Computer automatic execution program and full-automatic system, server and method linked with 6G technology

Abstract

According to the automatic program running device of the computer, an instruction is automatically output through voice recognition of a user and is input as a program control instruction, so that the user can automatically start a required program only by voice without operating through an icon on a display screen. Therefore, the program starting time is shortened, and an accurate program execution result can be obtained. Besides, a result obtained through a program device is integrated into a virtual environment platform linked with the 6G technology, and a user can experience a more real simulation test in a non-face-to-face virtual space, so that safer purchase compared with reality is realized, and automatic service of the whole process from shopping to distribution is obtained. Particularly, by introducing various input modes, the blind, the deaf-mute and the old can conveniently use the computer, and meanwhile, the working efficiency of an enterprise can be greatly improved by utilizing a program automation system.

Description

Technical Field

[0001] The present invention relates to a computer automatic program execution device, specifically, a computer automatic execution program that utilizes artificial intelligence technology, voice recognition technology, and 6G technology to extract user-required programs and automatically run them, as well as a fully automatic program system that is linked to 6G technology. Background Art

[0002] Unless otherwise expressly stated in this specification, the contents described in this section are not prior art for the claims of this application. Even if included in this section, they are not considered prior art.

[0003] A computer system is an organic combination of hardware, software, and data. It is a collection of components that enables information processing, storage, and transmission based on computer technology. Computer systems can be designed for different purposes and applications and are widely used in a variety of fields, including information processing and storage, problem solving, communications, and entertainment. Such systems operate through the coordinated interaction between hardware and software, enabling users to efficiently utilize computers. Advances in information and communications technology (ICT) and the ubiquity of smartphones have made computing systems an integral part of daily life. Traditional computers mostly rely on users searching for the desired program or application and then directly running it to execute it. This traditional computer program execution process is particularly difficult for the visually impaired and the elderly who lack computer proficiency. As a result, people with computer difficulties often rely on others for assistance, causing significant inconvenience. Summary of the Invention

[0004] Problems to be solved by the present invention The computer automatic program execution device and program automatic system according to the embodiment are designed to extract the target program that the user wants to run from multiple programs stored in the computing system through the user's voice recognition technology, and execute the extracted program based on the voice recognition results. In a specific embodiment, the system recognizes the user's voice, uses the STT (Speech-to-Text) model to extract program execution instructions from the voice, and converts them into program control commands. This allows the user to automatically start the required program by voice alone, without having to manually operate through the icons on the display screen.

[0005] Furthermore, in this embodiment, based on 6G technology, the program output results are presented in at least one of the following formats: the metaverse, virtual reality (AR), and extended reality (XR), allowing users to experience the output more vividly. Furthermore, by integrating the output generated by the computer's automated program execution into the virtual environment platform of the 6G-based fully automated program system, simulated testing can be performed in a remote virtual space. If the user is satisfied with the results, the system provides an automated, one-stop service, from purchase selection and automatic payment to product delivery.

[0006] The technical problems to be solved by the present invention are not limited to the contents described above. Other technical problems not clearly described should be clearly understood by those skilled in the art based on the following description.

[0007] Solutions to the Problem According to an embodiment, a computer program automatically executes a program apparatus comprising a memory storing at least one instruction and a processor. The program executes the at least one instruction to implement the following functions: Converting user speech to text using a language model including an STT model; Extracting the program the user intends to use from the converted text; Detecting program control instructions from the converted text and inputting them into the program to control it. The program control integrates the output results with a virtual reality platform for a fully automated program system linked to 6G technology and projects them into a virtual space in the digital world. This allows users to immerse themselves in a more vivid experience. Because the testing environment is identical to the actual environment, it is more secure than reality. The entire system is fully automated, allowing anyone to conveniently purchase the output results. Once a purchase is made, the integrated payment server automatically performs the identity verification process to complete payment. The confirmation button automatically outputs the delivery address, providing an automated service from payment to delivery. This voice input technology allows users to easily operate the computer system using only voice commands.

[0008] Effects of the present invention The computer automatic program execution device according to the embodiments can be installed in various computer systems, such as smartphones and smart tablets. The device automatically runs user-requested programs and functions using a variety of input methods, including direct user input, voice input, gesture input, and Braille typing. This makes computer use more convenient for the visually impaired, hearing-impaired, elderly, and those with limited computer skills.

[0009] Furthermore, users no longer need to manually search for programs or enter commands on the touchscreen. The system automatically extracts operating instructions through voice, enters them into the program, and executes them. This reduces computer execution time while ensuring the accuracy of program execution results. These results are then mapped to a virtual space in the digital world using 6G technology. Users can intuitively perceive the results through augmented reality (AR), enabling tactile communication and testing, making purchases safer than in real life. Because the entire system is automated, anyone can project any object on Earth into a virtual space for realistic viewing using voice input, and easily operate the computer using only voice input. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 Schematic diagram of a fully automatic system for programs linked to 6G technology according to an embodiment.

[0011] Figure 2 A diagram showing a computer automatically executing a program.

[0012] Figure 3 Schematic diagram of a data processing architecture for a computer automatically executing a program according to an embodiment of the present invention.

[0013] Figure 4 is an example diagram illustrating modules or models stored in a memory according to an embodiment.

[0014] Figure 5 It is a schematic diagram of the data processing structure of the processor.

[0015] Figure 6 It is a schematic diagram of functions that can be executed by the program (100) through the processor (130).

[0016] Figure 7 It is a schematic diagram of a fully automatic program system that integrates a computer automatic program execution device with 6G technology.

[0017] Figure 8 Schematic diagram of the BERT model structure used in the embodiment.

[0018] Figure 9 This is an oblique view of the space stereo imaging mobile phone.

[0019] Figure 10 This is a schematic diagram of the spatial stereoscopic imaging display of a cosmic stereoscopic imaging mobile phone.

[0020] Figure 11 This is an oblique view of a brooch-shaped space stereoscopic video phone.

[0021] Figure 12 This is an oblique view of a space stereo video phone made in the form of a necklace. DETAILED DESCRIPTION

[0022] Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the attached drawings. However, regardless of the symbols in the drawings, the same or similar constituent elements are assigned the same reference numbers, and repeated descriptions thereof are omitted. In the following description, the suffixes "module" and "section" used for the constituent elements are only assigned or mixed for the purpose of facilitating the writing of the specification, and they themselves do not have meanings or functions that distinguish one from another. In addition, when describing the embodiments disclosed in this specification, if it is considered that the specific description of the relevant known technology may obscure the core gist of the embodiments disclosed in this specification, its detailed description will be omitted. In addition, the attached drawings are only used to help understand the embodiments recorded in this specification, and the technical ideas disclosed in this specification should not be limited by the drawings. It should be understood that the ideas and technical scope of the present invention include all changes, equivalents and even substitutes.

[0023] Terms containing ordinal numbers, such as "first" and "second," may be used to describe various components, but these terms do not limit the components. These terms are used solely to identify and distinguish various components.

[0024] When a component is referred to as being "connected" or "connected to" another component, it should be understood that the component can be directly connected or connected to the other component, but there may be other intermediate components between the two components. If it is explicitly stated as "directly connected" or "directly connected to", it should be understood that there are no intermediate components between the two components.

[0025] In this application, terms such as "including" or "having" should be understood as specifying the presence of features, numbers, steps, actions, components, parts or their combinations recorded in the specification; and should not be understood as pre-excluding or adding one or more other features, or pre-excluding or adding the possibility of numbers, steps, actions, components, parts or their combinations.

[0026] In this specification, "unit" includes units implemented by hardware, units implemented by software, and units implemented by a combination of the two. In addition, one unit can be implemented by two or more hardware components, and two or more units can be implemented by one hardware component.

[0027] In this specification, actions or functions described as being performed by a terminal, device, or equipment may be performed by a server connected to the terminal, device, or equipment. Similarly, actions or functions described as being performed by a server may also be performed by a terminal, device, or equipment connected to the server.

[0028] Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

[0029] Figure 1 Schematic diagram of a fully automatic program system linked to 6G technology according to an embodiment.

[0030] Reference Figure 1 According to an embodiment, a program automatic execution system using 6G technology may include a program automatic execution device (100) and a program automatic execution server (200). In an embodiment, the program automatic execution server (200) generates an application for program automatic execution and distributes it to the program automatic execution device (100); in an embodiment, the program automatic execution device (100) installs the program automatic execution application distributed from the server (200), and uses the installed application to recognize multiple applications and programs pre-installed in the automatic execution device (100) through the user's voice and action.

[0031] In an embodiment, the program automatic execution device (100) can realize a computer that can connect to a remote server or terminal via a network. Here, the computer can include, for example, a notebook computer equipped with a navigation system and a web browser, a desktop computer, a laptop computer, etc. In this case, at least one purchaser terminal (100) can realize a terminal device that can connect to a remote server or terminal via a network. At least one program automatic execution device (100) can be, for example, a wireless communication device that ensures portability and mobility. This type of device includes all types of mobile communication terminals, such as: navigation equipment, personal communication system (PCS), global system for mobile communications (GSM), personal digital cellular (PDC), personal handyphone system (PHS), personal digital assistant (PDA), International Mobile Telecommunications-2000 (IMT-2000), Code Division Multiple Access-2000 (CDMA-2000), Wideband Code Division Multiple Access (W-CDMA), wireless broadband Internet terminal (Wibro), smartphone, smart pad, tablet PC and other types of handheld wireless communication devices.

[0032] Figure 2 It is a diagram showing a computer automatically running a program.

[0033] refer to Figure 2As shown, the computer automatic execution program may include a communication unit, a memory, and a processor. The communication unit converts the received user voice into text through the STT conversion module, extracts keywords and generates control instructions. At the same time, it accumulates and learns the user's personalized program control history data to update the program automatic control model. The updated program automatic control model generates a user-customized automatic control processor. The generation unit of the processor stores program control instructions that match the recognized voice and text. The keyword matching instructions are combined through the instruction generation model to generate computer program automatic execution commands. The generated commands are input into the program, and the computer is automatically controlled to automatically execute the program. By automatically controlling the computer to automatically execute the program, the generated program outputs the results, and the corresponding model and data are output to generate at least one of the 3D metaverse, virtual reality, and augmented reality.

[0034] Figure 3 FIG. 1 is a schematic diagram of a data processing structure of a computer automatically executing a program according to an embodiment.

[0035] refer to Figure 3 As shown, the computer automatic execution program (100) according to the embodiment may include a communication module (110), a memory (120) and a processor (130).

[0036] The communication unit (110) can support a variety of communication methods such as wired and wireless, including different communication network architectures such as personal area networks (PANs) and wide area networks (WANs). In addition, the communication unit can operate based on the public World Wide Web (WWW) and support wireless transmission technologies suitable for short-range communication such as infrared (IrDA) or Bluetooth. For example, the communication unit (110) can be responsible for executing the data transmission and reception required by the technical method in the embodiment of the present disclosure. The memory (120) can refer to any type of storage medium, for example, it can include at least one type of storage medium selected from flash memory type, hard disk type, micro multimedia card type, card type memory (such as SD or XD memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, and optical disk. In this memory (120), it is possible to configure Figure 1 The user information database shown.

[0037] The memory (120) can store at least one instruction executable by the processor (130). In addition, the memory (120) can store any form of information generated or determined by the processor (130), as well as any form of information received by the program automatic execution system (100). For example, as described later, the memory (120) may store a verb evaluation table or a plurality of reference answers.

[0038] like Figure 4 As shown, there may be multiple types of modules or models stored in the memory (120).

[0039] Figure 4 is an example diagram of modules and / or models stored in memory according to an embodiment.

[0040] like Figure 4 As shown, the memory (120) may store an STT conversion module (121), a user information database (122), a keyword extraction model (123), an instruction generation model (125), an SSL / TLS security model (126), a program automatic control model (127), a 6D output generation model (128), a feedback provision model (129), and a unified payment server (139). Each module and / or model may be in the form of an application program executable by the processor (130).

[0041] Among the modules or models, the STT conversion module (121) is a module designed for speech-to-text conversion. Since the STT conversion module (121) itself is an existing public technology, its detailed description is omitted.

[0042] All models stored in the memory (120) according to the embodiment may be models trained by machine learning or deep learning. In this case, such models may be trained by transfer learning. The following briefly describes transfer learning.

[0043] like Figure 8 As shown in Figure 2, transfer learning involves transferring a pre-trained model to a specific task and fine-tuning it to obtain a model. BERT is a typical example of a pre-trained model. Since the structure and characteristics of BERT are well-known, a detailed description is omitted here.

[0044] In this embodiment, all models are pre-trained, employing language models. This type of language model is trained using a large-scale corpus through semi-supervised learning using at least one known machine learning technique, such as MLM (Masked Language Model) or NSP (Next Sentence Prediction). Since MLM and NSP are well-known technologies, detailed descriptions thereof are omitted here.

[0045] The pre-trained model is then combined with a layer for evaluating the response, and the combined model is trained. This training fine-tunes the pre-trained model to perform the task. This is why it is called fine-tuning.

[0046] During fine-tuning, multiple training data sets are used. The training input data includes answers to interview questions, while the training annotated data includes the evaluation results of the corresponding answers.

[0047] On the other hand, according to the embodiment, the learning, authentication and input data may include at least one of the user's gender, age, fingerprint, password, pattern SNS authentication, unified authentication, multiple biometric information, telephone number, address, date of birth, and name, in addition to the identity information and voice obtained from the user stored in the user information database (122).

[0048] The keyword extraction model (123) is a model designed specifically to extract keywords from answer articles. Since the technology of extracting keywords from articles is already well known, detailed description will be omitted.

[0049] The instruction generation model (125) matches the instructions corresponding to the extracted keywords. To this end, the instruction generation model (125) stores instructions for program control that match the recognized voice and text. In the embodiment, instructions refer to commands for controlling the program. In this embodiment, various codes store their own instructions, and the instruction generation model (125) generates commands for program control by combining the instructions matched by each keyword. For example, when the stored keywords are "shopping", "commodity", and "order", the instruction generation model (125) can generate corresponding commands by extracting the home shopping URL, product information, and other instructions that the user attempts to access. The SSL / TLS security protocol (126) implements an error handling mechanism for errors that may occur in communication. In addition, according to the protocol specification, the error code or status processing method is defined and adapted. According to the embodiment, the computer automatic execution program adopts security protocols such as SSL / TLS or performs additional security measures such as data encryption and authentication to ensure data security. In addition, according to the embodiment, the computer automatic execution program also tests and debugs the linkage system to find and fix potential problems.

[0050] The program automatic control model (127) implements program automatic control by inputting the generated instructions into the program. In an embodiment, the program automatic control model (127) is obtained by fine-tuning a pre-trained model. The pre-trained model is trained in a semi-supervised learning manner by applying at least one of MLM (masked language model) and NSP (next sentence prediction) on multiple corpora. In an embodiment, the pre-trained model is supervised learning multiple training data sets including training input data and training label data to obtain fine-tuning. The multiple training input data may include program extraction and program control processes for multiple user voice inputs and the user's respective program control history records.

[0051] The output object generation model (128) is to generate the output of the program automatic control result into an artificial neural network model based on 6G technology 5-dimensional output. To this end, the output object generation model (128) stores the program output result as a model and data for generating at least one of the 5D-based metaverse, virtual reality (VR), augmented reality (AR), and mixed reality (MR).

[0052] The feedback model (129) refers to: an artificial neural network model that evaluates and updates the artificial neural network model and the deep learning model after learning. In an embodiment, the feedback model (129) can evaluate the artificial neural network model through at least one of accuracy, precision and recall. Accuracy is an indicator that measures the degree of match between the predicted results of the artificial neural network model and the actual results. Precision is an indicator that measures the proportion of actual positive results among the predicted positive results. Recall is a detection indicator of the proportion of actual positive results predicted by the model. In an embodiment, the feedback model (129) can calculate the accuracy, precision and recall of the artificial neural network model, and then evaluate the artificial neural network model based on at least one of the calculated indicators.

[0053] In an embodiment, the feedback model (129) can detect the accuracy of the artificial neural network model by using an evaluation dataset.

[0054] The evaluation dataset consists of data not used in model training and is used to objectively evaluate the performance of the model.

[0055] In an embodiment, the feedback model (129) uses the evaluation data set to run the artificial neural network model and compares the predicted value of the artificial neural network model for each input data with the actual true value of the corresponding data.

[0056] The accuracy of the model predictions can then be tested by comparing the results. For example, in a feedback model (129), the accuracy can be calculated by the proportion of data that the model correctly predicts.

[0057] In addition, the feedback model (129) calculates a precision index calculated as the harmonic mean of precision and recall, and an F1 score (F1 Score) index indicating a balance between precision and recall, and evaluates the artificial neural network model based on the calculated F1 score, generating an AuC-RoC curve that visualizes the performance of the classification model. Based on the generated AuC-RoC curve, the artificial neural network model can be evaluated. In an embodiment, when the area under the RoC curve (AuC) is closer to 1, it can be evaluated that the performance of the feedback model (129) is better.

[0058] In addition, the feedback model (129) can evaluate the interpretability of the artificial neural network model. In an embodiment, the feedback model (129) evaluates the interpretability of the artificial neural network model using the SHAP (SHapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explanations) methods. SHAP (SHapley Additive Explanations) is a library that provides explanations for model prediction results, and the feedback model (129) extracts SHAP values ​​from the library. In an embodiment, the feedback model (129) can predict the degree of influence of the feature information input to the model on the prediction results by extracting the SHAP values.

[0059] The LIME (Local Interpretable Model-agnostic Explanations) method is a method for explaining the prediction results of a model for individual samples. In an embodiment, the feedback model (129) approximates the sample to an interpretable model through the LIME method and calculates the importance of each feature information. In addition, the feedback model (129) can evaluate the influence of each feature variable by analyzing the internal weight value and bias value. The feedback model (129) will perform improvement operations when the fairness of the artificial neural network model is low or discrimination occurs. In an embodiment, when the data of a specific group is lower than a certain level, the feedback model (129) will additionally collect data representing the specific group and perform a data preprocessing process. In an embodiment, the feedback model (129) prevents the model from learning unnecessary patterns through a data preprocessing process including data normalization, outlier removal and data ratio adjustment. In addition, by adding specific conditions to the learning algorithm, discrimination can be prevented or fairness can be guaranteed.

[0060] In order to ensure fairness, the feedback model (129) compares the model's predicted results with the actual results through confusion matrix analysis to evaluate the model's performance. The confusion matrix is ​​a matrix used to evaluate the model's classification performance in guided learning. The confusion matrix displays the classification results by comparing the model's predicted results with the actual results. The feedback model (129) can calculate the accuracy and misclassification rate of each classification through confusion matrix analysis, thereby evaluating the model's performance. In addition, in an embodiment, the feedback model (129) can confirm the data distribution through visual analysis of the training data. For example, for image data, the diversity and fairness of the data can be evaluated through visual processing of image samples corresponding to each category. In addition, the feedback model (129) verifies the fairness and diversity of the learning data through fairness verification and evaluation index calculation, thereby improving the artificial neural network model. Fairness verification refers to confirming whether the learning information and the artificial neural network model show differences in specific attributes of the data. The feedback model (129) can confirm the difference of a specific attribute by comparing the number of samples corresponding to each attribute or by evaluating the classification performance of each attribute. In addition, the feedback model (129) calculates a variety of indicators to evaluate the performance of the artificial neural network model. For example, the performance of the model can be evaluated by calculating indicators such as accuracy, precision, recall rate, F1 score, etc. At this time, by calculating the indicators of each category, the fairness and diversity of the model can be evaluated. In addition, the feedback model (129) will collect feedback on problems encountered by the artificial neural network model in the actual application environment, and reflect the collected feedback information to the artificial neural network model, thereby continuously optimizing the model. On the other hand, the aforementioned models may all be inferred models obtained through artificial neural network learning. Next, we will briefly explain these inferred models.

[0061] In this specification, the inferred model may refer to any form of computer program that runs based on a network function, an artificial neural network and / or a neural network. In this specification, the terms model, neural network, network function and neural network (neural network) may be used interchangeably. A neural network consists of one or more nodes, interconnected by one or more connections, thereby forming a relationship between input nodes and output nodes within the neural network. The number of nodes and links in a neural network, the relationship between nodes and connections, and the weight value assigned to each connection may determine the characteristics of the neural network. A neural network can be composed of a collection of one or more nodes. A subset of the nodes that make up a neural network can constitute a layer.

[0062] A deep neural network (DNN) refers to a neural network that contains multiple hidden layers in addition to an input layer and an output layer. Deep neural networks may include convolutional neural networks (CNN), recurrent neural networks (RNN), autoencoders, generative adversarial networks (GAN), restricted Boltzmann machines (RBM), deep belief networks (DBN), Q networks, U networks, twin networks, generative adversarial networks (GAN), transformers, and the like. The aforementioned description of deep neural networks is merely an example, and the present disclosure is not limited thereto. Neural networks can learn through at least one of supervised learning, unsupervised learning, semi-supervised learning, self-supervised learning, or reinforcement learning. Learning a neural network can be described as the process of applying the knowledge required to perform a specific operation to the neural network.

[0063] Neural networks can learn by minimizing output error. This involves repeatedly feeding training data into the network and calculating the error between the network's output and the target. To minimize this error, the error is backpropagated from the output layer to the input layer, updating the weights of each node in the network.

[0064] In supervised learning, all learning (training) data uses data labeled with the correct answer (labeled data); in unsupervised learning, data without correct answer labels (unlabeled data) can be used. The amount of change in the connection weights of each node is updated, and the learning rate determines the change. The neural network's calculations on the input data and the backpropagation of errors constitute a learning cycle (epoch). The learning rate can be varied depending on the number of repetitions in the neural network learning cycle. Furthermore, to prevent overfitting, methods such as increasing the learning data, regularization, dropout techniques (randomly discarding some nodes), and batch normalization layers can be used.

[0065] In one embodiment, the inference model can leverage at least a portion of a transformer. A Transformer model can consist of an encoder and a decoder. The encoder encodes the embedded data, while the decoder decodes the encoded data. A Transformer can receive a sequence of data and, through encoding and decoding stages, output a sequence of data of different types. Its architecture can accommodate this capability. In one embodiment, the sequence of data can be processed into a form that the Transformer can operate on. The process of converting a sequence of data into a form that the Transformer can process may include embedding. Terms such as data tokens, embedding vectors, and embedding tokens may refer to data embedded in a form that the Transformer model can process. To encode and decode a sequence of data, the Transformer can utilize an attention algorithm to process the encoder and decoder within the Transformer. An attention algorithm calculates the similarity between a given query and one or more keys, then applies these similarities to the values ​​associated with the keys, and takes a weighted sum of the similarity values ​​to calculate an attention value. Different types of attention algorithms can be distinguished based on how the query, key, and value are set. For example, if the query, key, and value are all set to be the same when calculating attention, this may indicate a self-attention algorithm. When attention is calculated by computing attention heads for each segmented embedding vector, in order to parallelize a series of input data and reduce the dimensionality of the embedding vector, this may indicate a "multi-head attention" algorithm. A Transformer can be composed of multiple modules that implement the Multi-Head Self-Attention Algorithm or the Multi-Head Encoder-Decoder Algorithm. In one embodiment, in addition to the attention algorithm, the Transformer model may also include additional components such as embedding layers, normalization layers, and a softmax function.Methods for constructing a Transformer architecture using an attention algorithm may include the method disclosed in Vaswani et al., "Attention Is All You Need," 2017 NIPS, which is incorporated herein by reference. The Transformer model can be applied to various data domains, such as embedded natural language, segmented image data, and audio waveforms, to convert a series of input data into a set of output data. To convert data with various data domains into a series of data that can be input to the Transformer model, the Transformer can embed data. The Transformer model can process additional data that represents the relative positional or topological relationships between the series of input data. Alternatively, the series of input data can be embedded by appending vectors that can express the relative positional or phase relationships between the series of input data. In one embodiment, the relative positional relationships between the series of input data may include, but are not limited to, word order within a natural language sentence, relative positional relationships within a segmented image, or the temporal order of segmented audio waveforms. The process of appending information expressing the relative positional or phase relationships between the series of input data can be referred to as positional encoding.

[0066] In one embodiment, the inferred model may include models such as RNN (recurrent neural network), LSTM (long short-term memory network), BERT (Bidirectional Encoder Representation based on Transformer), or GPT (Generative Pre-trained Transformer). In one embodiment, the inferred model may be a model trained by transfer learning. Here, transfer learning refers to a machine learning method that uses semi-supervised learning or self-supervised learning to pre-train with large-scale unlabeled training data to obtain a pre-trained model for a first task. The pre-trained model is then fine-tuned to adapt to a second task, and trained using labeled training data in a guided learning manner to achieve a target model.

[0067] Reference again Figure 3, let's take a look at the processor (130). First, according to one embodiment, the processor (130) can implement the technical features of the embodiments of the present disclosure described later by executing at least one instruction stored in the memory (120). In one embodiment, the processor (130) can be composed of at least one core and can include a central processing unit (CPU) for automatically executing the computer program (100), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), and other processors for data analysis and / or processing.

[0068] Reference below Figure 5 , illustrating data processing in the processor (130).

[0069] Figure 5 FIG. 1 is a schematic diagram of a data processing structure of a processor according to an embodiment.

[0070] Reference Figure 5 According to the embodiment, the processor 130 may be composed of a collection unit (131), a preprocessing unit (132), a learning unit (133), a generation unit (134), an automatic control unit (135), an output generation unit (136), a feedback unit (137) and an evaluation unit (140). The term "unit" used in this specification should be interpreted as including software, hardware or a combination thereof, depending on the context in which the term is used. For example, software can be machine language, firmware, embedded code and application software. In another example, hardware can be a circuit, a processor, a computer, an integrated circuit, an integrated circuit core, a sensor, a micro-electro-mechanical system (MEMS), a passive device or a combination thereof.

[0071] The collection unit (131) is responsible for collecting learning data of the deep learning model and user information for automatic program execution. The user information may include user voice and action information, program control history information of each user, and a user information database.

[0072] The preprocessing unit (132) preprocesses the collected artificial intelligence learning data to remove biased or discriminatory data.

[0073] The preprocessing unit (132) preprocesses the collected data into a form suitable for artificial intelligence model learning. For example, the preprocessing unit (132) can perform processing processes such as noise elimination, outlier removal, and missing value processing. In addition, the preprocessing unit (132) can effectively prevent the model from learning unnecessary patterns by performing data standardization, outlier removal, or data scaling through data preprocessing.

[0074] The learning unit (133) uses the pre-processed training data to train the artificial neural network model. The learning unit (133) selects algorithms suitable for the required fields of different industries or positions by analyzing the industries, positions and metadata, and detects the optimal hyper parameters in the selected algorithms. In the embodiment, hyper parameters refer to parameters that can be adjusted during the training process of the artificial intelligence model. Hyper parameters may include, but are not limited to, the learning rate, batch size, and number of training cycles (epochs) of the artificial neural network model.

[0075] The learning unit (133) extracts the keywords contained in each industry or position as metadata, and compares the extracted metadata with the keywords describing the algorithm. According to the comparison results, an algorithm with a similarity above a certain level can be selected. In addition, the learning unit (133) uses input data and correct answer data to train the artificial neural network model. In an embodiment, the learning unit (133) inputs the input data into the model, and the model outputs the predicted value of the input data. Afterwards, the learning unit (133) calculates the prediction error by comparing the predicted value with the correct answer data. The learning unit (133) minimizes the prediction error by adjusting the weight value (weight) and bias (bias) of the model, thereby advancing the learning process. In this process, the speed and accuracy of the model training can be controlled by adjusting hyperparameters including the learning rate. The learning unit (133) repeatedly executes the aforementioned artificial neural network model training process so that the model can learn the relationship between the input data and the correct answer data. The learning unit can perform hyperparameter tuning to detect the lowest value of the hyperparameter through experiments and verification. In addition, the learning unit samples the collected training data and learns the sampled data. "Sampling" refers to the process of extracting part of the data from the data set. In the embodiment, when the data set reaches a certain capacity and has diversity, part of the data is extracted for model training by a sampling method. By doing this, the amount of data required for model learning can be reduced and the learning speed can be improved. In addition, when the data set is unbalanced, the balance between categories can be adjusted by the sampling method. In addition, the learning unit generates training data through data augmentation technology and uses the generated data to train the artificial neural network model. Data augmentation refers to the technology of generating new data by deforming the collected training data. For example, the learning unit (133) generates new image data by applying left-right flipping, rotation, size adjustment, color transformation, etc. to the image data in the collected training data, and uses the newly generated image data as training data. In this way, in this embodiment, by expanding the scale of the data set, the model performance can be improved in diverse scenarios. In the embodiment, data augmentation can be used not only for image data, but also for text data. For example, the learning unit (132) can generate new learning data by applying word replacement, synonym replacement, sentence inversion and other methods to text data.

[0076] The generation unit (134) matches the instructions corresponding to the extracted keywords. To this end, the generation unit (134) stores program control instructions that match the recognized voice and text. Instructions are command statements used to control the program. In this embodiment, each code stores corresponding instructions. The generation unit (134) matches the keywords through the instruction generation model and combines the instructions to generate command statements for program control. For example, when the stored keywords are "TV shopping", "product", and "order", the generation module (134) can extract the TV shopping URL, product information, and other instructions that the user wants to access to generate corresponding commands.

[0077] The automatic control unit (135) automatically controls the program by inputting the generated instructions into the program. In subsequent embodiments, the output object generation unit (136) converts the program output results into at least one of a 3D-based metaverse, virtual reality, and augmented reality, and outputs the model and data required to generate the form. The output object generation unit (136) generates a model and data to be output in a virtual environment based on the program execution results. The model can include elements such as 3D objects, environments, and characters, and its data can include texture animation, physical properties, and other content. Thereafter, the output object generation unit (136) determines what type of output object should be generated by analyzing the program execution results. For example, when the program outputs a simulation result, the simulation object and its environment will be projected into the virtual environment. Subsequently, the output generation module (136) will use 3D modeling tools or design software based on the generated model and data to construct and design a 3D model for practical application in a virtual reality environment. At this time, the shape, color, material, etc. of the model can be defined, and animation can be added. Afterwards, the output generation unit (136) will integrate the metaverse, virtual reality, and augmented reality platforms: the generated 3D model and data will be integrated into the selected virtual environment platform (metaverse / virtual reality / augmented reality). Each platform may have its own unique format and technology, so the model and data need to be converted or formatted according to the specifications of the platform. Afterwards, the output generation unit (136) will output the generated model and data into the selected virtual environment. In this way, the user can view or interact with the 3D model. Thus, the output generated by the output generation unit (136) can be operated and interacted with by the user. When the user moves or operates the model in the virtual environment, the output generation module will perceive and reflect this operation.

[0078] The feedback unit (137) evaluates the trained artificial neural network model through the feedback model. In an embodiment, the feedback unit (137) can evaluate the artificial neural network model through at least one of accuracy, precision and recall. Accuracy is an indicator that measures the degree of consistency between the predicted results of the artificial neural network model and the actual results. Precision is an evaluation indicator used to measure the proportion of results predicted to be positive that are actually positive. Recall is an indicator that measures the proportion of actual positive samples correctly predicted by the model. In an embodiment, the feedback unit (137) can calculate the accuracy, precision and recall of the artificial neural network model, and evaluate the artificial neural network model based on at least one of the calculated indicators. The feedback unit (137) can detect the accuracy of the artificial neural network model by using an evaluation data set. The evaluation data set is composed of data not used in model training and is used to objectively evaluate model performance.

[0079] In an embodiment, the evaluation unit (140) runs the artificial neural network model using the evaluation data set, and compares and analyzes the predicted value of each input data by the artificial neural network model with the actual true value of the corresponding data. The accuracy of the model prediction can be measured by the subsequent comparison results. For example, in the evaluation unit (140), the accuracy can be calculated by the proportion of data correctly predicted by the model to the total data. In addition, the feedback (137) calculates the F1 score, which is an indicator reflecting the balance between precision and recall, and is obtained by calculating the harmonic mean of the two. Based on the calculated F1 score, the artificial neural network model is evaluated, and a visual graphical indicator of the classification model performance, namely the AUC-ROC curve, is generated. The artificial neural network model can be further evaluated by the generated AUC-ROC curve. In an embodiment, the evaluation unit (140) can evaluate the model performance by the area under the ROC curve (AUC). The closer the AUC value is to 1, the better the model performance. In addition, the feedback module (137) can evaluate the interpretability of the artificial neural network model. The feedback unit (137) evaluates the interpretability of the artificial neural network model through methods such as SHAP and LIME. SHAP is a library for providing interpretability analysis of model prediction results, and the evaluation unit (140) extracts SHAP values ​​from the library. By extracting the SHAP values, the evaluation unit (140) can predict the degree of influence of the feature information of the input model on the model prediction. The LIME method is a method for interpreting the prediction of a single sample by the model. The evaluation unit (140) approximates the sample to an interpretable model through the LIME method, thereby calculating the importance of each feature information. In addition, the evaluation unit can also infer the influence of each feature variable by analyzing the internal weight value and bias value of the model.

[0080] Figure 6It is a functional diagram of the computer automatically executing the program (100) through the processor (130).

[0081] Figure 6 This is only an illustrative description and the concept of the present invention should not be construed as being limited to Figure 6 shown.

[0082] refer to Figure 6 As shown, in step S100, a language model including an STT (speech-to-text) model converts user speech into text. In the next step, the program the user intends to use is extracted from the converted text. In step S200, instructions for program control are detected from the converted text. In step S300, the program is controlled by inputting these instructions into the program. In step S400, the output of program control is used to generate virtual reality, augmented reality, and the metaverse based on 6G technology. In this embodiment, an STT (speech-to-text) conversion module is stored in memory. The STT conversion module converts user voice control commands into text, extracts keywords, and generates corresponding instructions. In this embodiment, individual user program control history records can be accumulated and learned to update the program automatic control model. The updated program automatic control model is then used to generate an automatic control flow tailored to each user. The program automatic control model is obtained by fine-tuning a pre-trained model. The pre-trained model is trained using semi-supervised learning on multiple corpora by applying at least one of the following: a masked language model (MLM) and next sentence prediction (NSP). Fine-tuning is the process of performing supervised learning on the pre-trained model using a training dataset consisting of multiple training input data and corresponding annotated data.

[0083] At this time, fine-tuning refers to the process of guiding the learning of a pre-trained model using a training data set containing multiple training input data and training label data to obtain an adjusted model. The multiple learning input data include program extraction and program control processes for multiple user voice inputs, as well as program control history records for each user. The memory stores program control instructions that match the recognized voice and text, and also stores the models and data required to generate the program output results based on 6D technology to at least one of the metaverse, virtual reality, and augmented reality. Below, combined with Figure 7 As shown, the application methods and changes of 5D-based models and data stored in the memory in the program automation system after access to 6G technology are explained.

[0084] Figure 7 A schematic diagram showing a fully automatic system for programs linked to 6G technology.

[0085] The fully automated program system linked with 6G technology is a method for applying 6G technology to the output of the above program (100). 6G is an intelligent communication infrastructure that connects virtual and real life without time and space restrictions by improving 5G performance, optimizing networks based on artificial intelligence, and expanding coverage at sea, in the air, and in the universe. In the 6G era, the network will integrate ground communication, satellite communication, and marine communication to achieve signal coverage in communication blind spots such as deserts, uninhabited areas, and oceans. It will be widely commercialized in the fields of space communication, intelligent interconnection, emotional and tactile interaction, multi-sensory mixed reality, equipment collaboration, and fully automatic transportation.

[0086] refer to Figure 7 As shown, the 6G technology-linked fully automatic program system may include a computer automatic program execution device (100), a program automatic execution server (200), a virtual environment platform (600), and an integrated payment server (BRICS PAY BRICS International Integrated Payment Server).

[0087] In the 6G technology-linked fully automatic program system, the program automatic execution server (200) may store websites, software, neutrino data sets, space stereo imaging systems, and various AI navigation systems.

[0088] For example, there may be a built-in shopping guide function for the blind (to address the problem that blind people cannot view the many products in shopping malls, by providing artificial intelligence functions in the shopping platform, the color, shape, performance, material and other attributes of the products can be converted into voice or Braille descriptions to assist the visually impaired group to complete the purchase of products); it is also possible to develop an AI deaf-mute dialogue assistance system (an artificial intelligence deaf-mute dialogue commentator that converts the text or sign language input by the deaf-mute person into voice and text and transmits it to the communication partner, and at the same time converts the other party's voice information into text or sign language and transmits it to the deaf-mute user).

[0089] AI painters who can automatically draw various types of paintings (oil paintings, watercolors, comics, etc.) may be built in, as well as AI consultants who can automatically provide consulting services, AI designers who can automatically complete various designs, AI directors who can automatically shoot and produce TV series by inputting stories, AI Chinese medicine practitioners who can automatically diagnose and treat pulses, and AI teachers who can teach various subjects. In addition, each keycap of a computer keyboard can be replaced with an independent musical instrument module, and a single keyboard can be turned into an AI music band by building in a variety of musical instruments. When a musical score is input, the system can automatically play the music. The program automatic execution server (200) can also be equipped with various AI navigation function modules such as AI composer, AI lyricist, AI nurse, AI lawyer, etc., which have been invented and built-in functions.

[0090] In an embodiment, a program automatic execution server (200) generates an application program for automatic program execution and distributes it to a computer automatic program execution device (100). The computer automatic program execution device (100) in the embodiment installs the program application program distributed from the server (200) and automatically executes a plurality of pre-installed applications and programs in the device (100) by using the installed application program and through user voice and motion recognition.

[0091] The program device (100) for realizing automatic execution through user speech and action recognition converts the user speech into text using a language model including an STT conversion model, and extracts the program that the user wants to use from the converted text. In addition, the program automatic execution server (200) detects the converted text instructions for program control. The detected instructions are input into the automatic control program (100) to control the program, and the results output by the program control are used to generate the required model and data in the 6G virtual reality and store them in the memory.

[0092] The processor can implement the technical features of the present disclosure according to the embodiment by executing at least one instruction stored in the memory.

[0093] In the above-mentioned processor, the acquisition unit is responsible for collecting learning data of the deep learning model and user information for automatic program execution.

[0094] The preprocessing department will pre-process the collected learning data to remove biased or differential data in the artificial intelligence learning data.

[0095] The Learning Department trains the artificial neural network model using pre-processed learning data. Based on the characteristics of each industry or job demand area, the department selects an appropriate algorithm by analyzing industry and job metadata. Based on this, the optimal hyperparameters are determined for the selected algorithm, and the artificial neural network model training process is completed using the input data and correct answer data.

[0096] In addition, the collected learning data is sampled and the sampled data is used for learning. Sampling is the process of extracting part of the data from the dataset.

[0097] Data augmentation is a method of generating new data by deforming the collected training data and using the generated data to train the artificial neural network model.

[0098] The learning unit uses the collected learning data to generate new image data and uses this new image data as training data. This allows the implementation to expand the size of the dataset and improve the model's performance in different scenarios.

[0099] The generation unit matches the extracted keywords with corresponding instructions. To this end, it stores program control instructions that match the recognized speech and text. In the embodiments, instructions refer to commands used to control the program. In the embodiments, corresponding instructions are stored in different code sections. The generation unit uses an instruction generation model to combine the instructions that match each keyword to generate the commands required to control the program.

[0100] The control unit automatically controls the program by inputting generated instructions into it. The generation unit outputs the program results as models and data based on 6D technology, used to generate at least one of the following: cosmic imaging, metaverse, virtual reality, or augmented reality. Based on the program's execution results, the output object generation unit generates models and data to be output to the virtual environment. These models can include model-based and model-free models, CAD models, and seen / unseen models; while data can include object perception levels (OBJVLS), large language models (LLMs), and diffusion models. The output object generation unit then analyzes the program's execution results to determine the type of output to be generated. For example, when the program outputs simulation results, the simulated objects and environment are projected into the virtual environment. The output object generation unit then uses BEXEL software to design the actual 6D model for the virtual environment based on the generated model and data. The model's shape, color, material, and other details can be customized, and animation can be added. In this embodiment, the output generation unit is then integrated into the cosmic imaging, metaverse, virtual reality, and augmented reality platforms: into the selected virtual environment platform (cosmic imaging, metaverse, augmented reality) for the generated 6D model and data. Each platform may have its own formats and technologies, so the model and data must be converted or formatted according to the platform's specifications. The output generation unit outputs the generated model and data into the selected virtual environment. Through this, users can view and interact with the 6D model. In addition, within the metaverse, users can share and interact with other users. Furthermore, the output content generated by the output generation unit is user-operable and interactive. As the user moves or operates in the virtual environment, the output generation unit not only detects and reflects the linked content, but also provides real-time collection and sharing of 6G sensor data and location information. This technology accurately determines the user's location and surrounding environment, and provides appropriate virtual information based on the user's movement. Furthermore, a system can be constructed using augmented reality devices or glasses to project virtual information into the user's field of view. This system can arrange and manipulate virtual information based on the user's location and line of sight.

[0101] To this end, the 6G-enabled fully automated program system, when in operation, visually presents relevant information to the user through the screen of the augmented reality device. This interactive presentation can be achieved by overlaying the relevant information in the user's field of view or adjusting its position.

[0102] Furthermore, in fully automated program systems linked to 6G technology, users can operate and control programs in an augmented reality environment through gestures or voice commands, thanks to the high interactivity supported by 6G technology. This approach provides an experience tailored to user convenience and the user experience. For example, in a virtual space, such as a company projected within the digital world, all employees can remotely adjust computer tasks without meeting in person. Even from a distance, a single person can simultaneously operate multiple computers, significantly improving system control efficiency. Furthermore, when purchasing a vehicle using an embodiment, the vehicle's driving interface can be fully replicated in a non-face-to-face virtual space, allowing users to realistically experience the vehicle's driving state and enabling fully autonomous driving through the embodiment. Furthermore, using an embodiment, the vehicle to be purchased in augmented reality (AR) can be converted into text. This allows users to visualize and interact with the vehicle through a 6D model. In the metaverse scenario, users can not only share with other users but also directly manipulate and interact with the output (vehicle) in the virtual space. When the model is moved or manipulated in the virtual environment, the generation unit perceives and reflects these activities in real time. Specifically, when images are moved or deformed in a virtual space to resemble real objects, or in a virtual environment that resembles reality, AI capabilities are essential to selecting objects that are nearly 100% accurate. To achieve this, the artificial neural network model learned through the feedback model is evaluated using at least one of accuracy, precision, and recall.

[0103] Accuracy is a metric that measures the degree to which the predictions of an artificial neural network model agree with the actual results. Precision is a metric that measures the proportion of predicted positive results that are actually positive. Recall is a metric that measures the proportion of actual positive results that are predicted by the model.

[0104] The feedback unit uses the evaluation dataset to test the accuracy of the artificial neural network model. It also evaluates the interpretability of the artificial neural network model. Using the LIME (Local Interpretable Model-agnostic Explanations) method, the evaluation unit approximates the sample as an interpretable model and calculates the importance of each feature. Furthermore, by analyzing the model's internal weights and biases, the evaluation unit can estimate the influence of each feature variable. By converting the extracted output (vehicle) into text, users can experience the same experience as when testing a vehicle before planning to purchase it.

[0105] After selecting the desired product through the test and clicking the purchase button (141), the computer automatically executes the BRICS PAY international integrated payment server stored in the program (100) and automatically connects to the user information database (122), completes the authentication process and makes the payment. After the purchase is completed (143), the computer automatically connects to the user information database (122), obtains the address, name, phone number and other data from the user's identity information, and automatically delivers the target product (output vehicle) to the delivery address. After the purchase is completed (143), the computer automatically logs out of the account on the product (output vehicle) purchase website to maintain a secure state.

[0106] After the purchase is completed (143), the user information database (122) is switched to a dormant mode. The user information database (122) will automatically resume power supply after the user is authenticated (such as face recognition) by the unified payment server, thereby safely protecting the user's identity information.

[0107] Furthermore, where data security is required, the program (100) embodies an error handling mechanism, applies security protocols such as SSL / TLS, or performs additional measures such as data encryption and authentication, thereby maintaining the complete security state of the integrated payment server (BRICS PAY International Integrated Payment Server).

[0108] (BRICS PAY International Integrated Payment Server can play the same role as an integrated payment server and is adopted in this system for the common economic development and stable currency circulation of countries around the world.) Figure 9 This is an oblique view of the space stereo imaging mobile phone.

[0109] Reference Figure 9 As shown, the space stereoscopic image mobile phone connected to 6G technology is installed with the computer automatic execution program device (100) of the present invention on the "iPhone" body (refer to Figure 9 -a), the phone can be activated by voice and gesture input without touch screen input. In addition, to address the most prominent battery problem of the "iPhone", the battery is removed and replaced with a "nuclear battery" with excellent performance, smaller than a coin and with a lifespan of 50 years, which greatly reduces the size of the phone and allows it to be designed as a brooch, necklace, bracelet, or ring (see Figure 11 、 Figure 12 This design eliminates the risk of battery explosion and further improves performance, making it convenient for the deaf, blind, elderly, and children to use while retaining the original display function.

[0110] And on the back of the phone (refer to Figure 9-b) Install an LED lens with a diameter of about 3cm, and send the image signal to the control tablet computer through wired / wireless means, and the image is sent out by the LED lens. The screen light emitted by each module passes through the outer surface of the LED lens ( Figure 9 -c lens side view) installed "photoresist projection system" for projection (refer to Figure 10 Specifically, light rays a1, a2, a3, and a4 emitted by any module on the LED lens are projected onto the center point P of the incident aperture. They are then reflected from the center point P' of the exit aperture by the "photoresist projection system." In image space, the reflected light rays a1", a2", a3", and a4" intersect the spatial imaging plane A'B' at points a1', a2', a3', and a4', generating a three-dimensional spatial image. A space-based three-dimensional imaging phone capable of space travel has also been proposed.

[0111] In addition, computers automatically run programs to test and debug linkage systems, discovering and correcting potential problems.

[0112] Testing can be performed in various scenarios using simulators or simulated environments. Furthermore, according to embodiments, a computer-automated program can provide program output in an augmented reality (AR) environment using 6G technology. 6G technology offers ultra-high-speed data transmission and ultra-low latency communications, enabling rapid transmission and reception of large amounts of data. The computer-automated program utilizes this technology to rapidly receive the data, images, videos, data streams, and other information required for program execution. Furthermore, because 6G technology provides the ability to collect and share sensor data and location information in real time, the computer-automated program can accurately grasp the user's location and surrounding environment and provide appropriate virtual information based on the user's movements. Furthermore, the computer-automated program also provides linkage capabilities. For example, a system can project virtual information into the user's field of view through augmented reality devices or glasses. In this way, the fully automated system can arrange and manipulate virtual information based on the user's location and line of sight. To this end, the fully automated program system analyzes the surrounding environment in real time to detect physical objects or images around the user. Based on this, it automatically identifies and executes virtual information or programs related to these objects or images. When the program is running, the fully automated system will visually display relevant information to the user through the augmented reality device's screen. This information can be overlaid on the user's field of view or adjusted within the screen to support user interaction. Furthermore, the 6G technology used in the fully automated system supports high interactivity, allowing users to program and control programs in an augmented reality environment through gestures or voice commands. This approach provides users with an experience that is both convenient and environmentally friendly. For example, in a virtual space, such as a company mapped onto the digital world, all employees can remotely adjust computer operations without meeting face-to-face. Even from a distance, one person can control multiple computers, significantly improving system control efficiency. Furthermore, through embodiments, when purchasing a vehicle, the vehicle's driving interface can be fully reproduced in a non-face-to-face virtual space, allowing users to experience the driving state immersively. This embodiment also enables fully autonomous driving. Furthermore, the user terminal device can fully reproduce all vehicle interior interfaces in a non-face-to-face virtual space. In particular, fully autonomous driving, achieved through augmented reality technology, can be achieved through mobile phones in a virtual space projected onto the digital world, allowing users to experience the sensation of actually riding in the vehicle they desire. This allows them to test the desired vehicle just as they would in real life. Specifically, the user's desired voice command is automatically executed by the computer through multiple modules within the program, including the learning module, generation module, automatic control module, output generation module, feedback module, and evaluation module. The resulting output (car) is the closest to the user's ideal expectations. Through these processes, the output is customized based on the user's preferences, material, color, and other factors, ensuring that the output (car) is the closest to the user's desired voice command.If the output (vehicle) is the desired product for the test, selecting the purchase button automatically connects the system to the payment server integrated with the computer's automated program execution system. Specifically, when the purchase button is selected, the output generation module detects and responds accordingly. Upon receiving the output generation module's response, the automated control unit automatically performs an identity authentication process through the unified payment server based on the user's identity information stored in the user information database, ultimately completing the payment. Once payment is complete, the "Complete" button automatically exits the output (vehicle) purchase website to maintain security.

[0113] Furthermore, the completion button is detected and reflected by the output creation unit. After receiving the detection and feedback from the output creation unit, the automatic control unit automatically connects to the user information database. Based on the user's identity information, the automatic delivery system receives the user's delivery address, phone number, name, and other information and automatically delivers the output (car) to the designated address.

[0114] Furthermore, the automatic control button switches the user information database to power saving mode while acquiring the delivery address input from the user information database. The user information database automatically connects to the power supply only when the user is authenticated (facial recognition), thus completely protecting the user's identity.

[0115] In addition, the 6G technology-linked fully automatic program system of the embodiment introduces a new electromagnetic wave blocking technology (i.e., aluminum (Al) is plated on the thinnest plastic, a layer of plastic is attached to the surface, and then a second aluminum (Al) plating process is performed. An extremely thin plastic film is then attached to the surface, and a third aluminum (Al) plating process is performed on top of it, and an ultra-thin plastic film is attached on top of it. Thus, aluminum (Al) foil for electromagnetic shielding is produced. This aluminum foil (Al) material for electromagnetic wave shielding can block wind, rain, cold, and ultraviolet rays outdoors, making it suitable for use as a protective layer. Using this aluminum (Al) foil specifically for electromagnetic wave shielding, the generator of the terminal device can be sealed to achieve electromagnetic wave shielding. A 6G technology-linked fully automatic program system including the above-mentioned thin aluminum foil that can achieve remote wireless charging of the terminal.

[0116] Furthermore, in this embodiment, by applying the blind typing function to a fully automated program system, blind people or ordinary users can conveniently perform automated searches and use computers. Furthermore, in this embodiment, keyboard input or dynamic input via sign language by a deaf-mute person can be converted into speech by the fully automated program system and transmitted to the other party. Furthermore, the other party's speech can also be converted into text or sign language and fed back to the deaf-mute person. This enables barrier-free daily life and free communication for the deaf-mute.

[0117] like Figure 8 As shown, BERT is an embodiment of a pre-trained model. Its specific structure and operation can be referred to Figure 4 Description.

[0118] Those skilled in the art will understand that for each embodiment disclosed herein, the various exemplary logic blocks, modules, processors, means, circuits, and algorithm steps described herein may be implemented by various forms of electronic hardware (referred to herein as software for ease of description), programs or design codes, or a combination of the two. In order to clearly illustrate this compatibility of hardware and software, the components, functional blocks, modules, circuits, and operational steps of various embodiments have been generally described above based on their functions. Whether these functions are implemented in the form of hardware or software depends on the specific application and the design constraints imposed on the entire system. In the technical field disclosed by the present invention, those skilled in the art may implement the described functions in a variety of ways for specific applications, but these implementation decisions should not be interpreted as departing from the scope of the claims of the present invention.

[0119] The various embodiments provided herein may be implemented in the form of methods, systems, or articles of manufacture implemented using standard programming and / or engineering techniques. The term article of manufacture includes computer programs, carriers, or media that can be accessed from any computer-readable storage system. For example, computer-readable storage media include, but are not limited to, magnetic storage systems (e.g., hard disks, floppy disks, magnetic strips, etc.), optical disks (e.g., CDs, DVDs, OPP DISCs, etc.), smart cards, and flash memory systems (e.g., EEPROMs, cards, USB flash drives, key drives, etc.). In addition, the various storage media provided herein include one or more systems and / or other machine-readable media for storing information.

[0120] It should be understood that the specific order or hierarchy of the steps in the provided procedures is merely an example of an exemplary method. It should be understood that the specific execution order or hierarchy of the steps involved in the procedures may be rearranged within the scope of this disclosure based on design priorities. The accompanying method claims, while providing elements of each step in an exemplary order, are not intended to be limited to the specific order or hierarchy presented.

[0121] [Explanation of Symbols] 100. Computer automatic execution program 120. Memory 130. Processor 200. Program Automatic Execution Server 300. Methods for applying 6G technology in program output 500. Fully automatic program system linked with 6G technology 600. Virtual Environment Platform

Claims

1. 6G technology-linked fully automated system features: The invention comprises a computer automatic program execution device (100), a program automatic execution server (200), a virtual environment platform (600), and an integrated payment server (142); in a program automatic system linked to 6G technology, the program automatic execution server (200) can be built into a website, a variety of software, and an AI navigation system; for example, after various AI navigation functions such as AI blind shopping navigation, AI deaf-mute dialogue commentator, AI painter, AI consultant, AI designer, AI Chinese medicine practitioner, AI music band, and AI composer and lyricist are invented, they can be stored through the program automatic execution server (200), which generates an application for computer automatic execution and deploys it to the computer automatic program execution device (100); in an embodiment, the computer automatic program execution device (100) installs a program automatic execution application distributed from the server (200); and utilizes the installed application to automatically run multiple applications and programs pre-installed on the device (100) through the user's voice and action recognition; A device (100) capable of automatically operating via a user voice signal converts a received voice signal into text via a language module containing an STT model; applies the converted text to a program automatic execution server (200) to detect instructions for program control; inputs the detected instructions into an automatic control program to control the program, and generates, through program control, an output result to generate a model and data required for at least one of a 6D metaverse, virtual reality, and augmented reality, and stores the result in a memory; The processor can implement the technical features of the embodiments of the present disclosure by executing at least one instruction stored in the memory. Specifically, the processor maps the recognized speech and text into instructions required for program control and stores them. The generation module generates instructions for program control by combining instructions matching various keywords through an instruction generation model. The automatic control unit implements automatic control of the program by inputting the generated commands into the program. The output object generation unit generates models and data required for a metaverse, virtual reality, or augmented reality (at least one of which) based on 6D technology based on the results of the automatic control. The output object generation unit generates models and data to be output in a virtual environment based on the program execution results. The models may include model-based and model-free models, CAD models, seen / unseen scenes, environments, characters, etc. The data may include objaversIVIS, large language models (LLMs), diffusion models, textures, animators, physical properties, etc. The output object generation unit is responsible for analyzing the program execution results and determining the type of output objects to be generated. Based on the generated models and data, the output object generation unit can use 3D modeling tools or BEXEL software to design 6D models in the actual virtual environment, define the model's shape, color, material, etc., and add animation; the output object generation unit can output the model and data to the virtual environment selected by the cosmic city and countryside (spatial imaging), virtual reality, and augmented reality platforms, allowing users to view or interact with the 6D model; in addition, the Metaverse supports sharing or interaction with other users, allowing users to operate and interact; therefore, when moving or operating the model in the virtual environment, the output object generation module detects and reflects environmental changes through the linkage function, and simultaneously realizes the collection and sharing of sensor data and real-time location information; based on the user's position and line of sight, the virtual information is arranged and adjusted, so the output content in the virtual environment can be tested, and the purchase and settlement of the required output objects are supported.

2. The computer automatic operation program according to claim 1, characterized in that: The system comprises a communication module, a memory and a processor, converts the user voice signal received by the communication unit into text through a language model including an STT (speech to text) conversion model, applies the converted text to a program automatic execution server (200) to detect instructions for program control; inputs the detected instructions into an automatic control program, and controls the system through the program; generates a model and data required for a metaverse, virtual reality or spatial imaging (including at least one of them) based on 6D technology through the result output by the program control, and stores it in the memory; the processor can realize the technical characteristics of the present disclosure according to the embodiment by executing at least one instruction stored in the memory; that is, stores program control instructions that match the voice and text recognized by the above-mentioned processor, and the generation unit combines the matching instructions corresponding to each keyword through the instruction generation model to generate instructions for program control; the automatic control unit inputs the generated instructions into the program, and realizes automatic control through the program; the result output by the automatic control is output by the output generation unit, and is used to generate a model and data required for at least one of the metaverse, virtual reality and augmented reality based on 3D.

3. The computer automatic program execution device according to claim 2, characterized in that: The memory stores an STT (speech-to-text) conversion model. The instruction converts the control command issued by the user in voice through the STT conversion model, and extracts keywords from the converted text to generate corresponding instructions.

4. The computer automatic program execution device according to claim 2, characterized in that: The computer automatically runs the program, accumulates and learns the program usage history of each user, updates the program automatic control model, and generates an automatic control processor for each user based on the updated program automatic control model.

5. The computer automatic program execution device according to claim 4, characterized in that: The above-mentioned program automatic control model is obtained by fine-tuning a pre-trained model; the above-mentioned pre-trained model is trained on multiple corpuses by applying at least one masked language model (MLM) and next sentence prediction (NSP) technology using semi-supervised learning; the above-mentioned fine-tuning is obtained by supervised learning of the pre-trained model, and the supervised learning uses multiple training data samples including training input data and training annotation data; the above-mentioned multiple learning input data include the process of program extraction and program control for voice input of multiple users, as well as the program control history records of each user.

6. The computer automatic operation program according to claim 1, characterized in that: The memory stores program control instructions for matching recognized voice and text; the memory stores models and data for generating a 5D-based metaverse, virtual reality, or augmented reality using program output results.

7. The program automatic operation server according to claim 1, characterized in that: The program automatic execution server (200) stores data such as websites, software, neutrino data sets, spatial stereoscopic imaging systems, and AI navigation. The server (200) generates an application for automatic program execution and deploys it to the program automatic execution device (100). The device (100) installs the program distributed from the server (200) and automatically runs the application. The installed application is used to automatically start multiple applications and programs pre-installed on the device (100) through user voice and motion recognition.

8. The fully automatic program system for 6G technology linkage according to claim 1 is characterized in that: Leveraging the high interactivity of 6G technology, the virtual environment platform enables users to operate and control programs in the virtual environment constructed by the platform through actions or voice commands. Therefore, in the remote virtual space connected to the virtual environment platform, product testing can be carried out. The Internet driving interface can be fully reproduced in the virtual space, allowing vehicle performance testing. The driving state can be realistically experienced, and fully autonomous driving can be achieved. By performing performance tests on the intended vehicle in augmented reality, once the product is confirmed to be a favorite, it can be purchased in the virtual space. In addition, in the company's remote virtual space connected to the virtual environment platform, all employees can work collaboratively through computers. Since one person can operate multiple computers remotely, new work systems and methods have been created for companies and various sectors of society. There are also non-face-to-face shopping malls and shopping methods in the virtual space where everything coexists.

9. The fully automatic program system for 6G technology linkage according to claim 8 is characterized by: In a non-face-to-face space connected to a virtual environment platform, an integrated payment server (BRICS PAY BRICS Integrated Payment System) includes an integrated payment server (BRICS PAY BRICS Integrated Payment), a user information database, and an SSL / TLS security protocol. In the virtual environment platform, users can operate and control the program in an augmented reality environment through gestures or voice commands, thereby being able to fully reproduce the driving interface in the Internet virtual space to test vehicle performance; being able to truly experience the driving state and achieve fully autonomous driving; in augmented reality, by performing a performance test on the vehicle to be purchased and confirming that it is a satisfactory product, when choosing to purchase, the user is automatically connected to the integrated payment server (BRICS PAY BRICS International Integrated Payment Server) stored in the computer automatic operation program (100), and automatically completes the payment using the personal information stored in the user information database; when it is necessary to ensure data security, the program constructs an error handling mechanism and applies security protocols such as SSL / TLS, or performs additional work such as data encryption and authentication; a system and automatic payment method for automatically setting an address based on the identity information stored in the user information database and automatically delivering to customers. The fully automatic system linked to 10.6G technology is characterized by: The multi-functional input function collects input information from all users, including the blind and mute, and based on this information, generates corresponding results on the virtual space platform and provides them to users; by recognizing the user's voice or movement input, it integrates multiple programs to achieve automatic control of the program; Provide a multifunctional input device that supports all operations required by disabled users, such as experience, driving, buying a house, commuting, etc.

11. The fully automatic program system for 6G technology linkage according to claim 1 is characterized by: After the computer automatic program running device (100) is installed on a "universe stereoscopic imaging phone" or a computer connected to 6G technology, it can be used; in the communication unit, the command (neutrino) that the user wants to execute is converted into text as a voice signal through a language model including an STT conversion model, and the converted text is applied to the program automatic running server (200) to detect the instructions required for the control program; the detected instructions are input into the automatic control program to control the above program; through the control of the above program, in order to output the output object (neutrino) to the 6D space stereoscopic imaging system, data is generated and stored in the memory; the output object generation model stored in the memory refers to: based on the six dimensions of 6D or adding "line" and "gravity" 10 n The infinite extension and vibration of dimensions connects the output of the program's automatic control results (neutrinos), extending wirelessly into the universe and generating an artificial neural network model of outputs (neutrinos) in different dimensions. To this end, the processor stores instructions for performing the following operations: realizing program control by recognizing voice and test signals to run data (neutrinos) stored in the memory; the generation unit combines the instructions matched by each keyword through the instruction generation model to generate commands (neutrinos) for program control; the automatic control unit automatically controls the program by inputting the generated instructions into the program; the output result (neutrino) output by the automatic control is generated through the output generation unit into data suitable for a 6D basic space stereoscopic imaging system; The output object generation department uses BEXEL software to design 6D models for actual operation in the virtual environment based on the data generated by the program execution results. The output object generation department transmits the neutrino conversion data to the spatial stereoscopic imaging system. As a result, users can freely access and observe the universe from any location on the earth. Combining the characteristics of neutrinos close to the speed of light and 6G technology containing 5D technology (360 zettabytes of storage space and 13.8 billion years of storage cycle), the fully automatic programmable system constructed can realize space travel. Space travel will make us realize that the universe is a 10 n dimensional space-time; in 10 n In 10-dimensional spacetime, the universe we know vibrates from a singularity, like a form made up of infinitely extended light waves, which continue to expand in an endless manner; it exists in a form similar to gravity, and the two coexist and extend infinitely to form the universe; like a bomb exploding, its form instantly fills the entire universe; because neutrinos are also made up of the same form, by connecting neutrino data to the cosmic travel shown to us by a fully automated system, it can be proved that the universe is 10 n The fact of space-time.

12. The fully automatic program system for 6G technology linkage according to claim 1 is characterized in that: It has electromagnetic wave blocking function, combined with energy harvesting function and wireless power transmission technology, and can realize remote wireless charging.

13. The ultra-thin aluminum coating film and the manufacturing method according to claim 10, characterized in that: The above-mentioned fully automatic system can achieve remote wireless charging by introducing new electromagnetic wave blocking technology into the above-mentioned terminal equipment; that is, an aluminum (Al) layer is plated in the thinnest plastic film, the thinnest plastic is pasted on it, and then a second aluminum (Al) layer is plated again, and then the thinnest plastic is pasted and a third aluminum (Al) layer is plated, and a thin plastic is covered on it to produce an ultra-thin aluminum coating that shields the electromagnetic waves of the terminal equipment; this electromagnetic wave shielding aluminum coating film can be used outdoors as a windproof and antifreeze protective film; when the generator of the terminal equipment is sealed and wrapped with this electromagnetic wave shielding aluminum coating film, electromagnetic waves can be shielded, thereby realizing remote wireless charging of the terminal.

14. The fully automated program system according to claim 10, characterized in that: The fully automatic program system linked to 6G technology is applied to the visually impaired groups who have difficulties in daily life. The visually impaired guidance function is implanted in the shopping platform and the fully automatic program system. Because the visually impaired cannot see the various products in the shopping mall when shopping, the shopping mall provides AI functions. The color, appearance, material, etc. of each product are explained through voice, Braille and sign language to help the visually impaired purchase the products they need.

15. The fully automatic system for 6G technology linkage according to claim 10, characterized in that: In order to help the deaf and mute people who have difficulty in daily life, a deaf and mute dialogue function is built in: the text input by the deaf and mute can be converted into voice and transmitted to the communication partner, and the other party's voice information can be converted into text or sign language and fed back to the deaf and mute.

16. The fully automatic program system for 6G technology linkage according to claim 9 is characterized in that: In a non-face-to-face space connected to a virtual environment platform, in an integrated payment server (BRICS PAY BRICS International Integrated Payment) - a virtual space unmanned mall that requires perfect security and coexistence of all things, when a user clicks on the option to purchase the desired product (141), the integrated payment server (142) automatically connects to the user information database (122); automatically executes the authentication procedure and completes the payment; after completing the purchase (143), it automatically connects to the user information database (122), extracts the address, phone number, name, etc. from the user's identity information, and automatically delivers the required product (output cart) to the designated delivery address; after completing the purchase (143), it automatically exits from the purchase website or shopping platform of the required product (output cart) to maintain a secure state; after completing the purchase (143), the user information database (122) is switched to power saving mode; the user information database (122) can only be powered on and started after passing the personal verification (such as facial recognition, etc.) of the integrated payment server, thereby safely protecting the user's identity information; Furthermore, when data security is required, the program implements error handling mechanisms, applies security protocols such as SSL / TLS, or performs additional tasks such as data encryption and authentication to ensure the complete security of the integrated payment server (BRICS PAY International Integrated Payment Server, which has the same structure as the integrated payment server and is therefore omitted for further explanation).

17. The structure and manufacturing method of the space 3D video phone according to claim 11, characterized in that: In a space stereoscopic imaging mobile phone connected to 6G technology, the computer automatic program running device (100) of the present invention is installed on the body of the "iPhone", and the mobile phone can be started by voice and action input without the need for touch screen input; in addition, the biggest problem of the "iPhone" - the battery - is removed, and a "nuclear battery" with better performance, smaller than a coin and sustainable use for 50 years is installed instead, so as to minimize the size of the mobile phone, so that the device can be designed into various wearable forms such as brooches, necklaces, bracelets, rings, etc., without worrying about the explosion of the mobile phone battery; while continuing to use the original display, a "photoresist projection system" including LEDs and lenses will be installed on the back of the mobile phone; when sending images to the controller via wired or wireless means, The image is sent out on the LED screen; the screen light emitted by each module is projected through a convex lens installed on the outer surface of the LED, thereby magnifying the image; the light of the magnified image is projected onto the center point of the incident hole of the "projection system" installed on the outer surface of the lens; through the projection system, the reflected line reflected from the center point of the exit hole intersects with the spatial imaging plane in the image space; an intersection is formed, thereby obtaining a spatial stereoscopic image that is significantly larger than the projected object; when the user moves or adjusts the model in the virtual environment, the output object generation unit will detect and reflect it in real time; so the user moves or deforms the three-dimensional spatial imaging like a real object, supports image shooting, and can travel in space; installed with a program device (100), the mobile phone can be automatically turned on. The fully automatic system linked to 18.6G technology is characterized by: Can be used practically (neutrinos).

19. The three-dimensional image achievable by a program device according to claim 11, characterized in that ; "Universe stereoscopic imaging mobile phone" can automatically control the program device (100) through the user's voice (action) command (spatial stereoscopic imaging) and connect to the "spatial stereoscopic imaging system" stored in the program automatic operation server; in the spatial stereoscopic imaging, the display source is used to transmit the light signal of the content to be presented to the LED controller; when the image is sent to the controller, the LED sends out the image; the screen light emitted by each display module is projected into the convex lens connected to the LED, and the image is magnified; the magnified image light (a1, a2, a3, a4) is projected onto the center point p of the incident hole through the "projection system" connected to the lens; after these light rays are reflected from the center point p' of the exit hole by the "projection system", the reflected light rays (a1", a2", a3", a4") will intersect with the spatial imaging plane in the image space, and by forming the intersection points (a1', a2', a3', a4'), a spatial imaging significantly magnified than the projected object is obtained; different light rays continuously released on the LED screen form new intersection points on another spatial imaging plane according to the release time; Therefore, the spatial imaging presented in the continuously generated different positions can present a stereoscopic image in the image space; the size of the image can be adjusted by the focal length of the lens; and while watching the spatial stereoscopic image, if the 'neutrino' is called through the mobile phone program device (100), it will automatically connect to the 'neutrino data set' stored in the program automatic execution server; through the neutrino data connection, the extremely small neutrino can connect us from the 6-dimensional space-time to the infinite 10 n Second-dimensional space-time, thus realizing space travel; this means entering the neutrino era from the 6G era.

Images