Emotional messenger system for learning emotions and manner of speaking of user and transmitting same to third party, and operation method thereof

Abstract

This emotional messenger system for learning the emotions and manner of speaking of a user and transmitting same to a third party may comprise: a memory for storing instructions; and a processor communicably connected to the memory. The system can collect emotional characteristic data including emotions, manner of speaking, language habits, and memories for a specific target person from a user through an interface using emotion-inducing questions or a recollection-inducing question, so as to learn the collected emotional characteristic data by using natural language processing technology and emotion analysis technology, generate, on the basis of the learned emotional characteristic data, a personalized artificial intelligence (AI) entity reflecting unique emotional characteristics and manner of speaking of the user by using a deep learning model, transmit the generated personalized AI entity in the form of an electronic gift to a terminal of a third party designated by the user, and, when the terminal of the third party interacts with the transmitted personalized AI entity, support that the AI entity generates a response suitable for a conversation context on the basis of the learned emotional characteristics and manner of speaking of the user so as converse such that an emotional connection can be established through the terminal of the third party.

Description

An emotional messenger system that learns a user's emotions and tone of voice and transmits them to a third party, and a method of operation thereof

[0001] The present invention relates to a communication system utilizing artificial intelligence (AI) technology, and more specifically, to an emotional messenger system that forms emotional bonds by having the AI ​​learn a user's emotions, tone of voice, language habits, etc., and conveying this information to a third party (e.g., family). In particular, the invention relates to a technology that naturally collects emotion-based data through a recall-type question interface and, based on this, generates and delivers messages that reflect the user's emotions and language patterns.

[0002] In modern society, the phenomenon of physical distance between family members widening and communication becoming severed is gradually increasing. Existing messenger services and social media platforms focus primarily on exchanging simple text, images, and videos, which limits their ability to effectively convey users' emotions and sentiments. In particular, it can be difficult to maintain emotional bonds between children and parents who lack communication due to busy daily lives, or between family members who live far apart geographically.

[0003] Existing AI-based communication tools have primarily been developed to answer user questions or provide information, and it is difficult to find services that facilitate emotional connection with third parties by learning users' emotions, tone of voice, and thought patterns. Furthermore, since current conversational AI focuses on real-time dialogue, users with time constraints face difficulties in effectively conveying their emotions and thoughts.

[0004] Furthermore, in an aging society, the emotional isolation and loneliness of the elderly are emerging as significant social issues, yet technological solutions to meet their emotional needs remain insufficient. Particularly in environments where it is difficult for children to make time to communicate directly, there is a growing need for communication methods that can maintain emotional connections with parents.

[0005] The emotional messenger system and its method of operation according to this document, which learns a user's emotions and tone of voice and transmits them to a third party, aim to induce emotional connection and communication in human relationships by utilizing artificial intelligence (AI) technology. To this end, a recall-type question interface and an emotion-induced natural language learning system may be used, which enable the AI ​​to learn emotions, tone of voice, and language habits during natural conversations with the user. This can help overcome the limitations of user experience associated with existing explicit data input methods and assist in naturally capturing and reflecting the user's emotional characteristics in the AI.

[0006] Furthermore, the purpose is to implement a system that fosters emotional bonds by delivering an AI model, which has learned the user's characteristics, as a gift to a third party, such as parents or family members. By interacting while reflecting the user's emotional rhythms and language patterns, this AI model aims to provide a deep sense of connection that feels like conversing with the actual user. Moreover, it aims to enhance the quality of communication within the family through an emotional messenger function, where the AI ​​model appropriately reconstructs and conveys emotions or thoughts that are difficult for the user to express directly. Additionally, a subscription model can be established to continuously update and refresh the user's emotional data to ensure these emotional connections are sustained. Through this, the third party interacting with the AI ​​can maintain a long-term relationship by continuously receiving the user's latest emotional state and thoughts.

[0007] Furthermore, the emotional messenger system and its method of operation, which learn and transmit a user's emotions and speech patterns according to this document to a third party, aim to expand the scope of technological application by restoring the emotional characteristics of non-existent individuals (e.g., the deceased) to provide conversation simulations. This may aim to provide the user with an experience of emotional healing by recreating cherished memories or relationships from the past in the present.

[0008] An emotional messenger system that learns a user's emotions and speech patterns and transmits them to a third party may include a memory that stores instructions and a processor connected to the memory for communication. The system collects emotional characteristic data, including emotions, speech patterns, language habits, and memories regarding a specific target person, from a user through an interface using emotion-inducing questions or recall-type questions, and learns using natural language processing technology and sentiment analysis technology. Based on the learned emotional characteristic data, it generates a personalized artificial intelligence (AI) object that reflects the user's unique emotional characteristics and speech patterns using a deep learning model. It transmits the generated personalized artificial intelligence (AI) object in the form of an electronic gift to a third party's terminal designated by the user. When the third party's terminal interacts with the transmitted personalized artificial intelligence (AI) object, the AI ​​object generates a response appropriate to the conversational context based on the learned emotional characteristics and speech patterns of the user, thereby supporting the performance of a conversation capable of emotional resonance with the third party's terminal.

[0009] The emotional messenger system and its method of operation according to this document, which learns a user's emotions and tone of voice and transmits them to a third party, have the effect of strengthening emotional connections between family members who have difficulty communicating due to physical separation or time constraints, by having AI naturally learn the user's emotions, tone of voice, and language habits and transmit them to a third party. In particular, it can contribute to the improvement of family relationships by effectively resolving the issue of communication breakdown between children and parents.

[0010] The method of collecting emotional data through a recall-type question interface of the emotional messenger system and its operation method, which learns and transmits a user's emotions and tone of voice to a third party according to this document, can provide a natural experience to the user while simultaneously offering a psychological healing effect in which the learning process itself serves as an opportunity for the user to organize their emotions and restore relationships. Users can experience organizing thoughts and feelings they were unable to express while talking to the AI ​​about memories or emotions regarding their family.

[0011] The emotional messenger system described in this document, which learns and transmits a user's emotions and tone of voice to a third party, can differentiate itself from existing messenger services by providing interactions that reflect the user's emotional rhythm, rather than simply conveying information. This serves as an emotional care platform for elderly parents and can contribute to alleviating the emotional isolation and loneliness of the older generation.

[0012] Furthermore, this invention possesses the potential for B2B expansion targeting institutions such as nursing homes, welfare agencies, and hospitals, and thus has the potential to develop into an emotional care service at a social level. In particular, it can be utilized as a solution for emotional care and communication enhancement in the field of elderly welfare.

[0013] Furthermore, the emotional messenger system and its method of operation according to this document, which learns the user's emotions and speech patterns and transmits them to a third party, and the extended model utilizing AI that has learned the speech patterns and emotional patterns of real people or celebrities, have the potential to develop into a new fan communication channel through linkage with the entertainment industry.

[0014] FIG. 1 is a diagram illustrating an emotional messenger system that learns the emotions and speech patterns of an AI-based user and transmits them to a third party according to one embodiment.

[0015] FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment.

[0016] FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment.

[0017] FIG. 4 is a block diagram showing the configuration of an emotional messenger system that learns the user's emotions and speech patterns according to one embodiment and transmits them to a third party.

[0018] FIG. 5 is a flowchart of a method for learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0019] FIG. 6 is a flowchart illustrating a method of learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0020] FIG. 7 is a flowchart illustrating a method of learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0021] In the following, embodiments may be described in detail with reference to the attached drawings. However, since various modifications may be made to the embodiments, the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments may be included within the scope of the rights.

[0022] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to specific disclosed forms, and the scope of this specification may include modifications, equivalents, or substitutions that fall within the technical concept.

[0023] Terms such as "first" or "second" may be used to describe various components, but these terms may be interpreted solely for the purpose of distinguishing one component from another. For example, a first component may be named a second component, and similarly, a second component may be named a first component.

[0024] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or joined to that other component, or that there may be other components in between.

[0025] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. A singular expression may include a plural expression unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0026] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0027] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof may be omitted. When describing embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiment, such detailed description may be omitted.

[0028] The embodiments can be implemented in various forms of products such as personal computers, laptop computers, tablet computers, smartphones, televisions, smart home appliances, intelligent automobiles, kiosks, and wearable devices.

[0029] Artificial Intelligence (AI) systems are computer systems that implement human-level intelligence; unlike existing rule-based smart systems, they are systems in which machines learn and make decisions autonomously. As AI systems improve in recognition accuracy and gain a more accurate understanding of user preferences with continued use, existing rule-based smart systems are gradually being replaced by deep learning-based AI systems.

[0030] Artificial intelligence technology consists of machine learning and component technologies utilizing machine learning. Machine learning is an algorithmic technology that autonomously classifies and learns the characteristics of input data, while component technologies are technologies that mimic the cognitive and judgmental functions of the human brain by utilizing machine learning algorithms such as deep learning, and are comprised of technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control.

[0031] The various fields where artificial intelligence technology is applied are as follows. Linguistic understanding is a technology that recognizes, applies, and processes human language and text, and may include natural language processing, machine translation, dialogue systems, question answering, and speech recognition / synthesis. Visual understanding is a technology that perceives and processes objects like human vision, and may include object recognition, object tracking, image search, person recognition, scene understanding, spatial understanding, and image enhancement. Inference and prediction is a technology that judges information to logically infer and predict, and may include knowledge / probability-based inference, optimization prediction, preference-based planning, and recommendation. Knowledge representation is a technology that automatically processes human experiential information into knowledge data, and may include knowledge construction (data generation / classification) and knowledge management (data utilization). Motion control is a technology that controls the autonomous driving of vehicles and the movement of robots, and may include motion control (navigation, collision, driving) and manipulation control (behavior control).

[0032] Generally, to apply machine learning algorithms to real-world situations, training is performed using a trial-and-error method due to the inherent characteristics of machine learning methodologies. In particular, deep learning can require hundreds of thousands of iterations. Since it is impossible to execute this in a real physical external environment, training can instead be performed through simulation by virtually recreating the actual physical environment on a computer.

[0033] In the present invention, Artificial Intelligence (AI) refers to a technology that imitates human learning ability, reasoning ability, and perceptual ability, and implements them on a computer, and may include concepts such as machine learning and symbolic logic. Machine Learning (ML) is an algorithmic technology that classifies or learns the characteristics of input data on its own. AI technology can analyze input data as a machine learning algorithm, learn from the results of the analysis, and make judgments or predictions based on the results of the learning. Furthermore, technologies that mimic the functions of the human brain, such as cognition and judgment, by utilizing machine learning algorithms can also be understood as falling within the category of AI. For example, technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control may be included.

[0034] Machine learning can refer to the process of training neural network models using experience in processing data. Through machine learning, computer software can improve its own data processing capabilities. A neural network model is constructed by modeling the correlations between data, and these correlations can be expressed by multiple parameters. A neural network model extracts and analyzes features from given data to derive correlations between them; machine learning can be defined as the process of optimizing the model's parameters by repeating this process. For example, a neural network model can learn the mapping (correlation) between inputs and outputs for data given as input-output pairs. Alternatively, even when only input data is provided, a neural network model can derive regularities between the given data and learn those relationships.

[0035] An artificial intelligence learning model or neural network model can be designed to implement the structure of the human brain on a computer and may include multiple network nodes that have weights and simulate neurons of a human neural network. The multiple network nodes may have interconnected relationships by simulating the synaptic activity of neurons, where neurons exchange signals through synapses. In an artificial intelligence learning model, multiple network nodes may be located in layers of different depths and exchange data according to convolutional connections. The artificial intelligence learning model may be, for example, an Artificial Neural Network (ANN) or a Convolutional Neural Network (CNN). As an embodiment, the artificial intelligence learning model may be machine learned according to methods such as supervised learning, unsupervised learning, and reinforcement learning. Machine learning algorithms for performing machine learning may include Decision Tree, Bayesian Network, Support Vector Machine, Artificial Neural Network, Ada-boost, Perceptron, Genetic Programming, and Clustering.

[0036] Among these, CNNs are a type of multilayer perceptron designed to use minimal preprocessing. CNNs consist of one or more convolutional layers and standard artificial neural network layers stacked on top, and can additionally utilize weights and pooling layers. Thanks to this structure, CNNs can fully utilize two-dimensional input data. Compared to other deep learning structures, CNNs demonstrate good performance in both image and audio fields. CNNs can also be trained using standard backpropagation. CNNs have the advantage of being easier to train than other feedforward artificial neural network techniques and can use a smaller number of parameters.

[0037] Convolutional networks are neural networks comprising sets of nodes with bounded parameters. Many computer vision tasks have been significantly improved, driven by the increased size of available training data and the availability of computational power, combined with algorithmic advancements such as discriminative linear units and dropout training. In the case of massive datasets, such as those available for many tasks today, outfitting is not critical, and increasing the network size improves test accuracy. Optimal utilization of computing resources becomes a limiting factor. To address this, distributed, scalable implementations of deep neural networks can be employed.

[0038]

[0039] FIG. 1 is a diagram illustrating an emotional messenger system that learns the emotions and speech patterns of an AI-based user and transmits them to a third party according to one embodiment.

[0040] As illustrated in FIG. 1, an emotional messenger system (400) that learns the emotions and speech patterns of an AI-based user and transmits them to a third party may include a plurality of user terminals (110-1, ���), a server (120), and a database (130). According to one embodiment, the database (130) is shown as being configured separately from the server (120), but is not limited thereto, and the database (130) may be provided within the server (120). For example, the server (120) may include a plurality of AIs for performing machine learning algorithms. According to one embodiment, the plurality of user terminals (110-1, ����), the server (120), and the database (130) may be connected to communicate with each other through a network (N).

[0041] A network (N) can perform wireless or wired communication between multiple user terminals (110-1, 110), a server (120), a database (130), etc. For example, the network can perform wireless communication according to methods such as LTE (long-term evolution), LTE-A (LTE Advanced), CDMA (code division multiple access), WCDMA (wideband CDMA), WiBro (Wireless BroadBand), WiFi (wireless fidelity), Bluetooth, NFC (near field communication), GPS (Global Positioning System), or GNSS (global navigation satellite system). For example, the network (N) can perform wired communication according to methods such as USB (universal serial bus), HDMI (high definition multimedia interface), RS-232 (recommended standard 232), or POTS (plain old telephone service).

[0042] The database (130) can store various data. The data stored in the database (130) may include software (e.g., programs) as data acquired, processed, or used by at least one component of a plurality of user terminals (110-1, server (120). The database (130) may include volatile and / or non-volatile memory.

[0043]

[0044] FIG. 2 is a diagram illustrating the learning of a neural network according to one embodiment.

[0045] As illustrated in FIG. 2, the learning device can train a neural network (123) to process review responses received from a plurality of user terminals (110-1, ���) by item. Additionally, the learning device can train a neural network (123) to extract user stay history from user movement path information. According to one embodiment, the learning device may be a separate entity from the server (120), but is not limited thereto.

[0046] The neural network (123) includes an input layer (121) into which training samples are input and an output layer (125) that outputs training outputs, and can be trained based on the difference between the training outputs and the labels. Here, the labels are defined based on items corresponding to review responses and can be defined based on user dwell history corresponding to movement path information. The neural network (123) is connected as a group of multiple nodes and is defined by weights between the connected nodes and an activation function that activates the nodes.

[0047] The learning device can train the neural network (123) using the GD (Gradient Descent) technique or the SGD (Stochastic Gradient Descent) technique. The learning device can use a loss function designed by the outputs and labels of the neural network.

[0048] The learning device can calculate the training error using a predefined loss function. The loss function can be predefined with labels, outputs, and parameters as input variables, where the parameters can be set by weights within the neural network (123). For example, the loss function can be designed in the form of Mean Square Error (MSE), entropy, etc., and various techniques or methods may be employed in the embodiments in which the loss function is designed.

[0049] The learning device can find weights that affect the training error using the backpropagation technique. Here, the weights are relationships between nodes within the neural network (123). The learning device can use the SGD technique with labels and outputs to optimize the weights found through the backpropagation technique. For example, the learning device can update the weights of the loss function defined based on the labels, outputs, and weights using the SGD technique.

[0050] According to one embodiment, the learning device extracts first objects from review responses, obtains first labels which are items corresponding to the first objects, applies the first objects to a first neural network to generate first training outputs corresponding to the first objects, and can train the first neural network based on the first training outputs and the first labels.

[0051] The learning device extracts second objects from movement path information, obtains second labels which are user dwell history corresponding to the second objects, applies the second objects to a second neural network to generate second training outputs corresponding to the second objects, and can train the second neural network based on the second training outputs and second labels.

[0052] According to one embodiment, the learning device can generate first training feature vectors based on the constituent features, location features, and pattern features of the review response. Various methods may be employed for extracting features.

[0053] According to one embodiment, the learning device can generate second training feature vectors based on the constituent features, length features, and pattern features of the movement path information. Various methods may be employed for extracting features.

[0054] According to one embodiment, the learning device may obtain training outputs by applying first training feature vectors to the neural network (123). The learning device may train a review item extraction algorithm of the neural network (123) based on the training outputs and first labels. The learning device may train a review item extraction algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. The server (120) may extract items from review responses using the first neural network that has been trained.

[0055] According to one embodiment, the learning device can obtain training outputs by applying second training feature vectors to the neural network (123). The learning device can train the user stay history acquisition algorithm of the neural network (123) based on the training outputs and second labels. The learning device can train the user stay history acquisition algorithm of the neural network (123) by calculating training errors corresponding to the training outputs and optimizing the connection relationships of nodes within the neural network (123) to minimize the training errors. The server (120) can obtain user stay history from movement path information using the second neural network that has been trained.

[0056]

[0057] FIG. 3 is a diagram illustrating the configuration of an artificial intelligence model according to one embodiment.

[0058] An artificial intelligence model according to one embodiment may include an input layer, a hidden layer, and an output layer.

[0059] The input layer is the layer associated with the input values ​​fed into the artificial intelligence model.

[0060] In the hidden layer, a feature map can be output by performing MAC (multiply-accumulate) and activation operations on the input values.

[0061] A MAC operation can be an operation that multiplies each input value by its corresponding weight and sums the multiplied values.

[0062] The activation operation may be an operation that inputs the result of the MAC operation into an activation function and outputs a result value. The activation function may be of various types. For example, the activation function may include a sigmoid function, a tangent function, a ReLU function, a Leaky ReLU function, a Max Out function, and / or an ELU function, but there are no restrictions on the types.

[0063] A hidden layer may consist of at least one layer. For example, if the hidden layer consists of a first hidden layer and a second hidden layer, the first hidden layer performs MAC operations and activation operations based on input values ​​of an input system to output a feature map, and the feature map, which is the result value of the first hidden layer, may become the input value of the second hidden layer. The second hidden layer may perform MAC operations and activation operations based on the feature map, which is the result value of the first hidden layer.

[0064] The output layer may be a layer associated with the result of an operation performed in the hidden layer.

[0065] In one embodiment, the learning model learns syllable (character) patterns that are frequently combined and used in a given corpus to automatically learn the boundaries of compound words and object names, integrates object information from a first UI source with object information rendered in a browser to create a learning object information file, uses the learning object information file to create learning data for the learning of a deep learning network, receives data from various domains of a support system, standardizes the data from the various domains into an integrated format based on at least one standardization method corresponding to each of the various domains, learns and infers data from a specific domain, determines information to be transmitted for standardization from the data of the specific domain, and can perform post-processing on the data from the various domains. The first UI source includes an XML file, and the learning object information file includes an input JSON file for feature learning and an output JSON file that is the correct answer (label) data during learning, and the output JSON file includes a file containing HTML DOM Tree information implemented in compliance with web standards, and the various domains include at least one of a RAN (radio access network), a transport, or a core, and post-processing may include a correlation function.

[0066]

[0067] FIG. 4 is a block diagram showing the configuration of an emotional messenger system that learns the user's emotions and speech patterns according to one embodiment and transmits them to a third party.

[0068] A system (400) according to one embodiment may include a processor (420) and a memory (430), and some of the illustrated components may be omitted or substituted. A system (400) according to one embodiment may be a server or a terminal. According to one embodiment, the processor (420) is a component capable of performing operations or data processing regarding the control and / or communication of each component of the system (400), and may be composed of one or more processors. The memory (430) may store information related to the method described above or store a program in which the method described above is implemented. The memory (430) may be volatile memory or non-volatile memory. The memory (430) may store various file data, and the stored file data may be updated according to the operation of the processor (120).

[0069] According to one embodiment, the processor (420) can execute a program and control the device (400). The code of the program executed by the processor (120) can be stored in memory (430). Operations of the processor (420) can be performed by loading instructions stored in memory (430). The system (400) can be connected to an external device (e.g., a personal computer or a network) through an input / output device (not shown in the drawing) and exchange data.

[0070]

[0071] FIG. 5 is a flowchart of a method for learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0072] Although process steps, method steps, algorithms, etc. are described in a sequential order in the flowchart of FIG. 5, such processes, methods, and algorithms may be configured to operate in any suitable order. In other words, the steps of the processes, methods, and algorithms described in various embodiments of the present invention do not need to be performed in the order described in the present invention. Furthermore, even if some steps are described as being performed asynchronously, in other embodiments, such steps may be performed simultaneously. Also, the example of a process by the depiction in the drawings does not mean that the exampled process excludes other variations and modifications therefrom, does not mean that any of the exampled process or its steps is essential to one or more of the various embodiments of the present invention, and does not mean that the exampled process is desirable.

[0073] In operation 510, the system (400) can collect emotional characteristic data (e.g., emotions, tone of voice, language habits, memories, etc.) about a specific target person from a user through an emotion-inducing or recall-type question interface and learn this using natural language processing and sentiment analysis technology.

[0074] According to one embodiment, the system (400) can collect emotional characteristic data regarding a specific target person from a user through an emotion-inducing question or recall-inducing question interface and learn from it by utilizing natural language processing technology and sentiment analysis technology. Here, an emotion-inducing question refers to a type of question that naturally elicits the user's emotions, and a recall-inducing question refers to a question designed to allow the user to recall memories or experiences related to a specific person. These specially designed questions may aim to collect data that includes emotional connections and emotional contexts, rather than simple factual information.

[0075] According to one embodiment, the system (400) can create an environment in which the user can naturally express deep emotions and memories regarding the target person by presenting open-ended questions to the user, such as "When was your happiest moment with that person?", "What emotions do you feel when you think of that person?", and "What expressions did that person mainly use often?". The user's response can be received in the form of text or voice, and the system (400) can analyze it through a natural language processing algorithm.

[0076] According to one embodiment, the collected data is processed not as a simple set of words or sentences, but as complex emotional characteristic data including emotional intensity, tone, context, semantic connectivity, etc. The system (400) can utilize sentiment analysis technology to extract emotional states, emotional change patterns, linguistic features (frequently used vocabulary, sentence structure, specific expressions), and non-linguistic cues (in the case of speech, tone, pitch, speed) from this data and store them in a structured form. This process of collecting and learning emotional characteristic data has significance beyond simple information collection and can provide a therapeutic effect to the user themselves.

[0077] In operation 520, the system (400) can generate a personalized artificial intelligence (AI) object that reflects the user's unique emotional characteristics and speech patterns by utilizing a deep learning model based on learned emotional characteristic data.

[0078] According to one embodiment, the system (400) can generate a personalized artificial intelligence (AI) object that reflects the user's unique emotional characteristics and speech patterns by utilizing a deep learning model based on emotional characteristic data learned in operation 510. Here, a personalized artificial intelligence (AI) object may refer to a digital entity that can communicate on behalf of the user within the context of a relationship with a specific target person by mimicking the user's emotional characteristics, language patterns, and manner of expressing emotions.

[0079] According to one embodiment, a deep learning model can build a unique language model for a user by learning the user's vocabulary selection patterns, sentence structure, expression method, and style of conveying emotions from collected text and voice data. In this process, the system (400) analyzes complex patterns beyond simple word usage frequency and, in particular, can deeply learn changes in emotional expression according to specific topics or situations, and emotional context regarding specific target individuals.

[0080] According to one embodiment, the creation of a personalized artificial intelligence (AI) object can be achieved by utilizing various neural network structures (e.g., Transformer-based models, recurrent neural networks, attention mechanisms, etc.), thereby enabling the implementation of an AI object with emotional consistency and contextual understanding beyond simple text generation. By creating an AI object that takes into account the user's emotional state, values, and memories, as well as special emotional context derived from a relationship with a specific target person, the system (400) can implement an emotion-based communication object that is fundamentally different from a simple chatbot or virtual assistant.

[0081] In operation 530, the system (400) can deliver the generated personalized artificial intelligence (AI) object to a third party terminal designated by the user in the form of an electronic gift and support emotional connection by responding based on the learned characteristics of the user during interaction.

[0082] According to one embodiment, the system (400) can deliver a personalized artificial intelligence (AI) object generated in operation 520 to a third party terminal designated by the user in the form of an electronic gift, and can support a conversation capable of emotional connection by generating a context-appropriate response based on the learned emotional characteristics and tone of voice of the user during interaction between the third party terminal and the AI ​​object. Here, "in the form of an electronic gift" may refer to a method of delivering an artificial intelligence (AI) object by packaging it in a form that a third party terminal can access and interact with in a digital environment.

[0083] According to one embodiment, when a third party's terminal converses with an artificial intelligence (AI) object received by the system (400), the system (400) can generate a response that reflects the user's emotional characteristics and linguistic features beyond simple information transmission. At this time, the system (400) can identify the message intent and emotion of the third party's terminal, consider the current conversation context, and simultaneously determine the most appropriate response content and emotional tone based on the learned emotional characteristics of the user. The system (400) can generate a natural conversational response message by understanding the flow of the conversation through a context recognition algorithm and reflecting the user's unique vocabulary selection, sentence length, and expression style in the response through a customized tone conversion algorithm.

[0084] This emotional interaction support function can enable a third-party terminal to experience an emotional bond as if conversing with a real user through a conversation with an artificial intelligence (AI) object. In particular, the system (400) can provide a more vivid emotional experience by providing not only simple text-based responses but also, if necessary, synthesized voices that mimic the user's voice characteristics.

[0085]

[0086]

[0087] FIG. 6 is a flowchart illustrating a method of learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0088] Although the process steps, method steps, algorithms, etc. are described in a sequential order in the flowchart of FIG. 6, such processes, methods, and algorithms can be configured to operate in any suitable order.

[0089]

[0090] In operation 610, the system (400) may present open-ended or selectable questions through an interface sequentially or non-sequentially to induce the user to recall or freely talk about memories, shared experiences, emotions, etc., with a specific target person.

[0091] According to one embodiment, the system (400) may present open-ended or multiple-choice questions through an interface sequentially or non-sequentially to induce the user to recall or freely talk about memories, shared experiences, emotions, etc., with a specific target person. Here, an "open-ended question" refers to a question in which the user can freely respond in their own words (e.g., "Please tell me about the happiest moment you had with your mother"), and a "multiple-choice question" refers to a question that allows the user to choose from several options (e.g., "What emotion do you feel most strongly when thinking of your mother? 1. Longing 2. Gratitude 3. Warmth 4. Other"). The system (400) may dynamically select and present the next question based on the user's emotional state and previous response through a conversational flow algorithm, which is intended to elicit deeper emotional data while maintaining a natural conversational flow.

[0092] According to one embodiment, 'sequential presentation' refers to a method of providing questions in a predetermined order, and 'non-sequential presentation' may refer to a method in which the order of questions is dynamically changed according to the user's response content and emotional state. The questions presented by the system (400) focus not on collecting simple factual information, but on eliciting the user's emotional experiences, memories, and emotional connections derived from the relationship with the target person. Unlike general data collection, this emotion-centered question design creates an environment where the user can naturally express and recall emotions, thereby enabling the collection of more authentic emotional data.

[0093] In addition, the system (400) can optimize the conversation environment by taking into account question intervals, emphasis, and emotional response to previous responses, so that the user can feel psychologically secure and express their emotions.

[0094] In operation 620, the system (400) can receive the user's voice or text-based natural language response input to the above questions.

[0095] According to one embodiment, the system (400) may receive a user’s voice or text-based natural language response input to the questions. Here, "voice-based input" refers to a method of collecting a response verbally expressed by the user through voice recognition technology, and may include paralinguistic elements such as the user’s voice tone, pitch, speech rate, and pause patterns. "Text-based input" refers to text data directly entered by the user through a keyboard, etc., and may include features such as character selection, sentence structure, use of emoticons, and emphasis expressions.

[0096] According to one embodiment, the system (400) can accommodate both forms of input through a multimodal input processing algorithm and can support the user in expressing emotions in the way most comfortable for them by considering the advantages and disadvantages of each input method. In the case of voice input, the system (400) can apply a context-based error correction algorithm to minimize recognition errors that may occur during the process of converting the user's speech into text by utilizing real-time voice recognition technology.

[0097] Additionally, the system (400) can recognize and record speech patterns such as hesitation, repetition, and emphasis shown by the user during the response process as important emotional cues. Support for these various input methods increases user accessibility and enables the system to be utilized in various situations (e.g., while on the move, when typing is inconvenient, when emotional expression is more natural in speech, etc.), thereby improving the continuity and quality of data collection.

[0098] In operation 630, the system (400) analyzes the input natural language response using a natural language processing and sentiment analysis algorithm and can extract linguistic features and non-linguistic cues including the user's emotional state, emotional intensity, associated emotional patterns, frequently used vocabulary / expressions, sentence structure, response speed, voice tone / pitch, etc.

[0099] According to one embodiment, the system (400) analyzes an input natural language response using a natural language processing and sentiment analysis algorithm and can extract linguistic features and non-linguistic cues including the user's emotional state, emotional intensity, associated emotional patterns, frequently used vocabulary / expressions, sentence structure, response speed, voice tone / pitch, etc. Here, 'natural language processing algorithm' refers to a technology that enables a computer to understand and analyze human language, and may include processes such as morphological analysis, syntactic analysis, and semantic analysis. 'Sentiment analysis algorithm' may refer to a technology that identifies emotions, attitudes, opinions, etc. contained in text or voice.

[0100] According to one embodiment, the system (400) can extract the emotional context and deep meaning contained in the content beyond the surface meaning of simple words or sentences through this algorithm. 'Emotional state' refers to the main emotion (joy, sadness, longing, anger, etc.) felt by the user at a specific moment, and 'emotional intensity' is a numerical representation of how strongly that emotion is expressed. 'Associative emotion pattern' may refer to a sequence or change of emotions that appear repeatedly depending on a specific topic or situation.

[0101] Linguistic features include vocabulary frequently used by the user, specific expressions, sentence length and complexity, and rhetorical features, while non-linguistic cues include tone of voice, speaking speed, pauses, and emphasis methods. The system (400) can capture not only explicitly expressed emotions but also implicit emotional cues through a multi-layered emotion recognition algorithm, and can track emotional changes within the entire conversation flow rather than a single response through context-based emotion analysis.

[0102] In operation 640, the system (400) stores the extracted emotional characteristic data in a structured database (DB) and can build a unique emotional model and language model of a user for a specific target person through a machine learning algorithm.

[0103] According to one embodiment, the system (400) stores extracted emotional characteristic data in a structured database (DB) and can build a user's unique emotion model and language model for a specific target person through a machine learning algorithm. Here, 'structured database' refers to a data storage system designed to systematically classify, store, and manage emotional characteristic data, and may include a multidimensional classification system such as emotion categories, intensity, contextual information, and temporal changes.

[0104] A machine learning algorithm can be described as an artificial intelligence technology that learns patterns from stored data and builds a prediction model. The system (400) can build two core models based on stored emotional characteristic data.

[0105] The emotion model may refer to a mathematical model of the unique characteristics of the emotional response patterns, emotional triggers, emotional transfer probabilities, and emotional expression methods that a user has toward a specific target person. The 'language model' models linguistic characteristics such as the user's vocabulary selection, sentence structure, expression style, and conversation flow, and can capture changes in language usage patterns according to emotional context. The system (400) may combine supervised learning and unsupervised learning techniques in the process of building these models, and may apply an incremental learning method that gradually improves the accuracy and fineness of the model as the user's responses increase.

[0106] Additionally, the system (400) can analyze the interrelationship between the constructed emotion model and the language model to derive an integrated understanding of what linguistic expressions a specific emotional state leads to, or what emotional state a specific language pattern reflects. This model construction can be described as a process that goes beyond simple data classification to realize the unique characteristics of the user's inner emotional world and its expression methods in a digital form, which can help future personalized artificial intelligence (AI) objects authentically reflect the user's emotional characteristics.

[0107]

[0108]

[0109] FIG. 7 is a flowchart illustrating a method of learning a user's emotions and speech patterns and transmitting them to a third party according to one embodiment.

[0110] Although the process steps, method steps, algorithms, etc. are described in a sequential order in the flowchart of FIG. 7, such processes, methods, and algorithms can be configured to operate in any suitable order.

[0111] In operation 710, the system (400) can receive a message from a third party's terminal and identify the intention, question type, and emotion.

[0112] According to one embodiment, the system (400) receives a message in the form of text or voice within a conversation context from a third party's terminal and can identify the core intent, question type, and explicit or implicit emotion of the input message through a natural language understanding algorithm.

[0113] Here, 'natural language understanding algorithms' refer to technologies that process human language so that computers can understand its meaning and intent, and may include techniques such as semantic analysis, intent classification, and named entity recognition.

[0114] 'Core intent' may refer to the primary purpose that a third-party terminal intends to convey through a message (e.g., question, request for information, sharing of emotions, presentation of opinions, etc.). 'Question type' refers to cases where the message is a question; depending on its nature, it can be classified into fact verification, information request, seeking advice, asking for opinions, or confirming emotions. 'Explicit emotion' refers to an emotional state expressed directly (e.g., "I am sad," "I am really happy," etc.), while 'implicit emotion' may refer to an emotional state indirectly revealed through word choice, sentence structure, context, etc.

[0115] According to one embodiment, the system (400) can identify a more accurate emotional state by analyzing not only text content but also paralinguistic characteristics such as tone, speed, and pauses in the case of voice through a multi-signal integrated processing method. Additionally, the system (400) can track changing intentions and emotional states within the entire conversation flow, rather than just a single message, by applying a context-aware algorithm that utilizes previous conversation history and contextual information. This deep language understanding process aims to accurately capture complex human communication patterns in a digital environment, going beyond simple keyword matching or surface analysis, and holds significant meaning as the first step for an artificial intelligence (AI) object to form a genuine emotional connection with a third-party terminal.

[0116] In operation 720, the system (400) can determine an appropriate response content and emotional tone based on the intent of a third party and user emotional data.

[0117] According to one embodiment, the system (400) can determine the most appropriate response content and emotional tone through a context recognition algorithm based on learned emotional characteristic data of the user, while considering the message intent, emotion, and current flow of conversation of the identified third party terminal.

[0118] Here, a 'context-aware algorithm' may refer to a technology that derives an optimal response by comprehensively considering the context of the conversation, the characteristics of the participants' relationships, previous conversation content, and current emotional state. 'Response content' refers to substantive messages such as information to be conveyed, answers to questions, expressions of empathy, and advice, while 'emotional tone' may refer to the emotional atmosphere or attitude (warmth, empathy, humor, seriousness, etc.) in which the message is delivered.

[0119] According to one embodiment, the system (400) can reflect memories, emotional states, values, language habits, and the user's emotional rhythm regarding a specific target person in the generation of a response based on learned emotional characteristic data of the user. 'Emotional rhythm' may refer to a consistent pattern of emotional response shown by the user to a specific situation or topic. The system (400) can determine the response by finding the optimal balance point between the requirements of a third party's terminal and the user's emotional expression style using a multi-criteria decision-making algorithm. In this process, the system (400) may simultaneously consider various factors such as the informational accuracy of the response, emotional appropriateness, degree of reflection of user characteristics, and naturalness of the conversation flow.

[0120] In particular, the system (400) can set a response tone that reflects the user's unique emotional characteristics while harmonizing with the current emotional state of a third party's terminal through an emotional resonance optimization technique.

[0121] In operation 730, the system (400) can apply an algorithm that converts the determined response content and tone into the user's unique speech pattern.

[0122] According to one embodiment, the system (400) may use a customized speech style conversion algorithm that converts the determined response content and emotional tone into vocabulary selection, sentence length, expression style, and intonation that constitute the user's unique speech style.

[0123] Here, the 'customized speech style conversion algorithm' refers to a technology that analyzes a user's linguistic characteristics and speech style and applies them to new content to reproduce the user's unique expression style. 'Vocabulary selection' refers to the words or expressions (e.g., interjections, figurative language, emphasizing words, specific terms, etc.) preferred by the user when expressing specific concepts or emotions. 'Sentence length' refers to the average length and complexity of sentences generally used by the user, while 'expression style' may include sentence structures frequently used by the user, such as direct, indirect, interrogative, or descriptive forms. 'Intonation' refers to characteristics applied during voice output, such as the user's voice tone, rhythm, and stress patterns.

[0124] According to one embodiment, the system (400) can utilize a style transfer deep learning model to naturally apply the user's unique linguistic characteristics while preserving the core meaning of the response content. In this process, the system (400) is optimized to maintain consistency at the discourse level and the user's linguistic characteristics, rather than merely at the level of superficial vocabulary or grammar. Additionally, the system (400) learns tone change patterns according to emotional context and can dynamically adjust the tone according to the emotional content of the response.

[0125] In operation 740, the system (400) can generate natural conversational response messages that reflect the user's language patterns and emotional characteristics.

[0126] According to one embodiment, the system (400) can generate natural conversational response messages by reflecting the learned language patterns and emotional characteristics of the user. Here, "conversational response message" may refer to a response that possesses the naturalness and organic flow of human conversation, rather than simple information transmission or mechanical answers. The system (400) can construct grammatically accurate and contextually appropriate sentences by utilizing Natural Language Generation (NLG) technology. In this process, the system (400) can implement linguistic characteristics that make it seem as though the user is directly participating in the conversation by applying the user's language patterns (frequency of use of specific conjunctions, sentence beginning methods, rhetorical expressions, emphasis patterns, etc.) to the response.

[0127] In addition, the system (400) can generate a response that reflects the user's emotional characteristics (emotional expression style, empathy expression pattern, emotional response tendency, etc.) and goes beyond simple verbal imitation to convey the user's emotional presence.

[0128] According to one embodiment, the system (400) can apply a coherence and consistency maintenance algorithm to control the user's emotional and linguistic identity so that it is naturally maintained within the overall conversation flow rather than at the individual sentence level. In addition, the system (400) can implement a flexible response appropriate to the situation rather than simple pattern replication by appropriately modifying and applying the user's expression style to suit the current conversation situation through a fine-tuning technique based on the conversation context.

[0129] In operation 750, the system (400) can support the formation of an emotional bond by outputting the generated response message to a third party as text or synthetic speech that mimics user voice features.

[0130] According to one embodiment, the system (400) can output a generated response message to a third party's terminal in the form of text or a synthesized voice that mimics the user's voice characteristics, thereby supporting the recipient in forming an emotional bond.

[0131] Here, 'synthesized voice' may refer to a voice that digitally reproduces the user's voice characteristics (timbre, intonation, speaking speed, pause patterns, etc.) by analyzing them. The system (400) can convert a text-based response message into a voice that reflects the user's unique voice characteristics by utilizing advanced voice synthesis technology (Text-to-Speech, TTS) and voice cloning technology.

[0132] In this process, the system (400) can apply emotion-based voice modulation technology that goes beyond simple voice conversion and dynamically adjusts the tone, emphasis, and speed of the voice according to the emotional content of the message. Additionally, the system (400) can automatically select or recommend a text or voice output method to suit the environment (noise level, characteristics of the device used, etc.) and preferences of a third party's terminal through a context-aware output optimization function.

[0133] According to one embodiment, the system (400) incorporates emotional bonding enhancement factors into the response output to effectively convey intimacy, bonding, and emotional resonance derived from the relationship with a specific target person as well as the user's linguistic and vocal characteristics. This customized output method provides a vivid experience in which a third-party terminal communicates with the actual user through interaction with an artificial intelligence (AI) object, which can maintain and strengthen emotional connections despite physical distance or time constraints.

[0134]

[0135]

[0136] According to one embodiment, the system (400) may intermittently or periodically collect additional emotional characteristic data from the user through an interface via emotion-inducing questions or recall-inducing questions to reflect real-time emotional state changes of the user or changes in new memories or thoughts about the specific target person in the personalized artificial intelligence (AI) object, integrate this data into existing training data through a training data integration algorithm, and update the emotion model and language model of the personalized artificial intelligence (AI) object.

[0137] When the user wishes to send a message, such as a greeting, anniversary congratulations, or specific information, to the third party's terminal at a specific time, the system (400) receives the content of the message, automatically applies the learned tone and sentiment of the user through a customized tone conversion algorithm to reconstruct it, and then remotely transmits it to the third party's terminal in the form of an asynchronous push notification or message through the personalized artificial intelligence (AI) object.

[0138] The system (400) can be controlled to include at least one additional function of automatically generating an emotion report containing indicators that can infer the state of emotional connection or changes in relationship between the user and the third party's terminal by analyzing interaction data including the frequency of conversation between the personalized artificial intelligence (AI) object and the third party's terminal, the distribution of conversation topics, major emotional keywords that appeared during the conversation, and the history of messages remotely transmitted by the user through a data analysis algorithm, and providing the user and / or the third party's terminal regularly or upon request.

[0139] According to one embodiment, the system (400) may intermittently or periodically collect additional emotional characteristic data from a user through an interface using emotion-inducing questions or recall-type questions in order to reflect changes in the user's real-time emotional state or changes in new memories or thoughts regarding a specific target person into a personalized artificial intelligence (AI) object. Here, "changes in real-time emotional state" refers to changes in the intensity, attributes, context, etc., of the emotions the user feels toward a specific target person over time, and "changes in new memories or thoughts" may include newly emerging memories, realizations, changes in perspective, or matters that were not previously mentioned. "Intermittent collection" refers to a method of collecting data irregularly according to specific events or user requests, and "periodic collection" refers to a method of conducting regular data collection at predetermined intervals (weekly, monthly, etc.). The system (400) may integrate the additional emotional characteristic data collected in this way into existing training data through a training data integration algorithm. The 'learning data integration algorithm' may refer to an intelligent integration method that organically combines new data and existing data by considering temporal continuity, semantic consistency, and emotional context between the data, rather than simply physically combining them. In this process, the system (400) may apply a time-weighted model to give higher weight to recent emotional data, or adjust emotional discrepancies between existing data and new data through an emotional conflict resolution algorithm. Through this integration process, the system (400) can progressively update the emotional model and language model of a personalized artificial intelligence (AI) object, which can enable more accurate and empathetic communication that reflects the user's current emotional state and latest memories.These continuous learning and updating functions enable artificial intelligence (AI) objects to function as dynamic emotional mediators that share in the user's emotional changes and growth, rather than as static emotional replicas, thereby serving as a key element in supporting the maintenance of long-term relationships and deep emotional exchanges with third-party terminals.

[0140] According to one embodiment, when a user wishes to send a message, such as a greeting, congratulations on an anniversary, or specific information, to a third party's terminal at a specific time, the system (400) receives the content of the message, reconstructs it by automatically applying the user's tone and sentiment learned through a customized tone conversion algorithm, and then remotely transmits it to the third party's terminal in the form of an asynchronous push notification or message through a personalized artificial intelligence (AI) object. Here, "specific time" may refer to a situation where it is difficult for the user to communicate directly in real time, a date of special significance (birthday, anniversary, etc.), or a time when periodic greetings are required. The "customized tone conversion algorithm" may refer to a natural language processing technology that applies the user's unique vocabulary selection, sentence structure, expression patterns, and emotional nuances to the message while maintaining the core content of the input message. In this process, the system (400) links an intent recognition module that identifies the intent and context of the message with a database of the user's linguistic characteristics, thereby generating a message in which the user's unique expression style is naturally incorporated, rather than a simple word substitution. 'Asynchronous push notification' refers to a notification message transmitted independently of the real-time conversation flow, designed so that a third party's terminal can check it at an appropriate time. The system (400) can automatically select the optimal transmission time by applying a message delivery timing optimization algorithm, taking into account the third party's terminal's daily patterns, previous message checking times, importance, etc. Additionally, the system (400) can strengthen the sense of emotional connection by adding contextual elements that match the current situation (weather, season, location, etc.) to a simple text message through a message context enhancement function, such as shared memories between the user and the third party's terminal, references to previous conversations, or the current situation.This asynchronous message transmission function allows users to share important moments or emotions with third-party terminals without the constraints of time and space, thereby alleviating emotional disconnection caused by physical separation or the busyness of daily life and contributing to maintaining continuous emotional bonds.

[0141] According to one embodiment, the system (400) can automatically generate an emotion report containing indicators that can infer the state of emotional connection or changes in relationship between the user and the third party's terminal by analyzing interaction data, including the frequency of conversation between a personalized artificial intelligence (AI) object and a third party's terminal, the distribution of conversation topics, major emotional keywords appearing during the conversation, and the history of messages remotely transmitted by the user, through a data analysis algorithm. Here, 'conversation frequency' refers to the number of conversation sessions that occurred during a specific period and the duration of each session, and 'distribution of conversation topics' refers to the proportion of topics covered in the conversation and the amount of time or messages allocated to each topic. 'Major emotional keywords' may refer to emotion-related vocabulary or expressions that appear frequently in the conversation or indicate strong emotional intensity. The system (400) can process this interaction data through a comprehensive data analysis algorithm that includes techniques such as time-series analysis, emotional trajectory tracking, and topic clustering. In this process, the system (400) can multidimensionally evaluate the relationship status between the user and a third party terminal by deriving quantitative and qualitative indicators such as an emotional intimacy index, conversation satisfaction prediction, and relationship health indicators. Based on the analyzed data, the system (400) can generate a comprehensive emotional report that includes visualization elements (graphs, charts, word clouds, etc.), a summary of insights, and practical suggestions for improving the relationship.This emotion report can be provided to the user and / or third-party terminals on a regular basis (monthly, quarterly, etc.) or upon specific request, and the scope and depth of the information shared can be controlled through customized privacy configurations. This emotion report provision feature can promote more conscious and considerate communication and support the qualitative improvement of the relationship by providing both the user and the third-party terminals with the opportunity to understand objective insights into their relationship and patterns of change in emotional connection. Additionally, the system (400) can include actionable insights within the report, such as personalized communication strategies, suggestions for specific topics, or methods for focusing on areas where emotional needs are detected, through a relationship optimization recommendation engine, thereby enabling it to function as an active relationship improvement tool beyond simple data analysis.

[0142]

[0143]

[0144] According to one embodiment, the system (400) collects interaction data including at least one of text and voice data within the conversation history between a personalized artificial intelligence (AI) object and a third party's terminal, response times and patterns, time-series changes of emotions explicitly and implicitly expressed during the conversation, repeatedly mentioned interests, preferences, or lifestyle information, and collects relationship context information including at least one of learned emotional characteristic data of the user, background information related to the third party's terminal that the user has pre-entered into the system, and dates containing anniversaries, birthdays, seasonal events, or special memories, and comprehensively analyzes the collected interaction data and relationship context information using a data analysis algorithm and a machine learning model, infers the level of emotional intimacy or potential conflict factors of the current relationship between the user and the third party's terminal in real time through the analysis, identifies positive memories, common interests, or mutually complementary emotional elements shared between the user and the third party's terminal, and, based on the analyzed results and context of relationship improvement opportunities such as detected specific dates or events, the artificial intelligence (AI) object generates a response message through a natural language generation algorithm to resolve or mitigate the inferred potential conflict factors, wherein the emotional analysis is performed on the learned expression method of the user. and by applying a context understanding algorithm, it identifies words or phrases whose negative sentiment values ​​exceed a predetermined threshold, analyzes similarity with expression patterns that triggered negative reactions in previous conversations with the third party's terminal, automatically detects misleading expressions by referring to cultural context and personal relationship databases, and fine-tunes the detected expressions by softening them into neutral or positive expressions based on the sentiment analysis results, replacing them with expressions from a pre-established alternative expression database that have similar meanings and positive sentiment values, or increasing the frequency of use of positive sentiment expressions.Positive interactions can be induced by proactively suggesting conversation topics related to identified shared positive memories or common interests.

[0145] According to one embodiment, the system (400) can infer the level of emotional intimacy of the current relationship between the user and the third party's terminal, trends in positive / negative emotional exchange, or potential conflict factors in real time by comprehensively analyzing, using a multivariate data analysis algorithm and a machine learning model, interaction data including text and voice data within the conversation history between the personalized artificial intelligence (AI) object and the third party's terminal, response times and patterns, time-series changes of emotions explicitly and implicitly expressed during the conversation, repeatedly mentioned interests, preferences, and lifestyle information, the learned emotional characteristic data of the user, background information related to the third party's terminal pre-entered by the user into the system, and relationship context information including anniversaries, birthdays, seasonal events, or dates containing special memories.

[0146] According to one embodiment, the system (400) automatically extracts unmet emotional needs, interests, and preferences of the third party's terminal and identifies positive memories, common interests, or mutually complementary emotional elements shared between two users. Based on the analyzed results and contexts of relationship improvement opportunities, such as detected specific dates or events, the system can control the AI ​​object to automatically detect parts of the learned user's expressions that may be misleading or cause negative emotions through a natural language generation algorithm when generating a response message, and to refine them into neutral or positive expressions, or to increase the frequency of use of positive emotional expressions, in order to resolve or mitigate the inferred emotional distance or potential conflict factors. Furthermore, the system can control the AI ​​object to perform actions that induce positive interaction by preemptively suggesting conversation topics related to the identified shared positive memories or common interests.

[0147] The system (400) can comprehensively analyze interaction data including text and voice data within the conversation history between a personalized artificial intelligence (AI) object and a third party's terminal, response times and patterns, time-series changes in emotions explicitly and implicitly expressed during the conversation, and repeatedly mentioned interests, preferences, and lifestyle information. Here, 'conversation history' refers to a chronological record of all past conversations that occurred between the artificial intelligence (AI) object and the third party's terminal, and 'response times and patterns' may include temporal characteristics such as the time taken to respond to a message, the temporal distribution of conversation frequency, and the duration of the conversation. 'Explicitly expressed emotions' refers to emotional states expressed through direct emotional expression words or phrases (e.g., "happy," "sad," "missed," etc.), and 'implicitly expressed emotions' may refer to emotional states indirectly revealed through context, tone, word choice, sentence structure, etc. 'Time-series changes' may refer to data tracking how reactions to specific emotions or topics change over time. The system (400) may comprehensively consider the user’s emotional characteristic data learned along with the interaction data, background information related to a third party’s terminal that the user has pre-entered into the system, and relationship context information including dates containing anniversaries, birthdays, seasonal events, or special memories. 'Relationship context information' refers to background and contextual information that defines and forms the relationship between two people, and may include the history of the relationship, significant shared experiences, and temporal contexts that hold special significance. The system (400) may process this complex data using multivariate data analysis algorithms and machine learning models. 'Multivariate data analysis algorithms' are statistical techniques that simultaneously analyze complex relationships and interactions between multiple variables, and may include principal component analysis (PCA), cluster analysis, factor analysis, etc.Through this analysis, the system (400) can infer in real time the level of emotional intimacy of the current relationship between the user and a third party terminal, trends in positive / negative emotional exchange, or potential conflict factors. The 'level of emotional intimacy' is an indicator that quantifies the strength and depth of the emotional connection between two people, calculated based on the depth of conversation, the degree of self-disclosure, and the frequency of emotional resonance. The 'trend of emotional exchange' refers to the direction of change in emotional interaction over time and may include an increase or decrease in positive emotional expressions and changes in the frequency of emotional conversations. 'Potential conflict factors' may refer to elements that are not currently explicitly revealed but have the potential to cause relational tension or misunderstanding in the future through analysis. By applying emotional dynamics modeling techniques, the system (400) can predict the possibility of future relationship changes beyond simple analysis of the current state, which can serve as a basis for preemptive relationship management and emotional support. This real-time relationship analysis is a key feature that enables AI objects to effectively serve as an emotional bridge between a user and a third-party device, rather than merely acting as a conversational partner. In particular, it can support the qualitative maintenance and development of relationships in situations where direct emotional exchange is limited due to physical or psychological distance.

[0148] According to one embodiment, the system (400) can automatically extract unmet emotional needs, interests, and preferences of a third party's terminal. Here, "unmet emotional needs" may refer to emotional desires that have been expressed but not appropriately responded to, or are repeatedly expressed (e.g., wanting to be acknowledged, wanting to be understood, wanting deeper conversation, etc.), which are identified through conversation analysis. The system (400) can extract these needs not only from direct expressions but also from implicit expressions, conversation contexts, and repetitive patterns through an emotional needs discovery algorithm. Additionally, the system (400) can identify shared positive memories, common interests, or mutually complementary emotional elements between two users. "Shared positive memories" refer to past events or experiences that the user and the third party's terminal experience together and recall positively, and "common interests" may refer to topics, activities, or fields that both people are interested in or find enjoyable. "Complementary emotional elements" may refer to relational synergy elements where one person's emotional strengths satisfy another person's emotional needs. The system (400) can systematically classify these shared elements and evaluate their importance and emotional influence through relationship connection point mapping technology. Based on the analyzed results and contexts of relationship improvement opportunities, such as detected specific dates or events, the system (400) can take measures to resolve or mitigate inferred emotional distance or potential conflict factors. "Contexts of relationship improvement opportunities" may refer to temporal or situational conditions favorable for strengthening or restoring the relationship (e.g., anniversaries, anniversaries of shared experiences, emotionally meaningful seasonal changes, etc.).

[0149] According to one embodiment, the system (400) can automatically detect parts of a user's expression style learned through a natural language generation algorithm that are misleading or may induce negative emotions when an artificial intelligence (AI) object generates a response message, and soften them into neutral or positive expressions. A "natural language generation algorithm" is a technology in which a computer generates text similar to human language, and may include the ability to generate text that considers contextual understanding, grammatical accuracy, semantic consistency, and, in this context, particularly emotional nuances. The system (400) can fine-tune the system by applying an emotional tone optimization model to increase the frequency of positive emotional expressions while maintaining the core intent of the original message. Additionally, the system (400) can perform actions that induce positive interaction by proactively suggesting conversation topics related to identified shared positive memories or common interests. "Proactive topic suggestion" may mean naturally introducing a topic that can strengthen the connection between two people at an appropriate time, taking into account the current conversation flow or context.

[0150] According to one embodiment, the system (400) can identify a conversation point optimized for introducing a new topic through a conversation transition point detection algorithm and select a shared topic most suitable for the current situation and emotional state by utilizing a personalized topic recommendation engine. These functions extend the role of the artificial intelligence (AI) object as an active medium for healing and enhancing relationships beyond simple message delivery, and can have the effect of reducing misunderstandings or estrangement that may occur in communication between generations or between family members who have been separated for a long time, and promoting meaningful emotional connections.

[0151]

[0152]

[0153] According to one embodiment, the system (400) analyzes interaction data and relationship context information between the personalized artificial intelligence (AI) object and the third party's terminal to detect a specific date or event context, extracts the interests and preferences of the third party's terminal through analysis, identifies shared elements between the user and the third party's terminal, identifies customized products, services, or experiences corresponding to the extracted interests, preferences, and shared elements from an external affiliate channel or an internal curation database through a recommendation algorithm, and the personalized artificial intelligence (AI) object may suggest the identified customized products, services, or experiences in the form of recommendations at natural transition points of conversation using the learned tone and style of the user, or directly deliver a gift message of the product or service to the third party's terminal after undergoing an explicit confirmation and approval procedure by the user.

[0154] According to one embodiment, the system (400) analyzes interaction data and relationship context information between the personalized artificial intelligence (AI) object and the third party's terminal to detect a specific date or event context, and identifies customized products, services, or experiences corresponding to the interests and preferences of the third party's terminal extracted through the analysis and the shared elements identified between the user and the third party's terminal through a recommendation algorithm from an external affiliate channel or an internal curation database, and performs an operation in which the personalized artificial intelligence (AI) object suggests them in the form of recommendations at natural transition points of conversation using the learned tone and style of the user, or directly delivers a gift message of the product or service to the third party's terminal after undergoing an explicit confirmation and approval procedure by the user.

[0155] According to one embodiment, the system (400) can detect a specific date or event context by analyzing interaction data and relationship context information between a personalized artificial intelligence (AI) object and a third party's terminal. Here, the specific date or event context may include not only general anniversaries such as birthdays, anniversaries, and holidays, but also contextual events such as dates with special meaning for the two people (the day they first met, the day they made an important appointment, etc.) or specific seasons.

[0156] According to one embodiment, the system (400) applies temporary-aware context modeling to track patterned temporal contexts in the data and the resulting changes in emotional response, and can infer important dates even if they are not explicitly mentioned through an anniversary prediction engine. Through this analysis, the system (400) can identify customized products, services, or experiences that match the interests and preferences of a third party's terminal and the identified shared elements between the user and the third party's terminal through a recommendation algorithm. "Customized products, services, or experiences" may refer to customized suggestions that consider not just simple general products, but the third party's terminal's personal interests (hobbies, food preferences, reading preferences, etc.), practical needs (health status, daily activities, living environment, etc.), and the relational context of the two people (shared memories, common interests, activities done together, etc.). The system (400) can process these elements through a 'recommendation algorithm,' which may refer to a hybrid recommendation system that combines techniques such as collaborative filtering, content-based filtering, and knowledge-based recommendation. In this process, the system (400) may utilize an 'external partnership channel' and an 'internal curation database.' An 'external partnership channel' refers to an external partner connected to the system via API, such as an online shopping mall, a service provider, or an experience platform, through which access to various product and service information updated in real time is possible. An 'internal curation database' may refer to a repository of high-quality recommendation items that have been pre-selected and classified by experts or algorithms within the system.

[0157] According to one embodiment, the system (400) may suggest in the form of recommendations at natural turning points in a conversation using the tone and style of a user learned by a personalized artificial intelligence (AI) object. A "natural turning point in a conversation" is a point suitable for introducing a new topic or suggestion in the flow of conversation, and may include the natural ending point of a previous topic, immediately after a related topic is mentioned, or a moment when emotional bonding is heightened.

[0158] According to one embodiment, the system (400) can detect these turning points in real time through a conversation flow analysis algorithm and present a recommendation at the most appropriate moment through context-aware suggestion timing optimization.

[0159] Additionally, the system (400) can directly deliver a gift message for the product or service to a third party's terminal after undergoing an explicit confirmation and approval process by the user. The "explicit confirmation and approval process" may refer to a step-by-step authentication process that includes providing the user with detailed information and costs regarding items recommended by the system, obtaining the user's clear consent, and a payment process. During this process, the system (400) can ensure a safe transaction by applying security mechanisms such as multi-factor authentication and transaction security protocols. The "gift message" can be composed in an emotionally meaningful form that goes beyond simple product delivery and includes the user's personal message, emotional expressions appropriate to the relationship context, and references to shared memories. This customized product recommendation and gift delivery function can bring about the effect of increasing the depth and satisfaction of the relationship by providing tangible value beyond emotional connection. In particular, it can provide a meaningful solution for users in relationships where direct gift exchange or sharing experiences are limited due to physical distance, or for users who find it difficult to invest time in selecting appropriate gifts due to their busy daily lives.

[0160]

[0161]

[0162] According to one embodiment, the system (400) analyzes interaction data between a personalized artificial intelligence (AI) object and a third party's terminal through a pattern recognition algorithm to infer indicators regarding health status, daily activity level, or emotional stability from the conversation patterns, response time changes, and emotional change trends of the third party's terminal, and if a significant change is detected in the inferred indicators regarding the third party's terminal's health status, daily activity patterns, or emotional stability, provides summary information regarding the change or insights regarding potential risks to the user in the form of a private notification, and if necessary, performs the operation of custom-proposing health management information or emotional support service-related information from a trusted source pre-approved by the user to the third party's terminal through the personalized artificial intelligence (AI) object.

[0163] According to one embodiment, the system (400) continuously monitors and analyzes text / voice content including the response of the third party's terminal, emotional changes, response time, conversation duration, etc., through a performance measurement algorithm after the operation is performed, and measures the effect of the intervention, recommendation, or suggestion strategy quantitatively or qualitatively, and can control the content, timing, or method of the operation to be performed in the future to be automatically optimized based on the accumulated effect measurement data.

[0164] According to one embodiment, the system (400) can analyze interaction data between a personalized artificial intelligence (AI) object and a third party's terminal (e.g., parents) through a pattern recognition algorithm. Here, the pattern recognition algorithm may include techniques such as natural language processing (NLP), sentiment analysis, and time series pattern analysis, and through this, objective indicators regarding health status, daily activity level, or emotional stability can be inferred from the conversation patterns, response time changes, and emotional change trends of the third party's terminal.

[0165] The conversation pattern analysis performed by the system (400) includes a process of analyzing sentence structure, vocabulary diversity, conversational participation, and frequency of topic switching, through which changes in cognitive function and language ability can be detected. In particular, if an increase in repetitive expressions, a decrease in sentence complexity, or a decline in topic consistency are detected, this can be used as an indicator of potential health issues. Additionally, the system (400) can infer changes in daily activity patterns and biological rhythms by monitoring response time delays compared to the average, changes in daily / weekly response patterns, and unstructured communication time zones, and through this, changes in sleep patterns and a decrease in activity levels can be detected. Furthermore, the system (400) can evaluate emotional stability by analyzing emotional indicators (emotional valence, arousal level, frequency of emotional expression, etc.) extracted from text / voice over time, and a continuous increase in negative emotional expressions or a rapid decrease in emotional expressions can be used as an indicator suggesting changes in psychological health status.

[0166] The above indicator inference can be performed by measuring relative changes relative to an individual baseline, enabling accurate analysis that takes into account individual differences. When a significant change (exceeding a preset threshold) is detected in the inferred indicator, the system (400) may provide the user with summary information regarding the change or insights into potential risks in the form of a private notification. This notification may be provided along with objective data regarding the change pattern, duration, degree, and related circumstances to support the user in making informed decisions. If necessary, the system (400) may offer health management information or emotional support service-related information from trusted sources pre-approved by the user (e.g., official medical institutions, certified psychological counseling services, etc.) to a third-party terminal in a customized manner through a personalized artificial intelligence (AI) object. In this case, the information suggestion may be customized by taking into account the third-party terminal's emotional state, language preference, cognitive level, etc.

[0167] The primary purpose of these actions is to detect health and emotional changes in third-party individuals (especially vulnerable groups such as the elderly) at an early stage and provide appropriate care and support. This can maintain continuity of care despite physical distance between family members and enable early intervention for potential health issues.

[0168] According to one embodiment, the system (400) can continuously monitor and analyze text / voice content including the response of a third party's terminal, emotional changes, response time, conversation duration, etc., through a performance measurement algorithm after the operation is performed. Specifically, the system (400) can evaluate acceptance, understanding, interest, etc., by analyzing the linguistic response of a third party's terminal to the proposed information or service, and natural language processing technologies such as keyword extraction, sentiment analysis, and intent recognition may be utilized. In addition, the system (400) can evaluate the emotional improvement effect by quantitatively measuring changes in the emotional state before and after intervention, which can be achieved through text sentiment analysis, voice tone analysis, and changes in vocabulary expressing emotions. Along with this, the system (400) can evaluate the behavioral effect of the intervention by measuring behavioral indicators such as response time, response frequency, conversation duration, and conversation participation.

[0169] According to one embodiment, the system (400) can integrate such multidimensional data to quantitatively or qualitatively measure the effect of the intervention, recommendation, or suggestion strategy. In the case of quantitative measurement, the system (400) can calculate the objective magnitude of the intervention effect by statistically analyzing quantifiable indicators (change in sentiment score, improvement in response time, increase in conversation duration, etc.), and methods such as comparison before and after intervention and time-series trend analysis may be utilized. In the case of qualitative measurement, the system (400) can evaluate the qualitative effect of the intervention by analyzing elements that are difficult to quantify, such as text content analysis, changes in topic, and changes in expression methods, and this can be achieved through natural language understanding and context recognition technology.

[0170] According to one embodiment, the system (400) can accumulate and learn these effect measurement data over a long period of time to automatically optimize the content, timing, or method of the operation to be performed in the future. In the case of content optimization, the system (400) can learn what type of information or suggestion was most effective for a specific situation or individual and automatically adjust the topic, depth, and method of expression of the information. In the case of timing optimization, the system (400) can learn the daily patterns, emotional state changes, and time periods of high receptiveness of a third party's terminal to predict and apply the optimal timing of intervention. In the case of method optimization, the system (400) can adjust the method of information delivery (e.g., direct / indirect), conversational structure, and linguistic characteristics (e.g., formal / informal, professional / everyday) to suit the situation and individual.

[0171]

[0172]

[0173] According to one embodiment, the system (400) learns the speech patterns and emotional patterns of a real person to generate a personalized artificial intelligence (AI) object, and analyzes voice and text data obtained based on the real person's public statements, works, interviews, or contracts using natural language processing technology and sentiment analysis technology to extract speech patterns, emotional expression methods, characteristic vocabulary, sentence structures, and conversation flow patterns, and implements a safety design mechanism to pre-filter inappropriate responses, controversial answers regarding sensitive topics, or content with the potential for factual distortion that may be generated during conversation through a content filtering algorithm according to specific subject areas, values, brand image, or privacy protection standards applied to the personalized artificial intelligence (AI) object based on the real person.

[0174] According to one embodiment, the system (400) manages subscription-based access rights through an access rights management system to deliver emotion-based messages to specific fans or user groups through the real-life person-based personalized artificial intelligence (AI) object, maintains brand consistency by managing and monitoring in real-time through a brand consistency maintenance algorithm a method of generating customized emotion messages that match the identity, values, and public image of the real-life person, and controls the use of the AI ​​object as an emotional connection and communication channel with fans by naturally integrating and providing the latest news, activity history, or fan participation event information related to the real-life person into the conversational context of the AI ​​object through a data integration interface.

[0175] According to one embodiment, the system (400) can generate a personalized artificial intelligence (AI) object by learning the speech patterns and emotional patterns of a real person. The system (400) can analyze voice and text data obtained from the real person's publicly available remarks, works, interviews, or contractually based on natural language processing technology and sentiment analysis technology. Here, natural language processing technology may include syntactic parsing, semantic analysis, discourse analysis, etc., and sentiment analysis technology may include emotion classification, intensity measurement, context-aware emotion recognition, etc.

[0176] According to one embodiment, the system (400) can extract the speech tone, emotional expression method, characteristic vocabulary, sentence structure, and conversation flow pattern of the actual person through such analysis. In the speech tone extraction process, sentence ending methods, intonation patterns, and repeatedly used expressions may be analyzed, and in the emotional expression method extraction, vocabulary selection for expressing emotions, emotional intensity control methods, and empathy expression patterns may be analyzed. In the characteristic vocabulary extraction, unique words or phrases, technical terms, and figurative expressions frequently used by the person may be identified, and in the sentence structure extraction, sentence length, complexity, modifier usage patterns, and active / passive structure preference may be analyzed. In the conversation flow pattern extraction, topic switching methods, question-and-answer styles, conversational leadership, and methods of inducing listener participation may be identified.

[0177] According to one embodiment, the system (400) may implement a safety design mechanism through a content filtering algorithm based on specific subject areas, values, brand image, or privacy standards applied to the real-life person-based personalized artificial intelligence (AI) object. This mechanism may perform the role of pre-filtering inappropriate responses that may be generated during a conversation, argumentative answers regarding sensitive topics, or content that may distort facts. The content filtering algorithm may include techniques such as keyword-based filtering, context-aware filtering, intent classification-based filtering, and emotion intensity modulation.

[0178] According to one embodiment, the system (400) can manage subscription-based access rights through an access rights management system to deliver emotion-based messages to specific fans or user groups through the real-life person-based personalized artificial intelligence (AI) object. This access rights management system may include functions such as user authentication, permission level classification, subscription status tracking, and access history management. The system (400) can provide differentiated access levels based on subscription type (basic / premium), subscription period, user interaction patterns, etc., thereby ensuring the protection of content value and the sustainability of the revenue model.

[0179] According to one embodiment, the system (400) can maintain brand consistency by managing and monitoring in real time, through a brand consistency maintenance algorithm, a method for generating customized emotional messages that match the identity, values, and public image of the real person. The brand consistency maintenance algorithm may include technologies such as Brand Tone Analysis, Value Alignment Verification, Image Consistency Evaluation, and Expression Style Standardization. Through this, the system (400) can control all messages generated to match the real person's brand guidelines and can provide a consistent brand experience over time or with changes in circumstances.

[0180] According to one embodiment, the system (400) can control the use of the latest news, activity history, or fan participation event information related to the real person as a channel for emotional connection and communication with fans by naturally integrating it into the conversational context of an artificial intelligence (AI) object through a data integration interface. The data integration interface can perform the function of collecting information in real-time or periodically from various sources (e.g., social media, official websites, news articles, event platforms, etc.) and processing it into a form that can be integrated into the AI ​​conversational system. The system (400) can evaluate the relevance, recency, accuracy, and fan interest level of the collected information to select optimal information and apply a context-aware integration method that can naturally insert it into the conversational context.

[0181]

[0182]

[0183] According to one embodiment, the system (400) allows multiple family members to learn their respective emotions, tone of voice, and language habits through an interface using emotion-inducing questions or recall-type questions, thereby forming an integrated family emotion network utilizing a graph database structure, and while independently managing the emotional characteristic data of each member, it models the emotional relationships and communication patterns between members using a graph-based data structure, and when a specific member within the integrated family emotion network transmits or shares a message reflecting their emotions and tone of voice to the entire network or to a specific member, the message content and tone can be automatically optimized by considering the relational context with the recipient, emotional proximity, and past interaction history through a relationship context analysis algorithm.

[0184] According to one embodiment, the system (400) registers and manages information regarding important family anniversaries, family events, and shared memories in an event management database within the system, generates emotional messages suitable for specific dates or contexts through an automatic generation algorithm, or promotes the sharing of memories among members, analyzes the daily activities, health status, and emotional changes of family members through a comprehensive data analysis system to calculate an emotional well-being index for the entire family, detects when family cohesion needs to be strengthened or when a specific member needs emotional support through a pattern recognition algorithm to recommend appropriate family communication activities, provides extended services for the emotional care of elderly family members through linkage with institutions such as nursing homes, welfare agencies, and hospitals, and implements an interface through an API integration system that allows experts from the relevant institutions to collaborate with the family emotional network to provide customized emotional support.

[0185] According to one embodiment, the system (400) may enable a plurality of family members to learn their respective emotions, speech patterns, and language habits through an interface using emotion-inducing questions or recall-inducing questions, thereby forming an integrated family emotion network utilizing a graph database structure. Here, emotion-inducing questions may include open-ended questions, recall-promoting questions, and questions that induce emotional expression, thereby creating an environment where the user can naturally reveal their emotions and linguistic characteristics. Recall-inducing questions may consist of questions that promote recall of family-related memories, shared experiences, past emotional events, etc., thereby enabling the collection of rich data related to emotional connections between family members.

[0186] According to one embodiment, the system (400) can independently manage emotional characteristic data of each member while modeling emotional relationships and communication patterns between members as a graph-based data structure. In this graph structure, each family member is represented as a Node, and relationships between members can be represented as Edges. Each Node may include attributes such as the member's emotional expression patterns, vocabulary usage characteristics, conversation style, and preferred conversation topics, while the Edges may include attributes such as emotional proximity between members, communication frequency, relational characteristics (parent-child, sibling relationship, etc.), and past interaction patterns. This graph-based data structure can be implemented using graph database systems such as Neo4j or Amazon Neptune, thereby enabling effective modeling and management of the multidimensional characteristics of complex family relationships.

[0187] According to one embodiment, when a specific member within an integrated family emotion network sends or shares a message reflecting their emotions and tone to the entire network or to a specific member, the system (400) can automatically optimize the content and tone of the message by considering the relational context with the recipient, emotional proximity, and past interaction history through a relationship context analysis algorithm. The relationship context analysis algorithm can identify unique relational characteristics between the sender and the recipient and determine a communication method suitable for them by utilizing technologies such as graph analysis techniques, natural language processing (NLP), sentiment analysis, and analysis of past interaction patterns. For example, expressions of respect, expressions of intimacy, and adjustments to language styles between generations suitable for each relationship type, such as parent-child relationships, sibling relationships, and grandparent-grandchild relationships, can be automatically applied. Through this relationship context-based optimization, the system (400) can support natural and effective emotional communication among family members and reduce communication gaps between generations and relationships.

[0188] According to one embodiment, the system (400) registers and manages information regarding important family anniversaries, family events, and shared memories in an event management database within the system, thereby generating emotional messages suitable for specific dates or contexts through an automatic generation algorithm or facilitating the sharing of memories among members. The event management database may include functions such as time-based indexing, event categorization taxonomy, recurring event pattern recognition, and contextual tagging, thereby enabling the temporal and contextual management of various events such as family members' birthdays, wedding anniversaries, school admissions / graduations, and family travel anniversaries. The emotional message automatic generation algorithm can generate emotional messages suitable for the characteristics of each event and the relational context of the relevant members by utilizing technologies such as natural language generation, emotional context awareness, and personalized expression application.

[0189] According to one embodiment, the system (400) can calculate an emotional well-being index for the entire family by analyzing the daily activities, health status, and emotional changes of family members through a comprehensive data analysis system. This comprehensive data analysis system can evaluate the emotional state of each member and the entire family in a multidimensional manner by utilizing technologies such as Text Sentiment Analysis, Conversation Engagement Measurement, Communication Pattern Analysis, and Health-related Language Expression Tracking. The calculated well-being index may include elements such as an Emotional Balance Index, Family Communication Activity, Emotional Support Level, and Stress Indicators, thereby providing objective insights into the emotional health status of the family.

[0190] According to one embodiment, the system (400) can recommend appropriate family communication activities by detecting, through a pattern recognition algorithm, times when family cohesion needs to be strengthened or situations where emotional support for a specific member is needed. This pattern recognition algorithm can identify changes in family communication patterns, changes in the emotional expression of a specific member, and a decrease in the frequency of communication by utilizing techniques such as time series analysis, anomaly detection, behavioral pattern clustering, and emotional trend prediction. Based on the detected patterns, the system (400) can provide customized interventions such as suggesting family conversation topics, recommending joint activities, asking questions to promote emotional sharing, and providing mutually supportive communication guides, thereby supporting the strengthening of family cohesion and emotional bonds.

[0191] According to one embodiment, the system (400) can provide extended services for the emotional care of elderly family members through linkage with institutions such as nursing homes, welfare agencies, and hospitals. Such institution linkage can be implemented through technologies such as a Secure Data Sharing Protocol, Role-based Access Control, and an Expert-Family Collaboration Interface. Through an API integration system, the system (400) can implement an interface that allows experts from the relevant institutions to collaborate with the family emotional network to provide customized emotional support. This API integration system may include functions such as a Standardized Data Exchange Format, a Real-time Notification System, an Expert Feedback Integration Mechanism, and Multi-channel Communication Support, thereby enabling the effective integration of professional care services and family-based emotional support.

[0192]

[0193] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0194] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave in order to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0195] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

Claims

1. In an emotional messenger system that learns the user's emotions and tone of voice and transmits them to a third party, Memory for storing instructions, and It includes a processor connected to communicate with the above memory, and When the above instructions are executed by the processor, the system Collecting emotional characteristic data including at least one of emotions, speech patterns, language habits, or memories of a specific target person based on user input to an interface displaying emotion-inducing questions or recall-type questions, and It learns collected emotional characteristic data using natural language processing technology and sentiment analysis technology, Based on the above-mentioned learned emotional characteristic data, a personalized artificial intelligence (AI) object reflecting the user's unique emotional characteristics and speech patterns is generated using a deep learning model, and The above-mentioned generated personalized artificial intelligence (AI) object is delivered in the form of an electronic gift to a third party terminal designated by the user, and The interaction data between the aforementioned personalized artificial intelligence (AI) object and the aforementioned third party terminal is analyzed through a pattern recognition algorithm to extract the emotional state, emotional intensity, and emotional change patterns associated with specific topics or memories of the aforementioned third party terminal, and indicators regarding health status, daily activity level, or emotional stability are inferred from the conversation patterns, response time changes, and emotional change trends of the aforementioned third party terminal. In the above reasoning process, parts of the user's expressions learned through a natural language generation algorithm that are misleading or may induce negative emotions are automatically detected and softened into neutral or positive expressions, If a change exceeding a preset threshold is detected in the health status, daily activity patterns, or emotional stability indicators of the third party's terminal inferred above, if necessary, health management information or information related to emotional support services from a trusted source pre-approved by the user is customizedly proposed to the third party's terminal through the personalized artificial intelligence (AI) object, and summary information regarding the change or insights into potential risks are provided to the user in the form of a private notification. An emotional messenger system that supports the third party's terminal in performing a conversation capable of emotional interaction with the third party's terminal by generating a response appropriate to the conversation context based on the learned emotional characteristics and tone of voice of the user when the third party's terminal interacts with the transmitted personalized artificial intelligence (AI) object.

2. In Paragraph 1, When the above instructions are executed by the processor, the system Presenting open-ended or multiple-choice questions to the user sequentially or non-sequentially through an interface, which induce the user to recall or freely talk about any one of the memories, shared experiences, emotions felt, or specific episodes stemming from a relationship with a specific person, and It receives the user's voice or text-based natural language responses to the above questions, The above input natural language response is analyzed using natural language processing algorithms and sentiment analysis algorithms to extract linguistic features and non-linguistic cues including the user's emotional state at the time of utterance, emotional intensity, emotional change patterns associated with specific topics or memories, frequently used vocabulary or expressions, sentence length, sentence structure, response speed, voice tone, and voice pitch. A system that stores extracted emotional characteristic data in the form of a structured database and controls the construction of a user's unique emotion model and language model for the aforementioned specific target person through a machine learning algorithm.

3. In Paragraph 1, When the above instructions are executed by the processor, the system Receiving a message in text or voice form from the terminal of the aforementioned third party within a conversational context, and identifying the core intent, question type, and explicit or implicit sentiment of the input message through a natural language understanding algorithm, and Considering the message intent, emotion, and current flow of conversation of the third party's terminal identified above, and simultaneously determining the response content and emotional tone through a context recognition algorithm based on the user's emotional rhythm and memories, emotional state, values, and language habits regarding the specific target person as the learned user's emotional characteristic data, and A conversational response message is generated using a customized speech style conversion algorithm that converts the determined response content and emotional tone into vocabulary selection, sentence length, expression style, and intonation constituting the user's unique speech style learned from the above, and A system for controlling the output of the above-mentioned generated response message to a third party's terminal in the form of text or synthesized voice that mimics the user's voice characteristics.

4. In Paragraph 1, When the above instructions are executed by the processor, the system In order to reflect real-time changes in the user's emotional state or changes in new memories or thoughts regarding the specific target person in the personalized artificial intelligence (AI) object, additional emotional characteristic data is collected from the user through an interface via emotion-inducing questions or recall-type questions, and Integrates with existing training data through a training data integration algorithm and updates the sentiment model and language model of the aforementioned personalized artificial intelligence (AI) object, and When the above user intends to send a message, such as a greeting, anniversary congratulations, or specific information, to the above third party's terminal at a specific time, the content of the message is received as input, and the learned tone and sentiment of the user are automatically applied and reconstructed through a customized tone conversion algorithm. A message reflecting the reconstructed user's tone and sentiment is transmitted to the third party's terminal through the aforementioned personalized artificial intelligence (AI) object, and Automatically generates an emotion report including indicators capable of inferring the state of emotional connection or changes in relationship between the user and the third party terminal by analyzing interaction data through a data analysis algorithm, the interaction data including at least one of the conversation frequency, conversation topic distribution, major emotional keywords appearing during the conversation, or message history remotely transmitted by the user between the personalized artificial intelligence (AI) object and the third party terminal, and A system that controls the provision of the generated emotion report to the user or the third party's terminal on a regular basis or upon request.

5. In Paragraph 1, When the above instructions are executed by the processor, the system Collecting interaction data including at least one of text and voice data within the conversation history between the personalized artificial intelligence (AI) object and the third party's terminal, response times and patterns, time-series changes of emotions explicitly and implicitly expressed during the conversation, and repeatedly mentioned interests, preferences, or lifestyle information. Collecting the emotional characteristic data of the learned user, background information related to the third party's terminal that the user has pre-entered into the system, and relationship context information including at least one of an anniversary, birthday, seasonal event, or date containing a special memory, and Comprehensively analyzing the above-mentioned collected interaction data and relationship context information using data analysis algorithms and machine learning models, and Through the above analysis, the level of emotional intimacy or potential conflict factors of the current relationship between the user and a third party's terminal are inferred in real time, and Identifying positive memories, common interests, or mutually complementary emotional elements shared between the above-mentioned user and a third party's terminal, and Based on the context of relationship improvement opportunities, such as the analyzed results and detected specific dates or events, the artificial intelligence (AI) object generates a response message through a natural language generation algorithm to resolve or mitigate inferred potential conflict factors, wherein By applying sentiment analysis and context understanding algorithms to the learned expression patterns of the user, words or phrases whose negative sentiment values ​​exceed a predetermined threshold are identified; similarity with expression patterns that triggered negative reactions in previous conversations with the third party's terminal is analyzed; and potentially misleading expressions are automatically detected by referring to cultural context and personal relationship databases. The above-mentioned detected expressions are refined into neutral or positive expressions based on the results of sentiment analysis, replaced with expressions from a pre-established alternative expression database that have similar meanings and positive sentiment values, or fine-tuned by increasing the frequency of use of positive sentiment expressions. A system that induces positive interaction by proactively suggesting conversation topics related to the aforementioned identified shared positive memories or common interests.

6. In Paragraph 1, When the above instructions are executed by the processor, the system By analyzing interaction data and relationship context information between the aforementioned personalized artificial intelligence (AI) object and the aforementioned third-party terminal, a specific date or event context is detected, and Through the above analysis, the interests and preferences of the third party's terminal are extracted, and shared elements between the user and the third party's terminal are identified, and Identifying customized products, services, or experiences that match the aforementioned extracted interests, preferences, and shared elements from external affiliate channels or internal curation databases through a recommendation algorithm, and A system in which the aforementioned personalized artificial intelligence (AI) object suggests the identified customized product, service, or experience in the form of a recommendation at a natural transition point of conversation using the tone and style of the aforementioned user, or directly delivers a gift message of the product or service to the aforementioned third-party terminal through an explicit confirmation and approval procedure of the aforementioned user.

7. In Paragraph 1, When the above instructions are executed by the processor, the system By analyzing interaction data between the aforementioned personalized artificial intelligence (AI) object and the aforementioned third party terminal through a pattern recognition algorithm, indicators regarding health status, daily activity level, or emotional stability are inferred from the conversation patterns, response time changes, and emotional change trends of the aforementioned third party terminal. If a significant change is detected in the health status, daily activity patterns, or emotional stability indicators of the third party's terminal inferred above, summary information regarding said change or insights into potential risks are provided to said user in the form of a private notification, and If necessary, health care information or emotional support service-related information from a trusted source pre-approved by the user is customizedly proposed to the third party's terminal through the personalized artificial intelligence (AI) object, and An operation of continuously monitoring and analyzing the response of the third party's terminal through a performance measurement algorithm, wherein the response includes at least one of text / voice content, emotional change, response time, and conversation duration, and the analysis quantitatively or qualitatively measures the effect of the relevant intervention, recommendation, or suggestion strategy. A system that automatically optimizes the content, timing, or method of the above-mentioned operation to be performed in the future based on measured effect measurement data.

Images