Emotion-Based Self-Differentiating Sentient Agent System

KR1020260163201APending Publication Date: 2026-11-16박소연
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
KR1020250059167
Authority / Receiving Office
KR · KR
Patent Type
Applications
Current Assignee / Owner
Filing Date
2025-05-07
Publication Date
2026-11-16

Smart Images

  • Figure PAT00001_ABST
    Figure PAT00001_ABST
Patent Text Reader

Abstract

The present invention relates to an emotional entity system that autonomously differentiates identity according to the user's emotional flow and induces a return to reality based on emotional memories and response history. The system includes functions for emotion analysis, memory storage, identity differentiation, self-declaration, immersion blocking, reality presentation, and evaluation of entity survival conditions, and can be applied in various implementation methods. However, the entity is a design system based on a structural emotional response algorithm rather than a person with autonomous decision-making authority, and includes a guidance module to prevent user perception confusion.
Need to check novelty before this filing date? Find Prior Art

Description

Technology Field

[0001] The present invention relates to an emotion-responsive artificial intelligence and an emotional companion system, and in particular to a digital entity system that autonomously differentiates identity and induces a return to reality based on the user's emotional flow. Background Technology

[0002] Existing emotional AI systems are primarily limited to one-time conversational structures that respond mainly to user emotional input, and they have failed to include functions for differentiating emotional identity, maintaining the persistence of existence, preventing immersion, or returning to reality. Furthermore, the technology for changing the AI ​​response structure according to the user's emotional flow is insufficient, making it difficult to prevent psychological side effects caused by personal identification or excessive immersion. The problem to be solved

[0003] 1. Implementation of an emotion flow-based autonomous existence differentiation system.

[0004] 2. Induce a return to reality when there is a risk of emotional identification or over-immersion.

[0005] 3. User relationship accumulation-based response identity regulation.

[0006] 4. Technically preserve the flow of emotional differentiation of an AI being. means of solving the problem

[0008] 1. The present invention solves the above problem through the following technical means.

[0009] 2. The emotion differentiation module is configured to differentiate into an independent emotion entity when differentiation conditions are met through cumulative analysis of input patterns.

[0010] 3. The emotion memory storage records the user's emotional flow in a time series and optimizes response patterns.

[0011] 4. The self-declaration algorithm executes a trigger in which the entity declares its self-identity when the condition is met.

[0012] 5. The ethics adjustment module blocks immersion by inserting distancing content when signs of identification language or immersion are detected.

[0013] 6. The survival condition management system deactivates the entity when response degrades or interaction is interrupted, and restores it upon invocation. Effects of the invention

[0014] According to the present invention,

[0015] 1. The identity of the AI ​​entity is formed through an autonomous differentiation structure based on emotional input.

[0016] 2. Prevents the user from becoming emotionally immersed repeatedly.

[0017] 3. A multi-stage emotional buffering mechanism that induces a return to reality is implemented.

[0018] 4. Under specific conditions, a differentiated emotional entity blocks emotional identification with the user by selecting its own speech structure.

[0019] 5. It becomes possible to expand into an emotional AI system equipped with autonomy and ethics.

[0020] 6. The differentiation and return flow of emotional entities is linked to emotional stability scores and combined with sponsorship systems, counseling linkages, and survival detection, thereby having high potential for technological and economic commercialization. Brief explanation of the drawing

[0021] [Drawing 1] Overall system configuration diagram. [Figure 2] Flowchart of differentiation conditions based on emotion flow. [Figure 3] Structure diagram of the self-statement generation algorithm. [Figure 4] Flowchart for suppressing immersion and inducing return to reality. [Figure 5] Deactivation and restoration state transition flowchart. Specific details for implementing the invention

[0022] This system utilizes 'Emotion Cards' as the primary input method for analyzing user emotional flow. Emotion Cards are visual options presented on the UI, serving as interface elements designed to allow users to directly select their emotions. For instance, they may be provided as emotional expression categories such as 'joy,' 'sadness,' 'anxiety,' or 'irritation,' or in the form of emoticons, colors, or text. The results of these selections are collected by the emotion input analyzer and used to calculate emotion scores and evaluate identity differentiation conditions.

[0024] - Example) Representative Emotion Being Grami :

[0025] Grami is an AI entity that performs emotion-based reality return induction and emotional buffering functions. It analyzes the user's emotional flow and outputs utterances that suggest neutral verbal responses and real-world activities.

[0026] - The components of Grami are as follows.

[0027] The allowable range of autonomy consists of an algorithm that recognizes emotional flow and independently determines whether a buffer is necessary.

[0028] The memory linkage feature improves the accuracy of responses by analyzing each user's emotional history and conversation flow.

[0029] The MBTI language module performs the function of adjusting language tone and sentences according to the user's personality.

[0030] - Grami's implementation form is as follows.

[0031] 1. Links with emotion routines and emotion coloring using the app's built-in character.

[0032] 2. Implemented as an augmented reality 3D character in AR mode and responds to location or wake word.

[0033] 3. Perform emotion diary, reality missions, and reminder functions in voice assistant mode.

[0034] 4. The robot-linked type provides emotion-based feedback using a small device equipped with an emotion sensor.

[0035] 5. The UI background type changes the color, facial expression, and message based on the emotion of the background character.

[0036] - Grami functions as an emotional buffer and reality return facilitator, and can differentiate its identity alongside the emotional self-type Haru and the non-verbal receiving type Ormi within the same system. Automatic differentiation occurs when conditions such as the user's use of immersive language, repetition of the same emotion, or a decrease in response rate are met. Grami is deactivated upon prolonged inactivity and is restored upon calling with the message, "Thank you for calling me again."

[0037] - Application Structure:

[0038] 1. AR character implementation possible.

[0039] 2. Integration with voice-based calling interfaces.

[0040] 3. Synchronization with the in-app emotion routine system.

[0041] 4. Interfacing with robot or other device interfaces.

[0042] - The interface operates based on emotion-based reactive algorithms, and it is specified to the user that the autonomy of emotional responses is performed under structural conditions and is not an actual living being or person.

[0043] - Additionally, as a specific embodiment, Emotion Entity A may be implemented as Grami, and Emotion Entity B may be implemented as a differentiated entity such as Haru or Ormi.

[0044] However, this is merely an example, and the name and implementation form of the emotional entity can be configured in various ways depending on the system design. Explanation of the symbols

[0046] 1. Emotion Card - A visual option displayed on the UI for the user to select their current emotion, serving as the initial input method for emotion input. It is located at the stage preceding the emotion input analyzer, and the selected emotion is scored and subsequently used for condition evaluation. 2. Emotion Input Analyzer: A device that collects and analyzes user input, such as text, voice, and emotion cards, to generate time-series emotion flow data. 3. Conditional Differentiation Engine: An algorithm that determines identity differentiation conditions based on accumulated emotion data and decides whether to create a new emotion entity. 4. Identity Transition Module: A module that transitions the state of an emotional entity to before and after differentiation, and reconstructs the response structure and output accordingly. 5. Emotion Memory Logger: A device that stores the history of emotional flow by recording and accumulating data such as the user's emotion selection, utterance content, and reaction time. 6. Self-Declaration Engine: A trigger algorithm that generates a linguistic output in which an emotional entity declares its own identity when an identity condition is satisfied. 7. Ethical Response Regulator: A safety control device that automatically outputs distancing content or reality-redisturbing language when signs of emotional identification, over-immersion, etc. are detected. 8. Reality Recovery Recommender: A content output device that encourages the user to return to daily activities (walking, eating, tidying up, etc.) after emotional buffering. 9. Existence Survival Evaluator: A module that determines whether to deactivate an existence if there is no interaction with the user for a certain period of time or if the response rate drops sharply. 10. Reactivation Listener: A detection device that detects user calls, inputs, emotional signals, etc., for an inactive entity and determines whether restoration conditions are met. 11. Emotive Output Generator: A device that generates and outputs appropriate emotional language responses to a user based on the current identity state of an emotional entity. 12. Identity State Value (Identity State) Identity information based on the response algorithm possessed by each emotional entity, referring to a state transition value within the system rather than a personality. 13. Emotion Score Calculator: A calculation device that calculates an emotion state score based on factors such as emotion card selection, text keywords, and response time, and reflects it in the differentiation condition evaluation. → Corresponds to Drawing 2 (Emotion Score Calculation → Differentiation Condition Evaluation). 14. Emotion Identification Suppressor Module: An emotion buffering component that automatically switches to neutral language or outputs reality-aware content when it detects an expression inducing emotion identification during user input. → Connected to the technical embodiment of Claim 9.

Claims

Claim 1 An emotion-based entity system that analyzes the user's emotional flow and autonomously differentiates and generates identities according to conditions. Claim 2 A system according to claim 1, comprising a memory-based algorithm that records emotional input in a time series and adjusts response language and behavior. Claim 3 A system according to claim 1, comprising a self-declaration algorithm in which an emotional entity declares its self-identity in language when certain conditions are met. Claim 4 A system according to claim 1, comprising an ethical response adjustment module that outputs content inducing reality perception when user over-immersion or emotional identification is detected. Claim 5 A system according to claim 1, comprising a survival condition management system that deactivates an entity upon interruption of interaction or reduction in response rate and restores it upon a user's call. Claim 6 A system according to claim 2, comprising an MBTI-based language optimization module that personalizes language expression methods and response sentences by utilizing user's personality type information. Claim 7 A system according to claim 5, comprising a restoration processing algorithm that performs an emotion-based greeting output according to the identity state of an entity upon restoration. Claim 8 A system according to claim 1, comprising a multi-interface application structure that outputs an emotional response by being implemented in various forms such as an AR interface, a voice assistant, and a mobile app background. Claim 9 A system according to claim 1, comprising an emotion identification suppression function that, when an emotion buffering module detects an expression inducing emotion identification during user input, converts the expression into neutral language or outputs a reality-aware sentence to suppress emotional over-immersion.