A mobile phone-based AI digital human automatic learning system and usage method

Abstract

The present invention provides an AI digital human automatic learning system and method based on a mobile phone, including an AI digital human control center, a receiving module, a morphology perception module and a feedback module; the receiving module receives a control signal sent by a control signal interface of a mobile phone terminal; the morphology perception module inputs the control signal into an AI digital human perception simulation model to obtain an embodied morphology perception result output by the AI ​​digital human perception simulation model; the feedback module returns the embodied morphology perception result through the perception module interface. The present invention controls the corresponding target intelligent body to perform corresponding actions in a target test environment based on the input control signal through the AI ​​digital human perception simulation model, and perceives it to obtain the corresponding embodied morphology perception result, thereby realizing the embodied intelligent body's cognition of the embodied morphology, and further improving the scalability and versatility of the embodied intelligent body.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a mobile phone-based AI digital human automatic learning system and method thereof. Background Art

[0002] At present, the explosion of large models has further integrated artificial intelligence into people's lives. On the road to general artificial intelligence (AGI), more and more research is being conducted on embodied agents. Embodied agents have a variety of embodied forms, and different forms often have different functions. Existing embodied agent research is often based on training or exploratory learning based on predetermined embodied forms. In order to train embodied intelligence, researchers have proposed different simulators to serve the purpose of "making the agent learn specific skills", covering skills such as "finding the way in indoor scenes", "finding specified objects", and "answering questions based on things in the scene".

[0003] However, the skills acquired through this training method often depend on the embodied form, have limited generalization capabilities, and there is little research on how agents understand their own embodied form, and even fewer simulation platforms for training this ability. Summary of the invention

[0004] The present invention provides an AI digital human automatic learning system and method based on a mobile phone, which are used to solve the defects in the prior art that the skills and generalization ability of the intelligent body are limited due to the fixed embodied form, and realize the embodied form recognition of the embodied intelligent body.

[0005] The technical solution of the present invention to solve the above technical problems is as follows:

[0006] An AI digital human automatic learning system based on a mobile phone, comprising an AI digital human control center, a receiving module, a morphological perception module and a feedback module; the AI ​​digital human control center is connected to the receiving module, the morphological perception module and the feedback module respectively to manage the data of each module;

[0007] The receiving module is used to receive the control signal sent by the control signal interface of the mobile terminal;

[0008] The morphology perception module is used to input the control signal into the AI ​​digital human perception simulation model to obtain the embodied morphology perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied morphology perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding actions in the target test environment to obtain the corresponding embodied morphology perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body morphology according to the control signal;

[0009] The feedback module is used to return the embodied form perception result through the perception module interface.

[0010] The present invention provides an AI digital human automatic learning method based on a mobile phone, comprising: receiving a control signal sent by a control signal interface; inputting the control signal into an AI digital human perception simulation model to obtain an embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on a control training signal and an embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control a corresponding target intelligent body to perform a corresponding action in a target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding action in the target test environment to obtain a corresponding embodied form perception result, wherein the target intelligent body is selected from a virtual intelligent body previously simulated based on the embodied intelligent body form according to the control signal.

[0011] According to an AI digital human automatic learning method based on a mobile phone provided by the present invention, the AI ​​digital human perception simulation model includes a test environment simulation layer, an environment perception layer and a perception simulation layer; a control signal is input into the AI ​​digital human perception simulation model to obtain an embodied form perception result output by the AI ​​digital human perception simulation model, including: inputting the control signal into the test environment simulation layer to control the corresponding target intelligent body to perform a corresponding action in a target test environment according to the control signal; wherein the target test environment is a test environment that is first selected based on the control signal and injected with preset environmental noise; using the environmental perception layer to perceive the target intelligent body after performing the corresponding action in the target test environment to obtain a first embodied form perception result; inputting the first embodied form perception result into the perception simulation layer to process the first embodied form perception result using the preset perception noise to obtain the embodied form perception result.

[0012] According to an AI digital human automatic learning method based on a mobile phone provided by the present invention, there are at least two environmental perception layers, and the target test environment is perceived from different angles respectively; the environmental perception layer includes first-level parameters and second-level parameters, the first-level parameters are used to control the perception ability and noise intensity of each perception channel, and the second-level parameters are obtained by encapsulating the corresponding first-level parameters based on preset encapsulation rules, and the preset encapsulation rules are used to limit the first-level parameters that are changed by the change of the second-level parameters to be encapsulated as corresponding second-level parameters.

[0013] According to an automatic learning method for AI digital humans based on a mobile phone provided by the present invention, the preset perception noise includes receiving noise, and the receiving noise is used to limit the interference in the process of inputting the first embodied form perception result into the perception simulation layer, and the receiving noise is obtained based on turning on or off at least one environmental perception layer; before inputting the first embodied form perception result into the perception simulation layer, it includes: controlling to turn on or off at least one environmental perception layer to generate receiving noise; inputting the first embodied form perception result into the perception simulation layer, it includes: based on the receiving noise, receiving the first embodied form perception result sent by the environmental perception layer in the turned-on state.

[0014] According to an automatic learning method for AI digital humans based on a mobile phone provided by the present invention, the preset perception noise includes environmental perception noise, and the environmental perception noise is used to limit interference in the process of obtaining a first embodied form perception result; the embodied form perception result is processed using the preset perception noise to obtain the embodied form perception result, including: adding environmental perception noise to the first embodied form perception result.

[0015] According to an automatic learning method for AI digital humans based on a mobile phone provided by the present invention, the embodied form perception result is returned through the perception module interface, including: returning at least two embodied form perception results through the perception module interface, so that the receiving end can perform quality evaluation on the at least two embodied form perception results received, and select the embodied form perception result with the highest quality evaluation result as the target form embodied form perception result; or, returning at least two embodied form perception results through the perception module interface, so that the receiving end can perform quality evaluation on the at least two embodied form perception results received, and based on the case where the error of each quality evaluation result is lower than a preset threshold, averaging all the received embodied form perception results to obtain the target form embodied form perception result; or, returning at least two embodied form perception results through the perception module interface, so that the receiving end can randomly select the at least two embodied form perception results received to obtain the target form embodied form perception result.

[0016] According to a mobile phone-based AI digital human automatic learning method provided by the present invention, the AI ​​digital human perception simulation model also includes a control simulation layer; before the control signal is input into the test environment simulation layer, it includes: processing the control signal based on a preset control noise; wherein the control noise is first configured based on interference in the process of receiving the control signal.

[0017] The present invention also provides an electronic device, comprising: a memory for storing a computer software program; a processor for reading and executing the computer software program, thereby implementing any of the above-mentioned mobile phone-based AI digital human automatic learning methods.

[0018] The present invention also provides a non-transitory computer-readable storage medium, in which a computer software program is stored. When the computer software program is executed by a processor, it implements any of the above-mentioned mobile phone-based AI digital human automatic learning methods.

[0019] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned mobile phone-based AI digital human automatic learning methods.

[0020] The embodiment of the present invention uses an AI digital human perception simulation model to control the corresponding target intelligent agent to perform corresponding actions in the target test environment based on the input control signal, and perceive it to obtain the corresponding embodied form perception result, thereby realizing the embodied intelligent agent's cognition of the embodied form and further improving the scalability and versatility of the embodied intelligent agent. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a structural schematic diagram of the mobile phone-based AI digital human automatic learning system provided by the present invention;

[0022] Figure 2 It is a flowchart of the mobile phone-based AI digital human automatic learning method provided by the present invention;

[0023] Figure 3 It is a schematic diagram of the architecture of the mobile phone-based AI digital human automatic learning method provided by the present invention;

[0024] Figure 4 A schematic diagram of an electronic device according to an embodiment of the present invention;

[0025] Figure 5 A schematic diagram of an embodiment of a computer-readable storage medium provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0027] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0028] In the description of the present invention, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in the present invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present invention.

[0029] Reference Figure 1 , Figure 1 It is a structural schematic diagram of the mobile phone-based AI digital human automatic learning system provided by the present invention. The mobile phone-based AI digital human automatic learning system includes an AI digital human control center, a receiving module, a morphological perception module and a feedback module; the AI ​​digital human control center is respectively connected to the receiving module, the morphological perception module and the feedback module to manage the data of each module.

[0030] Optionally, the receiving module receives a control signal sent by a control signal interface of the mobile phone terminal.

[0031] Optionally, the morphology perception module inputs the control signal into the AI ​​digital human perception simulation model to obtain the embodied morphology perception result output by the AI ​​digital human perception simulation model.

[0032] Among them, the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent agent to perform corresponding actions in the target test environment according to the input control signal, and perceive the target intelligent agent after performing the corresponding action in the target test environment to obtain the corresponding embodied form perception result. The target intelligent agent is selected from the virtual intelligent agent previously simulated based on the embodied intelligent agent form according to the control signal.

[0033] The AI ​​digital human in this embodiment refers to a virtual human or digital character created using artificial intelligence technology and digital technology. These digital humans can have human-like appearance, language and behavior, and are usually used for various purposes, including virtual assistants, virtual speakers, virtual sales representatives, etc. They can interact with humans through text, voice or visual interfaces to provide users with information, answer questions, perform tasks, etc.

[0034] Optionally, the feedback module returns the embodied form perception result through the perception module interface.

[0035] The embodiment of the present invention uses an AI digital human perception simulation model to control the corresponding target intelligent agent to perform corresponding actions in the target test environment based on the input control signal, and perceive it to obtain the corresponding embodied form perception result, thereby realizing the embodied intelligent agent's cognition of the embodied form and further improving the scalability and versatility of the embodied intelligent agent.

[0036] Figure 2 : is a flow chart of the mobile phone-based AI digital human automatic learning method provided by the present invention, such as Figure 2 As shown, the mobile phone-based AI digital human automatic learning method includes the following steps:

[0037] Step 11, receiving a control signal sent by a control signal interface of the mobile terminal;

[0038] Step 12, inputting the control signal into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; controlling the corresponding target intelligent body to perform the corresponding action in the target test environment according to the input control signal, and perceiving the target intelligent body after performing the corresponding action in the target test environment to obtain the corresponding embodied form perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body form according to the control signal;

[0039] Step 13: Return the embodied form perception result through the perception module interface.

[0040] It should be noted that the step numbers "Step 1N" in this manual do not represent the order of the automatic learning method of the mobile phone-based AI digital human. Figure 3 The present invention describes the method for automatic learning of AI digital human based on mobile phone. Optionally, the AI ​​digital human in this embodiment refers to a virtual human or digital character created using artificial intelligence technology and digital technology. These digital humans can have human-like appearance, language and behavior, and are generally used for various purposes, including virtual assistants, virtual speakers, virtual sales representatives, etc. They can interact with humans through text, voice or visual interfaces to provide users with information, answer questions, perform tasks, etc.

[0041] Step 11, receiving a control signal sent by a control signal interface of a mobile phone terminal.

[0042] It should be noted that the mobile terminal of the control signal can be a simulated embodied agent. In addition, the mobile terminal can select the corresponding control signal according to the target agent to be called and send it to the control signal interface, so as to select the corresponding target agent based on the control signal to perform the corresponding action.

[0043] Step 12, input the control signal into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and perceive the target intelligent body after performing the corresponding action in the target test environment to obtain the corresponding embodied form perception result, and the target intelligent body is selected from the virtual intelligent body based on the embodied intelligent body form simulation according to the control signal.

[0044] Specifically, the AI ​​digital human perception simulation model includes a test environment simulation layer, an environment perception layer, and a perception simulation layer. Accordingly, the control signal is input into the AI ​​digital human perception simulation model to obtain the embodied form perception results output by the AI ​​digital human perception simulation model, including:

[0045] Step A, inputting a control signal into the test environment simulation layer to control the corresponding target intelligent agent to perform corresponding actions in the target test environment according to the control signal; wherein the target test environment is a test environment that is previously selected based on the control signal and injected with preset environmental noise.

[0046] It should be added that the preset environmental noise includes at least one of Gaussian distribution and Poisson distribution. The preset environmental noise can uniformly affect all objects in the target test environment or affect some objects in the target test environment based on the actual noise characteristics involved. For example, when the preset environmental noise is white noise, it can uniformly affect all objects in the target test environment; for example, when the preset environmental noise is light intensity noise, objects in the target environment can be randomly sampled or specific objects can be specified. No further limitation is made here. In addition, the preset environmental noise can be set in advance, and at least one previously configured environmental noise can be directly called during actual training or application.

[0047] It should be noted that the above-mentioned objects include the target agent and other objects in the target test environment. In addition, the target agent is a virtual agent that selects a corresponding embodied form based on the control signal, and the embodied form is a simulated form that is previously configured based on the embodied form of the actual embodied agent.

[0048] In an optional embodiment, in order to improve robustness, the preset environmental noise may also include other intelligent agents, etc., to perform corresponding tasks according to the preset control logic, thereby causing interference to the target intelligent agent when performing corresponding actions according to the control signal. The specific settings can be based on actual design requirements and will not be further elaborated here.

[0049] In an optional embodiment, the AI ​​digital human perception simulation model also includes a control simulation layer; before the control signal is input into the test environment simulation layer, it includes: processing the control signal based on a preset control noise; wherein the control noise is first configured based on the interference in the process of receiving the control signal.

[0050] It should be noted that since the actual process of receiving the control signal is easily interfered with, the success rate of controlling the target intelligent body of the corresponding embodied form according to the received control signal is affected or the strength of the received control signal is reduced. Therefore, a preset control noise is added to the received control signal to realize the simulation of the control ability of the embodied intelligent body, increase the difficulty of the perception process and improve the perception accuracy. The preset control noise can be set based on the interference that may be encountered in the actual control process, such as when the control signal is actually called, and no further limitation is made here.

[0051] In addition, environmental perception noise can be pre-set and directly called during actual training or application to control the injection of noise of different types and intensities, making the entire system more adjustable, thereby simulating a variety of complex scenarios and providing test environments of different difficulty levels to more fully test the ability boundaries of embodied agents in the task of recognizing their own embodied form.

[0052] Step B, using the environmental perception layer, perceives the target intelligent agent after performing the corresponding action in the target test environment to obtain the first embodied form perception result.

[0053] In this embodiment, there are at least two environmental perception layers, and they perceive the target test environment from different angles respectively; the environmental perception layer includes first-level parameters and second-level parameters, the first-level parameters are used to control the perception ability and noise intensity of each perception channel, and the second-level parameters are obtained by encapsulating the corresponding first-level parameters based on preset encapsulation rules, and the preset encapsulation rules are used to limit the first-level parameters that are changed by the change of the second-level parameters to be encapsulated into the corresponding second-level parameters.

[0054] It should be added that the perception types of the environmental perception layer include visual perception, audio perception, lighting perception, posture perception, and object quantitative state perception, etc., so as to simulate the perception ability of embodied intelligent agents from different dimensions. The specific settings can be based on the design requirements of the actual embodied intelligent agent to recognize its own embodied form. The environmental perception layer can be of the same perception type, but for perception at different angles. For example, there are three visual environmental perception layers, and their perception angles relative to the target test environment are different to avoid perceiving the same information.

[0055] It should be noted that the environmental perception layer provides a noise setting interface for setting the noise function type and noise intensity, etc. For example, the noise intensity can be represented by the signal-to-noise ratio. The noise function type and noise intensity can be set through the noise setting interface to make the entire system more adjustable, so as to facilitate the simulation of various complex scenes and realize the cognition of the embodied form of the embodied intelligent body in various scenes. In addition, the preset encapsulation rules can be determined according to the actual association between the first-level parameters and the second-level parameters. For example, when the second-level parameters are changed, the corresponding first-level parameters will be affected, and the affected first-level parameters will be encapsulated into the second-level parameters.

[0056] For example, the second-level parameter is attentiveness. When this parameter is adjusted, the corresponding first-level parameters of visual information, audio information, and noise intensity of object posture will all undergo corresponding actions. Therefore, it is necessary to encapsulate the visual information, audio information, and noise intensity of object posture into attentiveness, while the first-level parameter object quantization state remains unchanged, so it does not need to be encapsulated into the attentiveness.

[0057] Step C: inputting the first embodied form perception result into the perception simulation layer, so as to process the first embodied form perception result by using preset perception noise to obtain the embodied form perception result.

[0058] In this embodiment, the preset perception noise includes receiving noise, and the receiving noise is used to limit the interference in the process of inputting the first embodied form perception result into the perception simulation layer, and the receiving noise is obtained based on turning on or off at least one environmental perception layer. The specific receiving noise can be set based on the interference that may be encountered in the actual perception process, and is not further limited here, so as to simulate the situation that the embodied form perception result may be lost or the embodied form perception result may be inaccurate in the actual perception process, and improve the accuracy of perceiving the embodied form of the embodied intelligent body itself.

[0059] Accordingly, before inputting the first embodied form perception result into the perception simulation layer, it includes: controlling to turn on or off at least one environmental perception layer to generate receiving noise; inputting the first embodied form perception result into the perception simulation layer, it includes: based on the receiving noise, receiving the first embodied form perception result sent by the environmental perception layer in the turned-on state.

[0060] In an optional embodiment, the preset perception noise includes environmental perception noise, and the environmental perception noise is used to limit interference in the process of obtaining the first embodied form perception result. Accordingly, the embodied form perception result is processed by using the preset perception noise to obtain the embodied form perception result, including: adding the environmental perception noise to the first embodied form perception result.

[0061] It should be noted that the environmental perception noise can be set based on the interference that may be encountered during the actual perception process, and no further limitation is made here. In addition, the environmental perception noise can be set in advance, and at least one pre-configured environmental perception noise can be directly called during the actual training or application process to control the injection of noise of different types and intensities, making the entire system more adjustable, thereby simulating a variety of complex scenarios and providing test environments of different difficulties to more fully test the ability boundaries of the embodied intelligent agent in the task of recognizing the self-embodied form.

[0062] In an optional embodiment, before inputting the control signal into the AI ​​digital human perception simulation model, it includes: training the AI ​​digital human perception simulation model. Specifically, training the AI ​​digital human perception simulation model includes: obtaining the control training signal and the embodied form perception label corresponding to the control training signal; dividing the acquired control training signal and the embodied form perception label corresponding to the control training signal according to a preset division ratio to obtain a training set and a test set; wherein the training set and the test set respectively include a certain number of control training signals and embodied form perception labels corresponding to the control training signal; based on a preset number of training iterations, using the training set, training the AI ​​digital human perception simulation model to obtain the trained AI digital human perception simulation model; using the test set, verifying the trained AI digital human perception simulation model, and determining the final AI digital human perception simulation model based on the verification result.

[0063] It should be noted that the specific steps of using the training set to train the AI ​​digital human perception simulation model and using the test set to verify the trained AI digital human perception simulation model and inputting the control training signal into the model can be referred to above and will not be repeated here.

[0064] Step 13, return the embodied form perception result through the perception module interface.

[0065] In this embodiment, when there are at least two environmental perception layers, and at least one environmental perception layer is turned on or off by the perception simulation layer, if the perception simulation layer receives the embodied form perception results sent by at least two environmental perception layers, the embodied form perception results are returned through the perception module interface, including: returning at least two embodied form perception results through the perception module interface, so that the receiving end can perform quality evaluation on the at least two embodied form perception results received, and select the embodied form perception result with the highest quality evaluation result as the target form embodied form perception result; or, returning at least two embodied form perception results through the perception module interface, so that the receiving end can perform quality evaluation on the at least two embodied form perception results received, and based on the error of each quality evaluation result being lower than a preset threshold, averaging all received embodied form perception results to obtain the target form embodied form perception result; or, returning at least two embodied form perception results through the perception module interface, so that the receiving end can randomly select the at least two embodied form perception results received to obtain the target form embodied form perception result. It should be noted that the receiving end is a mobile phone terminal of the control signal, and the details can be referred to above, and no repetition is made here.

[0066] In summary, the embodiment of the present invention uses the AI ​​digital human perception simulation model to control the corresponding target intelligent agent to perform corresponding actions in the target test environment based on the input control signal, and perceive it to obtain the corresponding embodied form perception result, thereby realizing the embodied intelligent agent's cognition of the embodied form and further improving the scalability and versatility of the embodied intelligent agent.

[0067] See also Figure 4 , Figure 4 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 4 As shown, an embodiment of the present invention provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor 420. When the processor 420 executes the computer program 411, the following steps are implemented:

[0068] Receive a control signal sent by a control signal interface of a mobile phone terminal;

[0069] The control signal is input into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding actions in the target test environment to obtain the corresponding embodied form perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body form according to the control signal;

[0070] The embodied form perception result is returned through the perception module interface.

[0071] See also Figure 5 , Figure 5 Schematic diagram of an embodiment of a computer-readable storage medium provided in an embodiment of the present invention. Figure 5 As shown, this embodiment provides a computer-readable storage medium 500, on which a computer program 511 is stored. When the computer program 511 is executed by a processor, the following steps are implemented:

[0072] Receive a control signal sent by a control signal interface of a mobile phone terminal;

[0073] The control signal is input into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding actions in the target test environment to obtain the corresponding embodied form perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body form according to the control signal;

[0074] The embodied form perception result is returned through the perception module interface.

[0075] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0076] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0077] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0078] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0079] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0080] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0081] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. An AI digital human automatic learning system based on a mobile phone, characterized in that: The mobile phone-based AI digital human automatic learning system includes an AI digital human control center, a receiving module, a morphological perception module, and a feedback module; the AI ​​digital human control center is connected to the receiving module, the morphological perception module, and the feedback module respectively to manage the data of each module; The receiving module is used to receive the control signal sent by the control signal interface of the mobile terminal; The morphology perception module is used to input the control signal into the AI ​​digital human perception simulation model to obtain the embodied morphology perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied morphology perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding actions in the target test environment to obtain the corresponding embodied morphology perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body morphology according to the control signal; The feedback module is used to return the embodied form perception result through the perception module interface; The AI ​​digital human perception simulation model includes a test environment simulation layer, an environment perception layer and a perception simulation layer; the control signal is input into the AI ​​digital human perception simulation model to obtain the perception result output by the AI ​​digital human perception simulation model, including: Inputting the control signal into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model includes: Inputting the control signal into the test environment simulation layer to control the corresponding target agent to perform corresponding actions in the target test environment according to the control signal; wherein the target test environment is a test environment that is previously selected based on the control signal and into which preset environmental noise is injected; Using the environmental perception layer, the target intelligent agent is perceived after performing a corresponding action in the target test environment to obtain a first embodied form perception result; Inputting the first embodied form perception result into the perception simulation layer, so as to process the first embodied form perception result by using preset perception noise to obtain an embodied form perception result; There are at least two environmental perception layers, which perceive the target test environment from different angles respectively; the environmental perception layers include first-level parameters and second-level parameters, the first-level parameters are used to control the perception ability and noise intensity of each perception channel, and the second-level parameters are obtained by encapsulating the corresponding first-level parameters based on preset encapsulation rules, and the preset encapsulation rules are used to limit the first-level parameters that are changed by the change of the second-level parameters to be encapsulated into corresponding second-level parameters.

2. A mobile phone-based AI digital human automatic learning method, characterized in that: include: Receive a control signal sent by a control signal interface of a mobile phone terminal; The control signal is input into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model; wherein the AI ​​digital human perception simulation model is trained based on the control training signal and the embodied form perception label corresponding to the control training signal; the AI ​​digital human perception simulation model is used to control the corresponding target intelligent body to perform corresponding actions in the target test environment according to the input control signal, and to perceive the target intelligent body after performing the corresponding actions in the target test environment to obtain the corresponding embodied form perception result, wherein the target intelligent body is selected from the virtual intelligent body previously simulated based on the embodied intelligent body form according to the control signal; Returning the embodied form perception result through the perception module interface; The AI ​​digital human perception simulation model includes a test environment simulation layer, an environment perception layer and a perception simulation layer; the control signal is input into the AI ​​digital human perception simulation model to obtain the perception result output by the AI ​​digital human perception simulation model, including: Inputting the control signal into the AI ​​digital human perception simulation model to obtain the embodied form perception result output by the AI ​​digital human perception simulation model includes: Inputting the control signal into the test environment simulation layer to control the corresponding target agent to perform corresponding actions in the target test environment according to the control signal; wherein the target test environment is a test environment that is previously selected based on the control signal and into which preset environmental noise is injected; Using the environmental perception layer, the target intelligent agent is perceived after performing a corresponding action in the target test environment to obtain a first embodied form perception result; Inputting the first embodied form perception result into the perception simulation layer, so as to process the first embodied form perception result by using preset perception noise to obtain an embodied form perception result; There are at least two environmental perception layers, which perceive the target test environment from different angles respectively; the environmental perception layers include first-level parameters and second-level parameters, the first-level parameters are used to control the perception ability and noise intensity of each perception channel, and the second-level parameters are obtained by encapsulating the corresponding first-level parameters based on preset encapsulation rules, and the preset encapsulation rules are used to limit the first-level parameters that are changed by the change of the second-level parameters to be encapsulated into corresponding second-level parameters.

3. The method for automatic learning of AI digital human based on mobile phone according to claim 2, characterized in that: The preset perception noise includes receiving noise, the receiving noise is used to limit interference in a process of inputting the first embodied form perception result into the perception simulation layer, and the receiving noise is obtained based on turning on or off at least one environment perception layer; Before inputting the first embodied form perception result into the perception simulation layer, it includes: Controlling to turn on or off at least one environment sensing layer to generate receiving noise; Inputting the first embodied form perception result into the perception simulation layer includes: Based on the received noise, a first embodied form perception result sent by the environment perception layer in the open state is received.

4. The method for automatic learning of AI digital human based on mobile phone according to claim 3 is characterized in that: The preset perception noise includes environmental perception noise, and the environmental perception noise is used to limit interference in the process of obtaining the first embodied form perception result; The embodied form perception result is processed by using preset perception noise to obtain the embodied form perception result, including: The environmental perception noise is added to the first embodied form perception result.

5. The method for automatic learning of AI digital human based on mobile phone according to claim 3, characterized in that: Returning the embodied form perception result through the perception module interface includes: Returning at least two of the embodied form perception results through the perception module interface, so that the receiving end can evaluate the quality of the at least two received embodied form perception results, and select the embodied form perception result with the highest quality evaluation result as the target form embodied form perception result; or, Return at least two of the embodied form perception results through the perception module interface, so that the receiving end can perform quality evaluation on the at least two received embodied form perception results, and based on the condition that the error of each quality evaluation result is lower than a preset threshold, average all received embodied form perception results to obtain the target form embodied form perception result; or At least two of the embodied form perception results are returned through the perception module interface, so that the receiving end can randomly select the at least two received embodied form perception results to obtain the target form embodied form perception result.

6. The method for automatic learning of AI digital human based on mobile phone according to claim 2, characterized in that: The AI ​​digital human perception simulation model also includes a control simulation layer; before the control signal is input into the test environment simulation layer, it includes: The control signal is processed based on preset control noise, wherein the control noise is configured based on interference in the process of receiving the control signal.

7. An electronic device, characterized in that: include: Memory for storing computer software programs; A processor is used to read and execute the computer software program, thereby implementing the mobile phone-based AI digital human automatic learning method as described in any one of claims 2 to 6.

8. A non-transitory computer-readable storage medium, characterized in that: The storage medium stores a computer software program, which, when executed by a processor, implements the mobile phone-based AI digital human automatic learning method as claimed in any one of claims 2 to 6.

Images