information.

[0119] S306 determines user intent based on voice commands.

[0120] For example, if it is confirmed that the text information corresponding to the voice command does not contain a preset field, it indicates that the user needs to directly control the smart device. In this case, the intent of the voice command and the converted text information can be parsed. The intent recognition module can identify the core purpose of the user's voice command, that is, determine the specific control effect or specific task that the user wants the smart device to achieve.

[0121] For example, when a user issues a voice command to "turn on the living room light," which does not contain any preset fields, the user's core intent is to control the living room light to turn on.

[0122] S307 controls the target device based on user intent.

[0123] For example, after clarifying the user's intent indicated by the voice command, the system can match the corresponding control logic and control instructions based on the parsed user intent, send control signals to the target device indicated by the user intent, and control the target device to perform the corresponding operation to meet the user's current usage needs.

[0124] In summary, in this embodiment, by first receiving the user's voice commands to control the smart device and converting them into text, it accurately determines whether the text contains a preset preference field. If the field is included, it directly constructs tuple data based on the voice command to obtain and store the user preference information. If the field is not included, it directly controls the target device according to the user's intention. This approach can efficiently and accurately learn and save user preferences from voice commands, achieving the effect of "speak once, remember forever," while also ensuring normal voice control even without explicit preference information. It balances the intelligence of user preference learning with the convenience of daily control, improving the practicality and personalization of smart device voice interaction.

[0125] In combination with the above Figure 2 or Figure 3 The method shown can be used to learn user preferences and obtain user preference information; based on this, it can be combined with the following... Figure 4The control method for the smart device shown applies user preference information to improve the control logic of the voice interaction system.

[0126] Figure 4 This is a schematic flowchart illustrating a control method for a smart device provided in an embodiment of this application. It should be understood that this method can be applied to... Figure 1 The voice interaction system shown.

[0127] For example, such as Figure 4 As shown, the method 400 includes: S401, Obtain target information, which includes voice commands and / or current environment information.

[0128] For example, the system monitors and acquires target information in real time to trigger the control of smart devices. This target information can be acquired in a flexible manner, such as acquiring the user's voice command alone, acquiring the current environment information alone, or acquiring both the voice command and the current environment information simultaneously.

[0129] Among them, voice commands are received by the voice acquisition module of the smart device and obtained after voice-to-text processing to reflect the user's active control needs; current environmental information is collected by the device's built-in or external environmental sensors (such as temperature and humidity sensors, light sensors, human body sensors, etc.), including but not limited to ambient temperature and humidity, light intensity, human presence status, time information, etc.

[0130] S402, when the target information meets the target control conditions, retrieve the target control action corresponding to the target control conditions from the preference memory.

[0131] For example, the system first retrieves pre-stored target control conditions, including voice trigger conditions and / or environmental trigger conditions. Then, it performs a precise comparison between the acquired target information and the target control conditions to determine whether the currently acquired target information meets the target control conditions. When it is determined that the target information meets the conditions, the system accesses the preference memory through a preference trigger and retrieves the corresponding target control action according to the target control conditions.

[0132] For example, if the target information is a single voice command, it is determined whether the voice command matches the voice triggering condition; if the target information is a single current environment information, it is determined whether the environment information matches the environment triggering condition; if the target information contains both voice commands and environment information, then both the voice triggering condition and the environment triggering condition must be met simultaneously for the target information to be deemed to meet the target control conditions.

[0133] S403 generates target control instructions based on target control actions.

[0134] For example, after obtaining the target control action from the preference memory, the system generates the target control instruction based on the specific content and execution requirements of the target control action, combined with the target device's model, functional parameters, and communication protocol.

[0135] For example, the target control command may include information such as the identification information of the target device, details of the control action, and execution priority, to ensure that the control command is executable and accurate.

[0136] Optionally, after generating the target control command, the generated control command can also be format-validated to avoid problems such as syntax errors or missing parameters, ensuring that subsequent control commands can be accurately parsed and executed by the target device.

[0137] S404, based on target control commands, controls the target device to perform target control actions.

[0138] For example, the system sends the generated target control command to the target device indicated by the target control command through a preset communication method. After receiving the control command, the target device parses the command and executes the corresponding target control action according to the control action details specified in the command.

[0139] In summary, in this embodiment, by acquiring voice commands and / or current environmental information as target information, and when the target information meets the target control conditions, the corresponding target control action is retrieved from the preference memory and a control command is generated, thereby controlling the target device to execute the action. This achieves automated intelligent control based on user preferences, eliminating the need for users to repeatedly issue control commands. It can fully utilize the stored user preference information and improve the convenience, automation, and personalized experience of intelligent device control, making the device response more in line with user habits and actual usage scenarios.

[0140] The above text combined Figures 1 to 4 The method for learning user preferences provided in the embodiments of this application is described in detail below; the following will be combined with Figure 5 and Figure 6 The apparatus embodiments of this application are described in detail below. It should be understood that the apparatus in the embodiments of this application can perform the various methods described in the foregoing embodiments of this application, that is, the specific working processes of the various products described below can be referred to the corresponding processes in the foregoing method embodiments.

[0141] Figure 5 This is a schematic diagram of the structure of a device for learning user preferences provided in an embodiment of this application.

[0142] For example, such as Figure 5 As shown, the device 500 includes: Interaction module 101 is used to receive user voice commands, which are used to control smart devices; The recognition module 502 is used to detect whether the voice command includes a preset field, the preset field being used to represent the user's preference information; The processing module 503 is used to obtain the user's user preference information for the smart device based on the voice command when the preset field is included in the voice command.

[0143] In one possible implementation, the processing module 503 is further configured to determine tuple data based on voice commands, the tuple data including: user identification information, target control action, and target control conditions required to execute the target control action; and obtain user preference information based on the tuple data.

[0144] In one possible implementation, the interaction module 101 is further configured to acquire target information, including the current user's voice command and / or the current environment information; the processing module 503 is further configured to determine whether the target information meets the target control conditions; when the target information meets the target control conditions, based on the target information, acquire the target control action corresponding to the target control conditions; based on the target control action, generate a target control instruction; and based on the target control instruction, control the target device corresponding to the target control instruction to execute the target control action.

[0145] In one possible implementation, the target control conditions include voice triggering conditions and / or environment triggering conditions. The processing module 503 is further configured to determine that the target information meets the target control conditions if the target information is a voice command and the voice command meets the voice triggering conditions; determine that the target information meets the target control conditions if the target information is current environment information and the environment information meets the environment triggering conditions; and determine that the target information meets the target control conditions if the target information includes both voice commands and current environment information.

[0146] In one possible implementation, the interaction module 101 is further configured to, when the voice command does not include a preset field, obtain the user's historical voice commands based on the user's voice command, wherein the user intent corresponding to the historical voice command is the same as the user intent corresponding to the voice command; and obtain the historical control action corresponding to the historical voice command based on the historical voice command; the processing module 503 is further configured to obtain the user's user preference information based on the historical control action and the voice command.

[0147] In one possible implementation, the processing module 503 is further configured to determine a target historical control action based on the execution frequency of various control actions in the historical control actions, wherein the target historical control action refers to a historical control action whose execution frequency reaches a preset frequency; based on the target historical control action, obtain historical environment information when the target historical control action is executed; and based on the voice command, historical environment information and target historical control action, obtain the user's user preference information.

[0148] In one possible implementation, the device further includes a memory module 504, which stores user preference information for the smart device in a preset preference memory.

[0149] In one possible implementation, the processing module 503 is further configured to monitor in real time the user's new voice commands and corresponding control actions for the smart device, analyze whether the new voice commands and control actions are consistent with the user preference information already stored in the preference memory; if they are inconsistent, update the corresponding user preference information in the preference memory based on the new voice commands and corresponding control actions; and record the update time and update content in order to trace the user's preference change trajectory.

[0150] In one possible implementation, the preset fields include preference keywords, preference degree descriptors, and target device association words. The preference keywords are used to characterize the user's tendency to control the smart device, the preference degree descriptors are used to characterize the user's preference intensity for the control effect, and the target device association words are used to specify the specific smart device corresponding to the user's preference.

[0151] It should be noted that the aforementioned device for learning user preferences is embodied in the form of functional units. The term "module" here can be implemented in software and / or hardware, without specific limitations.

[0152] For example, a "module" can be a software program, a hardware circuit, or a combination of both that implements the above functions. The hardware circuit may include an application-specific integrated circuit (ASIC), electronic circuits, a processor (e.g., a shared processor, a proprietary processor, or a group processor) and memory for executing one or more software or firmware programs, integrated logic circuits, and / or other suitable components that support the described functions.

[0153] Therefore, the units of the various examples described in the embodiments of this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0154] Figure 6 This is a schematic diagram of the structure of a smart device provided in an embodiment of this application.

[0155] For example, such as Figure 6 As shown, the smart device 600 includes a memory 601 and a processor 602. The memory 601 stores executable program code 603, and the processor 602 is used to call and execute the executable program code 603 to perform a method for learning user preferences.

[0156] Furthermore, embodiments of this application also protect an apparatus that may include a memory and a processor, wherein the memory stores executable program code, and the processor is used to call and execute the executable program code to perform a method for learning user preferences provided in embodiments of this application.

[0157] This embodiment can divide the device into functional modules based on the above method example. For example, each module can correspond to a separate function, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0158] When the functional modules are divided according to their respective functions, the device may also include an interaction module, a processing module, etc. It should be noted that all relevant content of each step involved in the above method embodiments can be referenced to the functional description of the corresponding functional module, and will not be repeated here.

[0159] It should be understood that the apparatus provided in this embodiment is used to perform the above-described method for learning user preferences, and therefore can achieve the same effect as the above-described implementation method.

[0160] When using integrated units, the device may include a processing module and a storage module. When applied to a smart device, the processing module can be used to control and manage the actions of the smart device. The storage module can be used to support the execution of relevant program code by the smart device.

[0161] The processing module may be a processor or a controller, which can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor may also be a combination of functions that implement computing capabilities, such as a combination of one or more microprocessors, a combination of digital signal processing (DSP) and a microprocessor, etc., and the storage module may be a memory.

[0162] In addition, the apparatus provided in the embodiments of this application may specifically be a chip, component or module. The chip may include a connected processor and a memory. The memory is used to store instructions. When the processor calls and executes the instructions, the chip can execute a method for learning user preferences provided in the above embodiments.

[0163] This embodiment also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer, it causes the computer to execute the above-described method steps to implement the method for learning user preferences provided in the above embodiment.

[0164] The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, Digital Video Discs (DVDs), Compact Disc Read-Only Memory (CD-ROMs), microdrives, and magneto-optical disks, read-only memory (ROMs), random access memory (RAMs), erasable programmable read-only memory (EPROMs), electrically erasable programmable read-only memory (EEPROMs), dynamic random access memory (DRAMs), video random access memory (VRAMs), flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.

[0165] This embodiment also provides a computer program product that, when run on a computer, causes the computer to perform the aforementioned steps to achieve a method for learning user preferences provided in the above embodiment.

[0166] In this embodiment, the device, computer-readable storage medium, computer program product, or chip are all used to execute the corresponding methods provided above. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods provided above, and will not be repeated here.

[0167] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0168] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0169] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for learning user preferences, characterized in that, The method includes: Receives voice commands from users, which are used to control smart devices; Detect whether the voice command includes a preset field, the preset field being used to represent the user's preference information; When the preset field is included in the voice command, the user's preference information for the smart device is obtained based on the voice command.

2. The method according to claim 1, characterized in that, The step of obtaining the user's preference information for the smart device based on the voice command includes: Based on the voice command, a set of tuple data is determined, which includes: the user's identification information, the target control action, and the target control conditions required to execute the target control action. Based on the tuple data, the user's user preference information is obtained.

3. The method according to claim 2, characterized in that, The method further includes: Acquire target information, which includes the current user's voice commands and / or the current environment information; Determine whether the target information meets the target control conditions; When the target information meets the target control conditions, the target control action corresponding to the target control conditions is obtained based on the target information; Based on the target control action, a target control instruction is generated; and based on the target control instruction, the target device corresponding to the target control instruction is controlled to perform the target control action.

4. The method according to claim 3, characterized in that, The target control conditions include voice triggering conditions and / or environmental triggering conditions. Determining whether the target information meets the target control conditions includes: If the target information is a voice command, and the voice command meets the voice triggering condition, then the target information is determined to meet the target control condition. If the target information is the current environment information, and the environment information meets the environment triggering condition, then the target information is determined to meet the target control condition. If the target information includes both voice commands and current environment information, then when the voice command meets the voice triggering condition and the environment information meets the environment triggering condition, the target information is determined to meet the target control condition.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: If the preset field is not included in the voice command, the user's historical voice commands are obtained based on the user's voice commands, and the user intent corresponding to the historical voice commands is the same as the user intent corresponding to the voice commands. Based on the historical voice commands, obtain the historical control actions corresponding to the historical voice commands; Based on the historical control actions and the voice commands, the user's user preference information is obtained.

6. The method according to claim 5, characterized in that, The process of obtaining the user's preference information based on the historical control actions and the voice commands includes: Based on the execution frequency of various control actions in the historical control actions, a target historical control action is determined. The target historical control action refers to a historical control action whose execution frequency reaches a preset frequency. Based on the target historical control action, obtain the historical environment information when the target historical control action is executed; Based on the voice command, the historical environment information, and the target historical control actions, the user's user preference information is obtained.

7. The method according to any one of claims 1 to 4, characterized in that, The method further includes: The user's preference information for the smart device is stored in a preset preference memory.

8. The method according to claim 7, characterized in that, The method further includes: Real-time monitoring of new voice commands and corresponding control actions of the user on smart devices, and analysis of whether the new voice commands and control actions are consistent with the user preference information already stored in the preference memory; If there is a discrepancy, the user preference information in the preference memory is updated based on the newly added voice command and the corresponding control action; and the update time and update content are recorded to facilitate tracing the user's preference change trajectory.

9. The method according to any one of claims 1 to 4, characterized in that, The preset fields include preference keywords, preference degree descriptors, and target device association words. The preference keywords are used to characterize the user's tendency to control smart devices, the preference degree descriptors are used to characterize the user's preference intensity for control effect, and the target device association words are used to specify the specific smart device corresponding to the user's preference.

10. A device for learning user preferences, characterized in that, The device includes: An interaction module is used to receive voice commands from users, which are used to control smart devices. The recognition module is used to detect whether the voice command includes a preset field, the preset field being used to represent the user's preference information; The processing module is configured to, when the preset field is included in the voice command, obtain the user's user preference information for the smart device based on the voice command.

11. A smart device, characterized in that, The intelligent device includes: Memory, used to store executable program code; A processor for calling and running the executable program code from the memory, causing the smart device to perform the method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed, implements the method as described in any one of claims 1 to 9.

13. A computer program product, characterized in that, The computer program product includes: computer program code, which, when executed, implements the method as described in any one of claims 1 to 9.

Images