Control Method, System, Device and Electronic Device for Smart Devices

Abstract

The present invention discloses a control method, system, device and electronic device for intelligent devices. Among them, the control method includes: receiving information data of a target user, respectively parsing the information data through the first type of models in the model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship between the first type of models, and processing the tag data through the second type of models in the model set to generate reply tag data and / or instruction tag data. The present invention solves the technical problem in the related art that in a serial semantic parsing model, it is easy to cause parsing errors of other associated models due to the parsing error of a certain model, resulting in interaction failure and reducing the user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent devices, and in particular, to a control method, system, device and electronic device for intelligent devices. Background Art

[0002] With the rapid development of intelligent device technology, more and more users begin to use intelligent devices for more convenient and efficient life and work. How to improve the intelligence of the dialogue system for controlling intelligent devices has become an important issue for enhancing the user experience. In related technologies, the existing dialogue system usually consists of three parts: a semantic parsing model, a dialogue management model, and a natural language generation model. The dialogue management model will generate corresponding reply messages and issue device control instructions based on the parsing results of the semantic parsing model and the running data of the current intelligent device.

[0003] As Figure 1 shown, the existing parsing model is usually serially composed of multiple semantic parsing models (i.e., semantic parsing models 1, 2... N). The multiple semantic parsing models have an upper and lower serial association relationship. In the dialogue system, through the parsing results of the serial semantic parsing model, conversations are carried out with users, and instructions are issued to intelligent devices. However, in the existing cascade serial semantic parsing model architecture, inevitable cascade errors will occur. That is, when a certain parsing model makes a prediction error, it will cause the associated next-level model to perform the parsing of the current model based on the above incorrect parsing result, easily resulting in incorrect parsing results, ultimately leading to the failure of device control and language interaction, and reducing the user experience.

[0004] In view of the above problems, no effective solution has been proposed yet. Summary of the Invention

[0005] Embodiments of the present invention provide a control method, system, device and electronic device for intelligent devices, so as to at least solve the technical problem in related technologies that the serial semantic parsing model is prone to cause incorrect parsing of other associated models due to the parsing error of a certain model, resulting in interaction failure and reducing the user experience.

[0006] According to one aspect of an embodiment of the present invention, a control method for a smart device is provided, including: receiving information data of a target user, where the information data includes voice information data and / or text information data; respectively parsing the information data through a first type of model in a model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship among the first type of models; processing the tag data through a second type of model in the model set to generate reply tag data and / or instruction tag data, where the reply tag data is used to reply interaction information to the target user, the instruction tag data is used to send an operation instruction to the smart device, and the device control parameter is carried in the instruction tag data.

[0007] Optionally, before receiving the information data of the target user, the control method further includes: inputting historical information data and original tag data, where the historical information data includes historical voice information data and / or historical text information data, and the original tag data is data obtained by parsing the historical information data and processing the parsed information data during a historical process; training the original tag data based on a third type of model in the model set to obtain trained tag data.

[0008] Optionally, after respectively parsing the information data through the first type of model in the model set to obtain tag data, the control method further includes: obtaining historical operation data of the smart device; updating the parsed tag data based on the historical operation data through a third type of model in the model set to obtain updated tag data.

[0009] Optionally, after respectively parsing the information data through the first type of model in the model set to obtain tag data, the control method further includes: if there is tag data with a repetition rate greater than a preset threshold in the updated tag data, retaining the tag data and deleting tag data with a repetition rate less than or equal to the preset threshold.

[0010] Optionally, the step of processing the tag data through the second type of model in the model set to generate reply tag data and / or instruction tag data includes: obtaining operation parameters of each smart device within the current set range; after processing the tag data through the second type of model in the model set, generating reply tag data and / or instruction tag data based on the operation parameters of the smart device.

[0011] Optionally, the step of generating reply label data and / or instruction label data based on the operating parameters of the intelligent device includes: judging the operating state of the intelligent device indicated in the label data based on the operating parameters of the intelligent device; generating instruction label data based on the operating state; the instruction label data includes at least one of the following: device-on control parameters, device-off control parameters, and device-adjustment control parameters.

[0012] Optionally, after generating the instruction label data based on the operating state, the control method further includes: if the instruction label data includes the device-on control parameters, generating reply-on label data, or, if the instruction label data includes the device-off control parameters, generating reply-off label data, or, if the instruction label data includes the device-adjustment control parameters, generating reply-adjustment label data.

[0013] Optionally, the type of the first type of model includes at least one of the following: deep learning model, probability model, supervised learning model, and unsupervised learning model.

[0014] According to another aspect of the embodiments of the present invention, there is also provided a control system for an intelligent device, including: a semantic parsing system for parsing information data of a target user and outputting label data, where the semantic parsing system includes a first type of model having no association relationship, and the number N of the first type of models is greater than 1; a dialogue management system for processing the label data to generate reply label data and / or instruction label data, where the reply label data is used to reply interaction information to the target user, the instruction label data is used to send an operation instruction to the intelligent device, and the instruction label data carries device control parameters; a natural language generation system for converting the reply label data into a reply language for interacting with the target user and converting the instruction label data into instruction data recognizable by the intelligent device, where the instruction data is used to control the intelligent device to execute corresponding instructions.

[0015] According to another aspect of the embodiments of the present invention, there is also provided a control device for an intelligent device, including: a receiving unit for receiving information data of a target user, where the information data includes voice information data and / or text information data; an analysis unit for respectively analyzing the information data through a first type of model in a model set to obtain label data, where the number N of the first type of models is greater than 1 and there is no association relationship between the first type of models; a processing unit for processing the label data through a second type of model in the model set to generate reply label data and / or instruction label data, where the reply label data is used to reply interaction information to the target user, the instruction label data is used to send an operation instruction to the intelligent device, and the instruction label data carries device control parameters.

[0016] Optionally, the control device further includes: a first input module, configured to input historical information data and original tag data before receiving information data of a target user, where the historical information data includes historical voice information data and / or historical text information data, and the original tag data is data obtained by parsing the historical information data and processing the parsed information data during a historical process; a first training module, configured to train the original tag data based on a third type of model in a model set to obtain trained tag data.

[0017] Optionally, the control device further includes: a first acquisition module, configured to acquire historical operation data of an intelligent device after parsing the information data through a first type of model in a model set to obtain tag data; a first update module, configured to update the parsed tag data through a third type of model in the model set based on the historical operation data to obtain updated tag data.

[0018] Optionally, the control device further includes: a first deletion module, configured to, after parsing the information data through a first type of model in a model set to obtain tag data, if there are tag data with a repetition rate greater than a preset threshold in the updated tag data, retain the tag data and delete the tag data with a repetition rate less than or equal to the preset threshold.

[0019] Optionally, the processing unit includes: a second acquisition module, configured to acquire operation parameters of each intelligent device within a current set range; a first generation module, configured to generate a reply tag data and / or an instruction tag data based on the operation parameters of the intelligent device after processing the tag data through a second type of model in the model set.

[0020] Optionally, the first generation module includes: a first judgment sub-module, configured to judge an operation state of the intelligent device indicated in the tag data based on the operation parameters of the intelligent device; a first generation sub-module, configured to generate an instruction tag data based on the operation state; the instruction tag data includes at least one of the following: an on-device control parameter, an off-device control parameter, and an adjust-device control parameter.

[0021] Optionally, the control device further includes: after generating the instruction tag data based on the operation state, a second generation module, configured to generate a reply on-tag data if the instruction tag data includes the on-device control parameter, or a third generation module, configured to generate a reply off-tag data if the instruction tag data includes the off-device control parameter, or a fourth generation module, configured to generate a reply adjust-tag data if the instruction tag data includes the adjust-device control parameter.

[0022] Optionally, the type of the first type of model includes at least one of the following: a deep learning model, a probability model, a supervised learning model, and an unsupervised learning model.

[0023] According to another aspect of the embodiments of the present invention, there is also provided a control system for an intelligent device, including: a semantic parsing system for parsing information data of a target user and outputting tag data, wherein the semantic parsing system includes a first type of model without an associated relationship, and the number N of the first type of models is greater than 1; a dialogue management system for processing the tag data to generate reply tag data and / or instruction tag data, wherein the reply tag data is used to reply to the target user with interaction information, the instruction tag data is used to send an operation instruction to the intelligent device, and the instruction tag data carries device control parameters; a natural language generation system for converting the reply tag data into a reply language for interacting with the target user and converting the instruction tag data into instruction data recognizable by the intelligent device, wherein the instruction data is used to control the intelligent device to execute corresponding instructions.

[0024] According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the control method for an intelligent device according to any one of the above by executing the executable instructions.

[0025] According to another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, the computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the control method for an intelligent device according to any one of the above.

[0026] In this application, information data of a target user is received, where the information data includes voice information data and / or text information data. The information data is respectively parsed by the first type of models in the model set to obtain tag data. The number N of the first type of models is greater than 1, and there is no correlation relationship among the first type of models. The tag data is processed by the second type of models in the model set to generate reply tag data and / or instruction tag data. The reply tag data is used to reply interaction information to the target user, and the instruction tag data is used to send operation instructions to the intelligent device. The instruction tag data carries device control parameters. This application optimizes the original serial semantic parsing model into a parallel multi-candidate dynamic parsing model through a dialogue system with parallel first type of models (which can be semantic parsing models), combines the running state of the current intelligent device, generates corresponding reply messages for interacting with the user, and sends corresponding device instructions to control the operation of the intelligent device, thereby reducing the risk of interaction failure caused by cascading errors of the serial semantic parsing model, enhancing the user experience, and further solving the technical problem in the related art that the serial semantic parsing model is prone to parsing errors in one of the models, resulting in parsing errors in other related models, causing interaction failure and reducing the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings described herein are used to provide a further understanding of the present invention, form a part of this application, and the illustrative embodiments and descriptions thereof of the present invention are used to explain the present invention, and do not constitute an improper limitation to the present invention. In the drawings:

[0028] Figure 1 is a schematic diagram of a serial semantic parsing model according to the prior art;

[0029] Figure 2 is a flowchart of an optional control method for an intelligent device according to an embodiment of the present invention;

[0030] Figure 3 is a schematic diagram of an optional dialogue system for controlling an intelligent device according to an embodiment of the present invention;

[0031] Figure 4 is a schematic diagram of an optional control system for an intelligent device according to an embodiment of the present invention;

[0032] Figure 5 is a schematic diagram of a control device for an intelligent device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] To enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0034] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or are inherent to these processes, methods, products or devices.

[0035] To facilitate the understanding of the present invention by those skilled in the art, the following explains some terms or nouns involved in the embodiments of the present invention:

[0036] Natural Language Processing (NLP) refers to the technology of using the natural language used by humans for communication to interact with machines.

[0037] A slot refers to an attribute that has been clearly defined for an entity and is composed of slot positions.

[0038] IOT: Internet of Thing, the Internet of Things, refers to the ubiquitous connection of things to things and things to people through various possible networks, and realizes the intelligent perception, identification and management of items and processes.

[0039] The following embodiments of the present invention can be applied to an intelligent device control dialogue system. Among them, the intelligent devices include but are not limited to: air conditioners, washing machines, refrigerators, sweeping robots, etc. The intelligent device control dialogue system receives the information data of the user (for example, voice data, text input data, etc.), parses the information data to identify the user's intention, controls the intelligent device to work according to the user's intention, and finally interacts with the user the control result.

[0040] Embodiment 1

[0041] According to an embodiment of the present invention, there is provided an embodiment of a control method for a smart device. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. And although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0042] Figure 2 is a flowchart of an alternative control method for a smart device according to an embodiment of the present invention, as Figure 2 shown, the method includes the following steps:

[0043] Step S202, receiving information data of a target user, where the information data includes voice information data and / or text information data.

[0044] Step S204, respectively parsing the information data through the first type of models in the model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship between the first type of models.

[0045] Step S206, processing the tag data through the second type of models in the model set to generate reply tag data and / or instruction tag data, where the reply tag data is used to reply interaction information to the target user, the instruction tag data is used to issue an operation instruction to the smart device, and the device control parameters are carried in the instruction tag data.

[0046] Through the above steps, the information data of the target user can be received, where the information data includes voice information data and / or text information data. The information data is respectively parsed through the first type of models in the model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship between the first type of models. The tag data is processed through the second type of models in the model set to generate reply tag data and / or instruction tag data, where the reply tag data is used to reply interaction information to the target user, the instruction tag data is used to issue an operation instruction to the smart device, and the device control parameters are carried in the instruction tag data. In the embodiment of the present invention, through a dialogue system of parallel first type of models (which can indicate semantic parsing models), the original serial semantic parsing model is optimized into a parallel multi-candidate dynamic parsing model. Combining with the running state of the current smart device, corresponding reply messages for interacting with the user are generated, and corresponding device instructions are issued to control the operation of the smart device. Thereby, the risk of interaction failure caused by cascading errors of the serial semantic parsing model can be reduced, the user experience can be increased, and further the technical problem in the related art that the serial semantic parsing model is prone to cause parsing errors of other associated models due to parsing errors of one of the models, resulting in interaction failure and reducing the user experience is solved.

[0047] The embodiments of the present invention will be described in detail below in combination with the above steps.

[0048] The execution subject of the following steps can be an intelligent device control dialogue system.

[0049] Step S202: Receive the information data of the target user, where the information data includes voice information data and / or text information data.

[0050] In the embodiments of the present invention, the information data can be the voice of the user directly interacting with the dialogue system (for example, the user can directly say to the dialogue system (which can run on a home control terminal or a cloud platform) "raise the air conditioner temperature"), or it can be the text information manually input by the user through the terminal (for example, open the input box of the dialogue system in the terminal and input "raise the air conditioner temperature"). Here, the terminal includes but is not limited to: mobile phones, iPads, PCs, tablets, etc.

[0051] Optionally, before receiving the information data of the target user, the control method further includes: inputting historical information data and original label data, where the historical information data includes historical voice information data and / or historical text information data, and the original label data is the data obtained by parsing the historical information data and processing the parsed information data during the historical process; training the original label data based on the third type of model in the model set to obtain the trained label data.

[0052] In the embodiments of the present invention, the historical information data refers to the information data of the user interacting with the dialogue system within a historical time period (for example, the voice data of the user, the text data input by the user, etc.), and the original label data refers to the data for annotating the historical information data of the user. By training the decision-making model (i.e., the third type of model in the model set), the trained label data can be obtained.

[0053] Step S204: Parse the information data respectively through the first type of model in the model set to obtain label data, where the number N of the first type of model is greater than 1, and there is no association relationship between the first type of models.

[0054] In the embodiments of the present invention, the first type of model in the model set refers to the semantic parsing model in the dialogue system. The number of the semantic parsing models can be multiple, and the multiple semantic parsing models are independent of each other and are in a parallel relationship of being candidates for each other. By parsing the information data of the user through the first type of model, a set of label data can be obtained.

[0055] Optionally, after parsing the information data by the first type of models in the model set to obtain label data, the control method further includes: obtaining historical operation data of the intelligent device; and updating the parsed label data by the third type of models in the model set based on the historical operation data to obtain updated label data.

[0056] In the embodiment of the present invention, the operation state of the intelligent device changes dynamically. For example, when the temperature of the air conditioner is adjusted by the user from 16 degrees to 26 degrees, the first type of models can parse the user's information data in real time according to the operation state data of the intelligent device to obtain label data, and then the trained decision-making model (i.e., the third type of models in the model set) updates the label data to obtain updated label data.

[0057] Optionally, after parsing the information data by the first type of models in the model set to obtain label data, the control method further includes: if there is label data with a repetition rate greater than a preset threshold in the updated label data, retaining the label data and deleting the label data with a repetition rate less than or equal to the preset threshold.

[0058] In the embodiment of the present invention, the results obtained by parsing the user's information data by multiple semantic parsing models may overlap. If the overlap rate is greater than a preset overlap rate (for example, there are 3 overlapping data), the label data is retained and the non-overlapping label data is discarded.

[0059] Another option is to sort the label data obtained by parsing by multiple semantic parsing models according to the overlap rate, set corresponding weights for the label data according to the sorting result, and further screen the label data according to the corresponding weights.

[0060] Step S206: Process the label data by the second type of models in the model set to generate reply label data and / or instruction label data, where the reply label data is used to reply interaction information to the target user, and the instruction label data is used to issue an operation instruction to the intelligent device, and the instruction label data carries device control parameters.

[0061] In the embodiment of the present invention, the label data is processed by the dialogue management model (i.e., the second type of models in the model set) to generate interaction information data that can be replied to the target user and operation instructions that can control the operation of the intelligent device. The operation instructions include but are not limited to: turning on device control parameters, turning off device control parameters, adjusting device control parameters, etc.

[0062] Optionally, the step of processing the label data by the second type of model in the model set to generate reply label data and / or instruction label data includes: obtaining the operating parameters of each intelligent device within the current set range; after processing the label data by the second type of model in the model set, generating reply label data and / or instruction label data based on the operating parameters of the intelligent device.

[0063] In an embodiment of the present invention, the operating parameters of the intelligent devices in the space where the user is located (i.e., the current set range, for example, the home, company, etc. where the user is located) are obtained, and based on the operating parameters, information for interacting with the user is generated, and the operating state of the intelligent devices is controlled to change.

[0064] Optionally, the step of generating reply label data and / or instruction label data based on the operating parameters of the intelligent device includes: judging the operating state of the intelligent device indicated in the label data based on the operating parameters of the intelligent device; generating instruction label data based on the operating state; the instruction label data includes at least one of the following: device-on control parameters, device-off control parameters, device-adjustment control parameters.

[0065] In an embodiment of the present invention, through the operating parameters of the intelligent devices stored in the IOT cloud, the operating state of the intelligent device that the user wants to control can be judged (for example, the user wants to increase the air conditioner temperature, and now the air conditioner is in the off state). Based on the current operating state of the intelligent device, corresponding instruction label data is generated to control the intelligent device to work. For example, according to the current off state of the air conditioner, an on instruction and an instruction to increase the air conditioner temperature need to be generated.

[0066] Optionally, after generating the instruction label data based on the operating state, the control method further includes: if the instruction label data includes device-on control parameters, generating reply-on label data, or, if the instruction label data includes device-off control parameters, generating reply-off label data, or, if the instruction label data includes device-adjustment control parameters, generating reply-adjustment label data.

[0067] In an embodiment of the present invention, if the issued instruction is an on instruction, a message indicating that the corresponding intelligent device has been turned on needs to be replied to the user. If the issued instruction is an off instruction, a message indicating that the corresponding intelligent device has been turned off needs to be replied to the user. If the issued instruction is an adjustment instruction, a message indicating that the corresponding intelligent device has been adjusted to the corresponding state needs to be replied to the user. The reply method and the format of the reply content of the information are not limited herein. For example, the user can be replied with "Master, the air conditioner has been turned on and adjusted to 26 degrees according to your instructions" in voice.

[0068] Optionally, the type of the first type of model includes at least one of the following: deep learning model, probability model, supervised learning model, and unsupervised learning model.

[0069] In the embodiments of the present invention, the type of the first type of model (i.e., the type of multiple semantic parsing models that are parallel to each other) can be a model that selects a deep learning algorithm, a model that selects a probability algorithm, a model that selects a supervised learning algorithm or an unsupervised learning algorithm.

[0070] In the embodiments of the present invention, by constructing an intelligent device control dialogue system architecture with a parallel semantic parsing model, the sub-models that were originally serially cascaded in the semantic parsing model are optimized into a parallel multi-candidate dynamic optimization path selection. Combining the current operating state of the intelligent device, an optimal decision is made in dialogue management, which can solve the cascading error problem of semantic parsing in the traditional intelligent device field interaction system and improve the user experience.

[0071] Embodiment Two

[0072] Figure 3 is a schematic diagram of an optional dialogue system for controlling an intelligent device according to an embodiment of the present invention, as Figure 3 shown, including: semantic parsing modules 1, 2... N, a dialogue management decision module, and a natural language generation module.

[0073] The semantic parsing modules 1, 2... N included in the dialogue system in the embodiments of the present invention are parallel parsing modules. The information data of the user is parsed through multiple parallel parsing modules, and then combined with the Internet of Things data. An optimal decision is made through the dialogue management decision module to obtain the information for interacting with the user and the instructions for controlling the intelligent device. Through the natural language generation module, a reply language that the user can understand is finally obtained for interacting with the user, and an instruction is issued to the intelligent device indicated by the information data of the user to control the intelligent device to work according to the user's intention.

[0074] In the embodiments of the present invention, by constructing an intelligent device control dialogue system architecture with a parallel semantic parsing model, the cascading error problem of semantic parsing in the traditional intelligent device field interaction system can be avoided, and the user experience can be improved.

[0075] Embodiment Three

[0076] Figure 4 is a schematic diagram of an optional control system for an intelligent device according to an embodiment of the present invention, as Figure 4 shown, including: a semantic parsing system 40, a dialogue management system 42, and a natural language generation system 44, where

[0077] The semantic parsing system 40 is used to parse the information data of the target user and output tag data. Among them, the semantic parsing system includes a first type of model without an associated relationship, and the number N of the first type of models is greater than 1;

[0078] The dialogue management system 42 is used to process the tag data and generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply to the target user with interactive information, and the instruction tag data is used to send an operation instruction to the intelligent device. The instruction tag data carries device control parameters;

[0079] The natural language generation system 44 is used to convert the reply tag data into a reply language for interacting with the target user, and convert the instruction tag data into instruction data that can be recognized by the intelligent device. Among them, the instruction data is used to control the intelligent device to execute the corresponding instruction.

[0080] In the embodiment of the present invention, the information data of the target user can be received. Among them, the information data includes voice information data and / or text information data. The information data is respectively parsed by the parallel semantic parsing models in the semantic parsing system 40 to obtain tag data. The tag data is processed by the dialogue management system 42 to generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply to the target user with interactive information, and the instruction tag data is used to send an operation instruction to the intelligent device. The instruction tag data carries device control parameters. Then, the reply tag data is converted into a reply language for interacting with the target user by the natural language generation system 44, and the instruction tag data is converted into instruction data that can be recognized by the intelligent device. Among them, the instruction data is used to control the intelligent device to execute the corresponding instruction. In the embodiment of the present invention, through the dialogue system of the first type of model (indicating the semantic parsing model) in parallel, that is, the original serial semantic parsing model is optimized into a parallel multi-candidate dynamic parsing model, combined with the running state of the current intelligent device, a corresponding reply language for interacting with the user is generated, and the corresponding device instruction is sent to control the operation of the intelligent device, so as to reduce the risk of interaction failure caused by the cascade error of the serial semantic parsing model, increase the user experience, and further solve the technical problem that in the related art, the serial semantic parsing model is prone to parsing errors in one of the models, resulting in parsing errors in other associated models, causing interaction failure and reducing the user experience.

[0081] Embodiment 4

[0082] A control device for an intelligent device provided in this embodiment includes a plurality of implementation units, and each implementation unit corresponds to each implementation step in Embodiment 1 above.

[0083] Figure 5 is a schematic diagram of a control device for an intelligent device according to an embodiment of the present invention, as Figure 5As shown in the figure, the authentication device may include: a receiving unit 50, a parsing unit 52, and a processing unit 54. Among them,

[0084] The receiving unit 50 is configured to receive information data of a target user. Among them, the information data includes voice information data and / or text information data;

[0085] The parsing unit 52 is configured to respectively parse the information data through the first type of models in the model set to obtain tag data. Among them, the number N of the first type of models is greater than 1, and there is no correlation relationship between the first type of models;

[0086] The processing unit 54 is configured to process the tag data through the second type of models in the model set to generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply interaction information to the target user, and the instruction tag data is used to issue an operation instruction to the intelligent device. The instruction tag data carries device control parameters.

[0087] The above control device can receive the information data of the target user through the receiving unit 50. Among them, the information data includes voice information data and / or text information data. The parsing unit 52 respectively parses the information data in the first type of models in the model set to obtain tag data. Among them, the number N of the first type of models is greater than 1, and there is no correlation relationship between the first type of models. The processing unit 54 processes the tag data in the second type of models in the model set to generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply interaction information to the target user, and the instruction tag data is used to issue an operation instruction to the intelligent device. The instruction tag data carries device control parameters. In the embodiment of the present invention, through a dialogue system with parallel first type of models (indicating semantic parsing models), that is, the original serial semantic parsing model is optimized into a parallel multi-candidate dynamic parsing model, combined with the running state of the current intelligent device, a corresponding reply language for interacting with the user is generated, and a corresponding device instruction is issued to control the operation of the intelligent device, thereby being able to reduce the risk of interaction failure caused by cascading errors of the serial semantic parsing model, increase the user experience, and further solve the technical problem in the related art that the serial semantic parsing model is prone to cause parsing errors of other associated models due to parsing errors of a certain model, resulting in interaction failure and reducing the user experience.

[0088] Optionally, the control device further includes: a first input module, configured to input historical information data and original tag data before receiving the information data of the target user. Among them, the historical information data includes historical voice information data and / or historical text information data, and the original tag data is data obtained by parsing the historical information data and processing the parsed information data during the historical process; a first training module, configured to train the original tag data based on the third type of models in the model set to obtain trained tag data.

[0089] Optionally, the control device further includes: a first acquisition module, configured to acquire historical operation data of the intelligent device after parsing information data through the first type of models in the model set to obtain label data; a first update module, configured to update the parsed label data through the third type of models in the model set based on the historical operation data to obtain updated label data.

[0090] Optionally, the control device further includes: a first deletion module, configured to, after parsing information data through the first type of models in the model set to obtain label data, if there is label data with a repetition rate greater than a preset threshold in the updated label data, retain the label data and delete the label data with a repetition rate less than or equal to the preset threshold.

[0091] Optionally, the processing unit includes: a second acquisition module, configured to acquire operation parameters of each intelligent device within the current set range; a first generation module, configured to generate reply label data and / or instruction label data based on the operation parameters of the intelligent device after processing the label data through the second type of models in the model set.

[0092] Optionally, the first generation module includes: a first judgment sub-module, configured to judge the operation state of the intelligent device indicated in the label data based on the operation parameters of the intelligent device; a first generation sub-module, configured to generate instruction label data based on the operation state; the instruction label data includes at least one of the following: device-on control parameters, device-off control parameters, device-adjustment control parameters.

[0093] Optionally, the control device further includes: after generating instruction label data based on the operation state, a second generation module, configured to generate a reply-on label data if the instruction label data includes device-on control parameters, or a third generation module, configured to generate a reply-off label data if the instruction label data includes device-off control parameters, or a fourth generation module, configured to generate a reply-adjust label data if the instruction label data includes device-adjustment control parameters.

[0094] Optionally, the type of the first type of models includes at least one of the following: deep learning model, probability model, supervised learning model, unsupervised learning model.

[0095] The above authentication device may further include a processor and a memory. The above receiving unit 50, parsing unit 52, processing unit 54, etc. are all stored in the memory as program units, and the processor executes the above program units stored in the memory to implement corresponding functions.

[0096] The above-mentioned processor includes a kernel, which retrieves corresponding program units from the memory. One or more kernels can be set, and reply tag data and / or instruction tag data are generated by adjusting kernel parameters.

[0097] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory includes at least one memory chip.

[0098] The present application also provides a computer program product, which is adapted to execute a program initialized with the following method steps when executed on a data processing device: receiving information data of a target user, respectively parsing the information data through the first type of models in a model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship between the first type of models, and processing the tag data through the second type of models in the model set to generate reply tag data and / or instruction tag data.

[0099] According to another aspect of the embodiments of the present invention, there is also provided an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to execute the control method for a smart device according to any one of the above through executing the executable instructions.

[0100] According to another aspect of the embodiments of the present invention, there is also provided a computer-readable storage medium, which includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the control method for a smart device according to any one of the above.

[0101] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages and disadvantages of the embodiments.

[0102] In the above embodiments of the present invention, the descriptions of the respective embodiments have their own emphases. For parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0103] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the division of the units can be a logical function division, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings or direct couplings or communication connections shown or discussed with each other can be through some interfaces, and the indirect couplings or communication connections of units or modules can be in an electrical or other form.

[0104] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed over multiple units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0105] In addition, each functional unit in various embodiments of the present invention can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0106] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs, etc., which can store program codes.

[0107] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A control method for a smart device, characterized in that, Including: Receiving information data of a target user, where the information data includes voice information data and / or text information data; Respectively parsing the information data through a first type of model in a model set to obtain tag data, where the number N of the first type of models is greater than 1, and there is no association relationship between the first type of models; Processing the tag data through a second type of model in the model set to generate reply tag data and / or instruction tag data, where the reply tag data is used to reply interaction information to the target user, and the instruction tag data is used to issue an operation instruction to an intelligent device, and the instruction tag data carries device control parameters; After respectively parsing the information data through a first type of model in the model set to obtain tag data, the control method further includes: obtaining historical operation data of the intelligent device; based on the historical operation data, updating the parsed tag data through a third type of model in the model set to obtain updated tag data; if there is tag data with a repetition rate greater than a preset threshold in the updated tag data, retaining the tag data and deleting the tag data with a repetition rate less than or equal to the preset threshold.

2. The control method according to claim 1, characterized in that Before receiving the information data of the target user, the control method further includes: Inputting historical information data and original tag data, where the historical information data includes historical voice information data and / or historical text information data, and the original tag data is data obtained by parsing the historical information data and processing the parsed information data during the historical process; Training the original tag data based on a third type of model in the model set to obtain trained tag data.

3. The control method according to claim 1, characterized in that The step of processing the tag data through a second type of model in the model set to generate reply tag data and / or instruction tag data includes: Obtaining operation parameters of each intelligent device within the current set range; After processing the tag data through a second type of model in the model set, generating reply tag data and / or instruction tag data based on the operation parameters of the intelligent device.

4. The control method according to claim 3, characterized in that The step of generating reply tag data and / or instruction tag data based on the operation parameters of the intelligent device includes: Judging the operation state of the intelligent device indicated in the tag data based on the operation parameters of the intelligent device; Generating instruction tag data based on the operation state; The instruction tag data includes at least one of the following: device-on control parameters, device-off control parameters, device-adjustment control parameters.

5. The control method according to claim 4, wherein After generating the instruction tag data based on the operation state, the control method further includes: If the instruction tag data includes the device-on control parameters, generating a reply device-on tag data, or, If the instruction tag data includes the device-off control parameters, generating a reply device-off tag data, or, If the instruction tag data includes the device-adjustment control parameters, generating a reply device-adjustment tag data.

6. The control method according to claim 1, wherein The type of the first type of model includes at least one of the following: deep learning model, probability model, supervised learning model, unsupervised learning model.

7. A control system for a smart device, characterized in that, Including: A semantic parsing system for parsing information data of a target user and outputting tag data. Among them, the semantic parsing system includes a first type of models without an associated relationship, and the number N of the first type of models is greater than 1; A dialogue management system for processing the tag data to generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply interactive information to the target user, and the instruction tag data is used to send an operation instruction to an intelligent device, and the instruction tag data carries device control parameters; A natural language generation system for converting the reply tag data into a reply language for interacting with the target user and converting the instruction tag data into instruction data recognizable by the intelligent device. Among them, the instruction data is used to control the intelligent device to execute corresponding instructions.

8. A control device for a smart device, characterized in that, Comprising: A receiving unit for receiving information data of a target user. Among them, the information data includes voice information data and / or text information data; A parsing unit for respectively parsing the information data through the first type of models in a model set to obtain tag data. Among them, the number N of the first type of models is greater than 1, and there is no associated relationship among the first type of models; A processing unit for processing the tag data through a second type of model in the model set to generate reply tag data and / or instruction tag data. Among them, the reply tag data is used to reply interactive information to the target user, and the instruction tag data is used to send an operation instruction to an intelligent device, and the instruction tag data carries device control parameters; The control device further includes: a first acquisition module for acquiring historical operation data of the intelligent device after respectively parsing the information data through the first type of models in the model set to obtain tag data; a first update module for updating the parsed tag data through a third type of model in the model set based on the historical operation data to obtain updated tag data; The control device further includes: a first deletion module for, after respectively parsing the information data through the first type of models in the model set to obtain tag data, if there are tag data with a repetition rate greater than a preset threshold in the updated tag data, retaining the tag data and deleting the tag data with a repetition rate less than or equal to the preset threshold.

9. An electronic device, characterized in that, Comprising: A processor; And A memory for storing executable instructions of the processor; Among them, the processor is configured to execute the control method for an intelligent device according to any one of claims 1 to 6 by executing the executable instructions.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program, wherein when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the control method for an intelligent device according to any one of claims 1 to 6.

Images