Large model-based home device control method, related device and program product

Abstract

The present disclosure provides a large model-based home device control method, related devices and program products, which relates to the fields of artificial intelligence technologies such as human-computer interaction, device control and large models. A specific embodiment of the method includes: reading a target device pointed by a control instruction from the received control instruction; generating a clarity of the control instruction based on a standard control instruction list associated with the target device, the standard control instruction list including standard control instructions for controlling the target device; in response to the clarity being less than a clarity threshold, generating an updated control instruction composed of the standard control instructions based on a semantic recognition result of the control instruction and the standard control instruction list by using a large model; and controlling the target device by using the updated control instruction. Thus, the user can control the home device more cost-effectively and flexibly while ensuring the stability and reliability of the device control process.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, specifically to the fields of human-computer interaction, device control, large-scale models and other artificial intelligence technologies, and particularly to home appliance control methods, devices, electronic devices, computer-readable storage media and computer program products based on large-scale models. Background Technology

[0002] With the continuous progress of society and the rapid development of information technology, people's requirements for the convenience, comfort and intelligence of their living environment are increasing day by day.

[0003] Against this backdrop, smart home control technology has emerged. This technology enables the networking and centralized management of various appliances, lighting devices, and security systems within a home and its rooms. Users can then manage these devices centrally and in batches via voice commands and terminal controls. Through this technology, users can manage and use home appliances more efficiently and conveniently, significantly improving the interactive experience and overall user experience.

[0004] In this context, how to manage and control equipment more broadly, and how to make it more convenient for users to control it while reducing their control costs, are issues that deserve attention and are urgently needed. Summary of the Invention

[0005] This disclosure presents a method, apparatus, electronic device, computer-readable storage medium, and computer program product for controlling home appliances based on a large model.

[0006] In a first aspect, embodiments of this disclosure propose a home appliance control method based on a large model, comprising: reading the target device pointed to by the control command from a received control command; generating a resolution of the control command based on a list of standard control commands associated with the target device, wherein the list of standard control commands includes standard control commands for controlling the target device; in response to a resolution less than a resolution threshold, generating an updated control command composed of standard control commands using the large model based on the semantic recognition result of the control command and the list of standard control commands; and controlling the target device using the updated control command.

[0007] Secondly, embodiments of this disclosure propose a home appliance control device based on a large model, comprising: a target device determination unit configured to read the target device pointed to by the control command from a received control command; a clarity generation unit configured to generate the clarity of the control command from a list of standard control commands associated with the target device, wherein the list of standard control commands includes standard control commands for controlling the target device; a command update unit configured to, in response to a clarity value being less than a clarity threshold, generate an updated control command composed of standard control commands using the semantic recognition result of the control command based on the large model and the list of standard control commands; and a first device control unit configured to control the target device using the updated control command.

[0008] Thirdly, embodiments of this disclosure provide an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to implement the large-model-based home device control method as described in any implementation of the first aspect.

[0009] Fourthly, embodiments of this disclosure provide a non-transitory computer-readable storage medium storing computer instructions that enable a computer, when executed, to implement a large-model-based home appliance control method as described in any implementation of the first aspect.

[0010] Fifthly, embodiments of this disclosure provide a computer program product including a computer program that, when executed by a processor, enables the implementation of a home appliance control method based on a large model as described in any of the implementations of the first aspect.

[0011] The home appliance control method, apparatus, electronic device, computer-readable storage medium, and computer program product based on a large model provided in this disclosure first reads the target device pointed to by the received control command; then, based on a list of standard control commands associated with the target device, a resolution of the control command is generated, wherein the list of standard control commands includes standard control commands for controlling the target device; next, in response to the resolution being less than a resolution threshold, an updated control command composed of standard control commands is generated using the semantic recognition result of the control command based on the large model and the list of standard control commands; finally, the updated control command is used to control the target device.

[0012] This disclosure enables users to control home appliances more cost-effectively and flexibly, while ensuring the stability and reliability of the appliance control process.

[0013] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0014] Other features, objects, and advantages of this disclosure will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is an exemplary system architecture to which this disclosure can be applied; Figure 2 A flowchart illustrating a home appliance control process based on a large model, provided for embodiments of this disclosure; Figure 3 A flowchart illustrating a process for generating the clarity of control instructions, provided for embodiments of this disclosure; Figure 4 This is a flowchart illustrating the home appliance control process based on a large model in a specific application scenario, as provided in an embodiment of this disclosure. Figure 5 A structural block diagram of a home appliance control device based on a large model provided in this disclosure embodiment; Figure 6 This is a schematic diagram of the structure of an electronic device suitable for implementing a large-model-based home device control method, provided as an embodiment of the present disclosure. Detailed Implementation

[0015] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding; these should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description. It should be noted that, unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0016] Furthermore, if the technical solutions disclosed herein involve or require the use of information that is of the nature of user personal information, such as the standard control instruction list of the devices involved in this disclosure, the candidate device list in the user's environment, etc., the process of acquiring, storing, using, processing, transporting, providing and disclosing such information shall comply with the provisions of relevant laws and regulations and shall not violate public order and good morals.

[0017] Figure 1An exemplary system architecture 100 is shown that can be applied to embodiments of the large-model-based home appliance control methods, apparatuses, electronic devices, and computer-readable storage media disclosed herein.

[0018] like Figure 1 As shown, system architecture 100 may include a terminal device 101 capable of receiving control commands (e.g., voice commands) sent by a user (not shown) and correspondingly controlling other terminal devices and home appliances based on these control commands, as well as home appliances (terminal devices) 102, 103, and 104 that can be controlled by the terminal device 101. The terminal device 101 and home appliances 102, 103, and 104 can be connected via a network 105 to communicate using the communication link formed and provided by the network 105, for example, communicating the aforementioned control commands. The network 105 may include various connection types, such as wired or wireless communication links or fiber optic cables, etc.

[0019] Accordingly, users can use terminal device 101 to control home devices 102, 103, and 104 via network 105, for example, by turning them on and off, and adjusting their operating modes and status. Various applications can be installed between terminal device 101 and home devices 102, 103, and 104 to enable information communication between them, such as smart home control applications and instant messaging applications.

[0020] Terminal device 101, as well as home appliances 102, 103, and 104, can all be hardware or software. For example, when terminal device 101 is hardware, it can be, for example, "smart glasses" with control command receiving function (e.g., voice command receiving function); similarly, when home appliances 102, 103, and 104 are hardware, they can be specific home appliances such as electric fans, air conditioners, table lamps, refrigerators, televisions, etc.; and when terminal device 101, as well as home appliances 102, 103, and 104 are software, they can be installed in the aforementioned home and electronic devices, and can be implemented as multiple software programs or software modules, or as a single software program or software module, without specific limitations here.

[0021] It should be understood that in some scenarios and embodiments, for the purpose of improving the processing capabilities of terminal device 101, terminal device 101 may actually be a "network device" composed of specific terminal devices and servers. For example, such a "network device" may be a terminal device receiving "natural language information" sent by the user and using a server to parse it. In such a case, the above architecture 100 may also include a "server" as an example. In some embodiments, if a server is included, when the server is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or it can be implemented as a single server; when the server is software, it can be implemented as multiple software programs or software modules, or it can be implemented as a single software program or software module, without specific limitations here.

[0022] Terminal device 101 can provide various services through built-in applications. For example, a smart home control application can control connected and managed home devices 102, 103, and 104 based on user control commands. When running this application, terminal device 101 can achieve the following: First, it receives control commands sent by the user; then, it reads the target device from the received commands; next, based on the standard control command list associated with the target device, it generates a resolution of the control commands, where the standard control command list includes standard control commands for controlling the target device; then, in response to a resolution lower than a resolution threshold, it uses a large model based on the semantic recognition results of the control commands and the standard control command list to generate an updated control command composed of standard control commands; finally, it uses the updated control command to control the target device (e.g., at least one of home devices 102, 103, and 104).

[0023] Typically, if the above process is completed solely by the terminal device 101, the home appliance control device based on the large model is also usually located in the terminal device 101, which will not be discussed further here. However, as discussed above, if a "server" is required in some embodiments or scenarios, the home appliance control device based on the large model can be deployed separately in the server, or jointly in the terminal device 101 and the server.

[0024] It should be understood that Figure 1 The number of terminal devices and networks shown is merely illustrative. Depending on implementation needs, any number of terminal devices and networks can be included.

[0025] Please refer to Figure 2 , Figure 2 A flowchart of a home appliance control process based on a large model is provided for embodiments of this disclosure, including process 200.

[0026] Process 200 specifically includes the following steps: Step 201: Read the target device pointed to by the control command from the received control command; In embodiments of this disclosure, this step is intended to be performed by the execution entity of a large-model-based home device control method (e.g., Figure 1 After receiving control instructions (e.g., control instructions in natural language form) provided by the user, the terminal device 101 shown reads the target device pointed to by the control instructions from the received control instructions.

[0027] Typically, the executing entity can choose to determine the target device to be controlled by analyzing the entities, or device entities, that are hit by the control command after receiving the control command.

[0028] For example, if the control command is to turn on the living room TV or the living room air conditioner, the executing entity can determine and read which air conditioner or TV in the living room is the target device that the control command intends to control by referring to "TV" (or "living room TV") and "Air conditioner" (or "living room air conditioner").

[0029] In some embodiments, to reduce the cost of providing user descriptions and control commands, the executing entity may choose to supplement entity parsing with semantic recognition to determine the target device the user wishes to control. This balances computational and identification resource consumption while ensuring that even when the user provides control commands using vague language with standard names (e.g., television, air conditioner), the executing entity can still read the desired target device.

[0030] In this scenario, after identifying the target device by splitting and reading entities, the executing entity can determine whether there is a lack of context in the control instructions, excluding those associated with the entities mentioned above. For example, the executing entity can detect whether there are any parts of the control instructions that are not associated with the target device identified through entity analysis. If so, the executing entity can "consider" that this part of the control instructions is not associated with the corresponding target device and that there is a "lack of context" (i.e., such a "lack of context" can be understood as a lack of "device" or "entity," for example, a lack of an object to perform the control action).

[0031] In practice, the executing entity can invoke pre-trained models such as semantic processing models and syntax structure recognition models to identify whether there is a lack of context.

[0032] If such a device exists, the executing entity can respond by identifying a target candidate device from the candidate device list that corresponds to the missing context. For example, the executing entity can identify a candidate device that exists and can perform the functions and purposes corresponding to the missing context control instructions based on the functions and purposes that the candidate devices in the candidate device list can perform (e.g., by semantic comparison to identify a candidate device that can provide the functions of the missing context control instructions), and use this target candidate device as a supplement to the controlled object.

[0033] In practice, the candidate devices in this candidate device list can be controllable home appliances located in the home or environment, identified by the executing entity through handshake communication after being connected by the user or authorized by the user. Therefore, this candidate device list not only supplements the "target devices" to achieve higher-quality control, but also constrains the executing entity's process of supplementing target devices, preventing errors caused by adding or identifying non-existent or uncontrollable "devices".

[0034] It should be noted that the candidate list devices can be obtained directly from local storage devices by the aforementioned executing entity, or from non-local storage devices. Local storage devices can be data storage modules located within the aforementioned executing entity, such as the local hard drive of a terminal device. In this case, the candidate list devices can be quickly accessed locally. Non-local storage devices can also be any other electronic devices configured to store data, such as user terminals. In this case, the aforementioned executing entity can obtain the required candidate list devices by sending an acquisition command to the electronic device.

[0035] Accordingly, if the target candidate device can be identified, the executing entity can use the target candidate device as the aforementioned target device to more fully and comprehensively determine the target device to be controlled.

[0036] In some embodiments, to improve control quality and avoid erroneously targeting devices that the user may not need or expect to control, the executing entity can actually provide target candidate devices to the issuer of the control command (i.e., the user mentioned above) before selecting them as target devices. For example, the executing entity can inform the user through voice announcements, sending prompt messages, etc., that based on the user's control command, in addition to the target device directly specified by the user, it believes that the user may need or expects to control these target candidate devices, and inquire whether this is the user's true thought or intention.

[0037] Accordingly, if a user deems it necessary and possible to control these target candidate devices simultaneously, they can provide feedback to the executing entity regarding the specific target candidate devices they wish to control (e.g., through natural language feedback). Upon receiving confirmation of the target candidate device from the sender, the executing entity can respond by designating the user-selected target candidate device as the target device. This approach, through user interaction, clearly and accurately provides supplementary services for target devices, avoiding erroneous control of devices the user does not wish to control, and improving control quality.

[0038] Step 202: Based on the list of standard control commands associated with the target device, generate the clarity of the control commands; In embodiments of this disclosure, after determining the target devices based on step 201 above, the executing entity can read the standard control instruction list associated with these target devices and generate the clarity of the control instructions based on the standard control instruction list. The standard control instruction list includes standard control instructions for controlling the target devices.

[0039] The standard control instruction list can record "standard control instructions" used to control the corresponding target device. For example, those instructions configured at the factory that can control the target device. For example, for a device like an "air conditioner", its standard control instructions could be, for instance, turn on the air conditioner, adjust the temperature to XX degrees, adjust the fan speed to Y level, and set the operating mode to K mode, etc.

[0040] Similarly, the list of standard control instructions can be similar to the list of candidate devices discussed above, provided by devices local to the executing entity or outside the executing entity, which will not be repeated here.

[0041] In some embodiments, the standard control instructions in the standard control instruction list can be either in the form of a programming language that can be directly controlled at the code level (e.g., Java, Python, etc.) or in the form of a standard language (one or both). This allows the standard control instruction list to record instructions that can be understood and used by the target device in a differentiated and comprehensive manner, based on the specific circumstances of the target device. This enables the executing entity to more accurately determine whether a control instruction is usable, thus avoiding wasting the rewriting computational resources discussed later.

[0042] Accordingly, in this step, the executing entity can use the standard control instruction list to generate the "clarity" of the aforementioned control instructions. For example, the executing entity can use these standard control instruction lists as a corpus to determine the "clarity" of the control instructions.

[0043] For example, the executing entity can use a list of standard control instructions as a corpus to determine whether each part of the control instructions can correspond to a standard control instruction in the list. Accordingly, for ease of understanding, the part of the control instructions that corresponds to a standard control instruction in the list can be called the first control sub-instruction, while the part that does not correspond can be called the second control sub-instruction.

[0044] In some embodiments, the executing entity may first break down the control instructions into multiple control sub-instructions (e.g., a combination of an action and a target device) using a “sentence structure” approach, and then determine whether they can be matched by the standard control instruction list to classify them as first control sub-instructions or second control sub-instructions.

[0045] Then, the executing entity can use the proportion of the first control sub-instruction to the control instruction (e.g., the proportion of the number of words, or the proportion of the total number of control sub-instructions involved) as the clarity of the aforementioned control instruction.

[0046] The executing entity can then compare this clarity with a clarity threshold. In practice, this clarity threshold can be set based on the assumption that the control commands can be well understood by the target device and have direct execution value.

[0047] Accordingly, if the resolution in this step is greater than or equal to the resolution threshold, the executing entity can choose to directly control the target device using control commands. This ensures that clear and compliant control commands can be directly used to control the target device, guaranteeing control efficiency.

[0048] In some embodiments, while the overall clarity of the control instructions may be satisfactory, they may still include second control sub-instructions. In such cases, considering that most of the content in the control instructions is "clear," indicating that the user may have sufficient experience in sending control instructions, the executing entity may choose to provide it to the issuer of the control instructions (e.g., the user) so that the user understands which control instructions have not been executed and decides whether adjustments are needed, rather than having the executing entity directly and proactively rewrite it as described below. This avoids unintended rewriting due to identification errors (e.g., identifying content in the control instructions that is not intended for control as a second control sub-instruction), which could lead to wasted resources and erroneous actions.

[0049] Correspondingly, if the sender receives the second control sub-instruction and provides the corresponding modification result, the executing entity can similarly use it as a new control instruction to execute process 200, which will not be repeated here.

[0050] If the clarity is less than the clarity threshold, it can be understood that its current execution value is low (for example, there are few actions that can be performed after execution, and it is difficult to fully achieve the user's expected purpose). In this case, the executing entity can respond by choosing to continue executing step 203.

[0051] Step 203: In response to a resolution less than the resolution threshold, use the semantic recognition results of the control instructions based on the large model and the standard control instruction list to generate an updated control instruction composed of the standard control instructions; In embodiments of this disclosure, as discussed above, if the sharpness determined in step 202 is less than the sharpness threshold, the executing entity may respond by using the semantic recognition results of the control instructions based on the large model and the standard control instruction list to generate an updated control instruction composed of standard control instructions.

[0052] For example, the executing entity can construct prompt words based on a pre-configured corpus template to indicate the semantic recognition results of the large model based on control instructions and the standard control instruction list. For example, a prompt word in the form of "based on the semantics of 'control instructions', using the standard control instructions in the standard control instruction list as corpus, 'rewrite the control instructions' to generate an updated control instruction composed of standard control instructions" could be used.

[0053] Then, the executing entity can invoke the large model and use the prompt word to instruct the large model to generate updated control instructions composed of standard control instructions based on the semantic recognition results of the control instructions and the standard control instruction list, so as to realize the rewriting action of control instructions in the form of standard control instructions.

[0054] Large models refer to a class of artificial intelligence models with a large number of parameters, constructed from artificial neural networks. Examples include generative models and Large Language Models (LLMs). Taking LLM as an example, it is an artificial intelligence model designed to understand and generate human language. Based on its understanding, LLM can perform corresponding processing operations to obtain the desired results. For instance, after obtaining control instructions and a list of standard control instructions, it can generate updated control instructions composed of standard control instructions based on the semantic recognition results of the control instructions and the list of standard control instructions.

[0055] LLMs can be trained on large amounts of text data and perform a wide range of tasks, including text summarization, translation, sentiment analysis, and more. A key characteristic of LLMs is their large scale; they typically include a large number of parameters to help them learn complex patterns in language data. These models are often based on deep learning architectures, such as transformers, which contributes to their superior performance across various natural language tasks.

[0056] Furthermore, for large models, the aforementioned "guide words" can be omitted through default configuration. This default configuration allows large models to stably and purposefully generate update control instructions composed of standard control instructions based on the semantic recognition results of control instructions and the standard control instruction list. This enables large models to execute the rewriting process of generating update control instructions composed of standard control instructions more efficiently and with higher quality.

[0057] In some embodiments, during the process of generating and rewriting control instructions, the rewriting capability of the large model can be enhanced by adding constraint information and reference information, enabling more efficient and comprehensive rewriting. For example, disabling conditions for certain target devices can be added, or scene-based standard control instruction selection references can be added. For instance, in a "night" scene, the standard control instruction to brighten the lights can be configured with reference information for the "brightness" level. Furthermore, corresponding order constraints can be added between standard control instructions to ensure they are executed in the correct order.

[0058] In practice, reference information can also be configured based on user-provided preference reference information. For example, specific reference information may include the user's control habits for various target devices in specific weather, room, time, and behavioral scenarios, as well as their preferred operating states of the target devices. This allows the large model to better understand the user's intentions and preferences when rewriting and determining control strategies. For instance, when a user provides a control command like "It's raining, the room is a bit damp, please dehumidify and cool down," the large model can use this reference information to determine the target devices the user needs to control and use to dehumidify and cool down their room on a rainy day, and the operating states these target devices need to be set to (e.g., if the air conditioner is set to dehumidification mode, its corresponding set temperature should be Z degrees; if the dehumidifier is turned on, its corresponding fan speed should be Q, etc.).

[0059] Step 204: Control the target device using update control commands.

[0060] In the embodiments of this disclosure, the control command is rewritten based on the above step 203 to generate an updated control command. The executing entity can send the corresponding updated control command (partial) to each target device to use the control command to completely and holistically control each target device involved.

[0061] The home appliance control method based on a large model provided in this disclosure first reads the target device indicated by the received control command; then, based on a list of standard control commands associated with the target device, a resolution of the control command is generated, wherein the list of standard control commands includes standard control commands used to control the target device; next, in response to a resolution less than a resolution threshold, an updated control command composed of standard control commands is generated using the semantic recognition result of the control command based on the large model and the list of standard control commands; finally, the updated control command is used to control the target device. This disclosure enables users to control home appliances more cost-effectively and flexibly while ensuring the stability and reliability of the device control process.

[0062] In some embodiments, during the step of classifying control sub-instructions to determine their clarity, the criterion for determining the clarity of a control sub-instruction includes not only judging whether its corresponding "action" conforms to or matches the actions in the standard control instruction list, but also considering whether the amplitude and limitation range of its action are appropriate. This helps to avoid more accurately identifying control sub-instructions whose actions meet the requirements but whose actual or target parameters cannot be achieved, thus improving the quality of clarity determination.

[0063] Correspondingly, when the standard control instructions in the standard control instruction list involve multiple amplitudes, gears, and limit ranges, these contents can also be recorded accordingly, so that the standard control instruction list can be used later to more accurately determine whether the acquired control instructions can be "understood" and "executed" by the target device.

[0064] For ease of understanding, the following will be referenced together. Figure 3 Let's have a discussion. Figure 3 A flowchart of a process for generating the clarity of control instructions provided in an embodiment of this disclosure includes process 300, for example, process 300 may be an alternative or alternative implementation of step 202 described above.

[0065] Process 300 specifically includes the following steps: Step 301: Determine the first control sub-instruction among the control instructions that can match the standard control instructions in the standard control instruction list; Specifically, as discussed above, the executing entity can first classify and determine the control sub-instructions that can be matched by the standard control instruction list at the "action" level as the first control sub-instruction.

[0066] Step 302: In response to the fact that the amplitude parameter associated with the first control sub-instruction is within the limit range of the standard control instruction corresponding to the first control sub-instruction, the first control sub-instruction is determined to be a clear control sub-instruction; Specifically, based on step 301 above, the executing entity can read the current (operating) parameters of the target device and then determine whether these parameters, after adjustment by the first control sub-instruction, still fall within the limitation range of the standard control instruction corresponding to the first control sub-instruction. For example, if the current set temperature is already at its lowest, the action of lowering the temperature in the first control sub-instruction will no longer fall within the limitation range of the standard control instruction. Another example is a user's control instruction that specifies a temperature, brightness, etc., that exceeds the minimum (or maximum) temperature, minimum brightness, etc., limit range corresponding to the standard control instruction.

[0067] If there is no situation exceeding the aforementioned limits, the executing entity may choose to actually use it to determine sharpness. For ease of distinction, such a first control sub-instruction actually used to determine sharpness can be described as a sharpness control sub-instruction.

[0068] Accordingly, the executing entity can continue to execute step 303 as discussed above, replacing the first control sub-instruction with the "clear control sub-instruction" to actually generate the clarity of the control instruction based on the proportion of the clear control sub-instruction to the control instruction. This will not be repeated here.

[0069] Step 303: Based on the proportion of clear control sub-instructions to control instructions, generate the clarity of control instructions.

[0070] Accordingly, in such an embodiment, if subsequent rewriting of control commands or generation of updated control commands is involved, during rewriting, the large model can, based on pre-defined constraints (e.g., based on adjustment tendencies, selecting usable parameters, limit values ​​in gears, and deleting the first control sub-command if none exist), correspondingly select to delete the unreachable first control sub-command (e.g., in lowering), or adjust the parameters it points to, so as to meet the user's adjustment needs to the greatest extent possible in the direction specified by the user, and when adjustment is objectively impossible, the first control sub-command will not be rewritten or issued, thereby saving rewriting resources and avoiding control errors.

[0071] In some embodiments, during the rewriting of control instructions, in addition to inputting the "control instructions" comprehensively into the large model as in process 200 above, so as to rewrite them as a whole using the large model and take into account the integrity of each part, considering that the first control sub-instruction in the control instructions is actually clear for the corresponding target device, the executing entity may also choose to rewrite only the second control sub-instruction in different embodiments due to different strategies and scenario requirements, so as to improve rewriting efficiency and save computing resources.

[0072] Accordingly, in such a case, when the executing entity generates an updated control instruction consisting of standard control instructions from the standard control instruction list based on the semantic recognition result of the control instructions and the standard control instruction list using the large model, in response to the resolution being less than the resolution threshold, as an alternative, it can choose to generate an updated control sub-instruction consisting of standard control instructions from the standard control instruction list based on the semantic recognition result of the second control sub-instruction, which is different from the first control sub-instruction, and the standard control instruction list using the large model. This process is essentially equivalent to the case described above where control instructions are used as input, and will not be repeated here.

[0073] Then, by combining the first control sub-instruction and the update control sub-instruction, the update control instruction is obtained, so as to achieve the purpose of obtaining the update control instruction by simply rewriting the second control sub-instruction.

[0074] Based on any of the above embodiments, if the executing entity rewrites the control instruction to obtain an updated control instruction, the executing entity may choose to establish a mapping relationship between the control instruction and the updated control instruction. This allows the executing entity to directly update the control instruction that has already been used when it receives the control instruction or a control instruction with a high degree of similarity to the control instruction (e.g., the semantic similarity between the two is greater than or equal to a predetermined similarity threshold), thus avoiding repeated rewriting. Alternatively, the executing entity may provide optimization suggestions or directions based on such a mapping relationship (e.g., providing suggestions for improving user expression or providing a reference for the configuration of standard control instructions).

[0075] Accordingly, in such cases, as an alternative or alternative, when the executing entity is reading the target device pointed to by the control command from the received control command, it may further choose to check whether there is a pre-established mapping relationship corresponding to the control command after receiving the control command (for example, a mapping relationship directly established with the control command, or a mapping relationship established with a similar control command whose similarity meets the above-mentioned similarity threshold requirement).

[0076] Then, if no pre-established mapping relationship exists between the control command and the target device, the executing entity, upon receiving the control command and finding no pre-established mapping relationship, selects to read the target device pointed to by the control command from the received control command. In other words, if such a mapping relationship exists, the executing entity can directly search for and reuse the previously generated update control command based on that mapping relationship.

[0077] In this way, users can issue control commands using their preferred expressions, and the executing entity does not need to frequently rewrite these commands so that the target device can accurately understand them. This not only saves resources required for command conversion but also improves overall control efficiency.

[0078] Based on any of the above embodiments, after the executing entity actually controls the target device using the update control command or control command, it can also provide the target device and the control result of the target device to the issuer of the control command (i.e., the user), so that the user can understand the controlled target device and the specific control content of it in a timely manner, so that the user can dynamically and timely adjust the control actions that do not meet the expectations according to actual needs, thereby improving the user's interactive experience.

[0079] To enhance understanding, this disclosure also presents a specific application scenario, outlining the control process for home appliances based on a large model implemented within that scenario. Please refer to this for further clarification. Figure 4 . Figure 4 This is a flowchart illustrating a home appliance control process based on a large model in a specific application scenario, as provided in an embodiment of the present disclosure, including process 400.

[0080] For ease of discussion, please refer to the following: Figure 1 The architecture shown in Figure 100 will be used for explanation.

[0081] For example, in process 400, a user 410 may be further included, who can send a control command 415 to the terminal device 101 by executing S401. For instance, the control command 415 may be a voice command specifically stating "Turn off the desk lamp and purify the indoor air".

[0082] Upon receiving the control command 415, the terminal device 101 can execute S402 to read the target device pointed to by the control command 415. For example, the executing entity targets the home appliance 103, executed as a "table lamp," based on the "entity" directly indicated by the control command 415. Then, it uses the candidate device list to target the home appliance 102, executed as an "electric fan," which has functions such as air supply and ventilation, and the home appliance 104, executed as an "air conditioner."

[0083] Then, terminal device 101 can continue to execute S403 to generate the clarity 430 of control instruction 415 based on the standard control instruction lists associated with home devices 102, 103, and 104 (e.g., standard control instruction list 421, standard control instruction list 422, and standard control instruction list 423).

[0084] For example, if the clarity 430 is less than the clarity threshold due to the "blur" present in "purifying indoor air", then the terminal device 101 can respond to this and continue to execute S404 to generate an updated control instruction 450 composed of standard control instructions by using the large model 440 based on the semantic recognition result of the control instruction 415 and the standard control instruction list 421, standard control instruction list 422, and standard control instruction list 423.

[0085] For example, after the large model 440 is "updated and adjusted", the ambiguous part of the control instruction 415, "purify indoor air", in the updated control instruction 450 can be adjusted to specify it as the "standard control instruction" in the standard control instruction list 422 and standard control instruction list 423, such as turning on home appliance 102 and turning on home appliance 104.

[0086] In addition, as discussed above, based on the purpose of "purifying indoor air", the large model 440 can further clarify the specific functions and working status of the home appliance 104 (e.g., "air supply" working mode, rather than "cooling" or "heating"), and rewrite them to update the control command 450.

[0087] Finally, the terminal device 101 can execute S405 to control the home appliances 102, 103 and 104 accordingly using the updated control command 450. For example, the corresponding part of the control command can be provided to the home appliances 102, 103 and 104 to "turn on the electric fan (home appliance 102)", "turn off the table lamp and electric fan (home appliance 103)" and "turn on the air conditioner (home appliance 104) and set it to the 'air supply' working mode".

[0088] Further reference Figure 5 As an implementation of the methods shown in the above figures, this disclosure provides an embodiment of a home appliance control device based on a large model. This device embodiment is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0089] like Figure 5 As shown, the home appliance control device 500 based on a large model in this embodiment may include: a target device determination unit 501, a clarity generation unit 502, an instruction update unit 503, and a first device control unit 504. The target device determination unit 501 is configured to read the target device indicated by the received control instruction; the clarity generation unit 502 is configured to generate the clarity of the control instruction from a list of standard control instructions associated with the target device, wherein the list of standard control instructions includes standard control instructions for controlling the target device; the instruction update unit 503 is configured to, in response to a clarity value less than a clarity threshold, generate an updated control instruction composed of standard control instructions using the semantic recognition result of the control instruction based on the large model and the list of standard control instructions; and the first device control unit 504 is configured to control the target device using the updated control instruction.

[0090] In this embodiment, the specific processing of the target device determination unit 501, the clarity generation unit 502, the instruction update unit 503, and the first device control unit 504 in the large-model-based home appliance control device 500, and the resulting technical effects, can be found in the following references: Figure 2 The relevant descriptions of steps 201-204 in the corresponding embodiments will not be repeated here.

[0091] In some optional implementations of this embodiment, the target device determination unit 501 is further configured to read the target device from the received control command by means of device entity parsing; the candidate device determination subunit is configured to determine the target candidate device with the missing context from the candidate device list in response to the presence of a missing context for the controlled device in the control command; and the target device supplementation subunit is configured to use the target candidate device as the target device.

[0092] In some optional implementations of this embodiment, the target device supplement unit includes: a candidate device providing subunit, configured to provide a target candidate device to the issuer of the control command; and a target device supplement subunit, configured to use the target candidate device as the target device in response to the issuer's feedback of a confirmation command for the target candidate device.

[0093] In some optional implementations of this embodiment, the sharpness generation unit 502 includes: a control instruction filtering subunit configured to determine a first control sub-instruction among the control instructions that can hit a standard control instruction in the standard control instruction list; a control instruction classification subunit configured to determine the first control sub-instruction as a sharp control sub-instruction in response to the amplitude parameter associated with the first control sub-instruction being within the limitation range of the standard control instruction corresponding to the first control sub-instruction; and a sharpness generation subunit configured to generate the sharpness of the control instruction based on the proportion of the sharp control sub-instruction to the control instruction.

[0094] In some optional implementations of this embodiment, the instruction update unit 503 includes: an update instruction generation subunit, configured to, in response to a resolution less than a resolution threshold, use a large model based on the semantic recognition result of a second control sub-instruction that is different from the first control sub-instruction in the control instructions and a standard control instruction list, to generate an update control sub-instruction composed of standard control instructions in the standard control instruction list; and a control instruction combination subunit, configured to combine the first control sub-instruction and the update control sub-instruction to obtain an update control instruction.

[0095] In some optional implementations of this embodiment, the standard control instruction list includes standard control instructions in programming language form and / or standard language form.

[0096] In some optional implementations of this embodiment, the apparatus 500 further includes a mapping relationship establishment unit, configured to establish a mapping relationship between control instructions and update control instructions.

[0097] In some optional implementations of this embodiment, the target device determination unit 501 is further configured to read the target device pointed to by the control command from the received control command in response to receiving a control command and there is no pre-established mapping relationship corresponding to the control command.

[0098] In some optional implementations of this embodiment, the apparatus 500 further includes: a second device control unit configured to control the target device using control commands in response to a resolution greater than or equal to a resolution threshold.

[0099] In some optional implementations of this embodiment, the apparatus 500 further includes: a control result providing unit, configured to provide a target device and a control result for the target device to the issuer of the control command, wherein the control result is determined based on an update control command or a control command.

[0100] This embodiment exists as a device embodiment in response to the above method embodiment. The home appliance control device based on a large model provided in this embodiment enables users to control home appliances more cost-effectively and flexibly, while ensuring the stability and reliability of the device control process.

[0101] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0102] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0103] like Figure 6 As shown, device 600 includes a computing unit 601, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 602 or a computer program loaded into random access memory (RAM) 603 from storage unit 608. RAM 603 may also store various programs and data required for the operation of device 600. The computing unit 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0104] Multiple components in device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of monitors, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0105] The computing unit 601 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as a large-model-based home appliance control method. For example, in some embodiments, the large-model-based home appliance control method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program can be loaded and / or installed on device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the large-model-based home appliance control method described above can be performed. Alternatively, in other embodiments, the computing unit 601 may be configured, by any other suitable means (e.g., by means of firmware), to perform a home appliance control method based on a large model.

[0106] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0107] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0108] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0109] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0110] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0111] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is established by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, also known as cloud computing servers or cloud hosts, which are hosting products within the cloud computing service ecosystem to address the management difficulties and weak business scalability inherent in traditional physical hosts and Virtual Private Servers (VPS) services. Servers can also be categorized as distributed system servers or servers incorporating blockchain technology.

[0112] According to the technical solutions of the embodiments of this disclosure, users can control home appliances more cost-effectively and flexibly while ensuring the stability and reliability of the appliance control process.

[0113] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution provided in this disclosure can be achieved, and this is not limited herein.

[0114] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A home appliance control method based on a large model, comprising: Read the target device indicated by the received control command; Based on the list of standard control instructions associated with the target device, the clarity of the control instructions is generated, wherein the list of standard control instructions includes standard control instructions for controlling the target device; In response to the resolution being less than a resolution threshold, an updated control instruction consisting of the standard control instructions is generated using a large model based on the semantic recognition results of the control instructions and the standard control instruction list; The target device is controlled using the update control command.

2. The method according to claim 1, wherein, The step of reading the target device pointed to by the received control command includes: The target device is read from the received control commands by parsing the device entity. In response to a lack of context for the controlled device in the control command, a target candidate device with the missing context is determined from the candidate device list; The target candidate device is selected as the target device.

3. The method according to claim 2, wherein, The step of selecting the target candidate device as the target device includes: The target candidate device is provided to the issuer of the control command; In response to the confirmation instruction from the sender regarding the target candidate device, the target candidate device is designated as the target device.

4. The method according to claim 1, wherein, The process of generating the clarity of the control commands based on the standard control command list associated with the target device includes: Determine the first control sub-instruction among the control instructions that can match the standard control instructions in the standard control instruction list; In response to the fact that the amplitude parameter associated with the first control sub-instruction is within the limit range of the standard control instruction corresponding to the first control sub-instruction, the first control sub-instruction is determined to be a clear control sub-instruction; The clarity of the control command is generated based on the proportion of the clarity control sub-instruction to the control command.

5. The method according to claim 4, wherein, In response to the sharpness being less than a sharpness threshold, the system utilizes a large model based on the semantic recognition results of the control instructions and the standard control instruction list to generate an updated control instruction composed of standard control instructions from the standard control instruction list, including: In response to the resolution being less than a resolution threshold, an updated control sub-instruction is generated by using a large model based on the semantic recognition result of a second control sub-instruction that is different from the first control sub-instruction in the control instruction and the standard control instruction list. This update control sub-instruction is composed of standard control instructions from the standard control instruction list. The first control sub-instruction and the update control sub-instruction are combined to obtain the update control instruction.

6. The method according to claim 1, wherein, The standard control instruction list includes the standard control instructions in programming language form and / or standard language form.

7. The method according to claim 1, further comprising: Establish a mapping relationship between the control command and the update control command.

8. The method according to claim 7, wherein, The step of reading the target device pointed to by the received control command includes: In response to receiving a control command and finding that there is no pre-established mapping relationship corresponding to the control command, the target device pointed to by the control command is read from the received control command.

9. The method according to claim 1, further comprising: In response to the resolution being greater than or equal to the resolution threshold, the target device is controlled using the control command.

10. The method according to any one of claims 1-9, further comprising: The target device and the control result of the target device are provided to the issuer of the control command, wherein the control result is determined based on the update control command or the control command.

11. A home appliance control device based on a large model, comprising: The target device determination unit is configured to read the target device pointed to by the received control command; A resolution generation unit is configured to generate the resolution of the control instructions from a list of standard control instructions associated with the target device, wherein the list of standard control instructions includes standard control instructions for controlling the target device. The instruction update unit is configured to, in response to the resolution being less than a resolution threshold, generate an update control instruction composed of the standard control instructions using a large model based on the semantic recognition result of the control instructions and the standard control instruction list; A first device control unit is configured to control the target device using the update control command.

12. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed, enable the at least one processor to perform the home appliance control method based on a large model as described in any one of claims 1-10.

13. A non-transitory computer-readable storage medium storing computer instructions for causing the computer to perform the home appliance control method based on a large model according to any one of claims 1-10.

14. A computer program product comprising a computer program that, when executed by a processor, implements the home appliance control method based on a large model according to any one of claims 1-10.

Images