Chatbot User Care System Using Biometric and Semantic Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional intelligent conversation robots are limited in determining user states as they rely on exact pattern matching and do not consider dialogue context or biometric data, making them inefficient and inconvenient for user health monitoring.
Innovation Solution
A user care system using a chatbot with a server-based system that generates and analyzes conversation content, compares user responses to determine states, and integrates biometric data from wearable devices to identify dangerous situations, triggering external help when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional chat robots use exact pattern matching to answer questions, then they can provide consistent fixed answers, but they require a large amount of dialogue examples and construction costs increase
Solution Approach 1:
The patent changes the matching parameter from exact pattern matching to similarity-based matching using cosine similarity calculation. This allows the system to find semantically similar questions without requiring exact matches, reducing the need for extensive dialogue examples while maintaining answer reliability.
Solution Approach 2:
The patent replaces the mechanical exact pattern matching system with an AI-based semantic understanding system using BERT embeddings and cosine similarity. This substitution enables the system to understand question intent rather than relying on pre-defined patterns, reducing database construction complexity.
2Device complexity
If chat robots answer each question with fixed responses without considering dialogue context, then they simplify the conversation processing, but they cannot understand past information or maintain conversation flow
Solution Approach 1:
The patent performs preliminary encoding of dialogue history using BERT embeddings before processing the current question. This allows the system to incorporate past conversation context into the similarity matching process, enabling understanding of dialogue flow without significantly increasing processing complexity.
Solution Approach 2:
The patent introduces an intermediary semantic embedding layer that bridges the gap between simple pattern matching and complex context understanding. The BERT embeddings act as intermediaries that capture semantic meaning and contextual information, allowing the system to maintain conversation flow while keeping processing manageable.
3Quantity of substance
If users directly input health information into mobile terminals for health care, then health information can be collected, but the method becomes inconvenient for users
Solution Approach 1:
The patent enables the wearable device to automatically collect and transmit health information without requiring user input. The system serves itself by autonomously gathering biometric data and sending it to the server for analysis, eliminating the need for users to manually enter health information.
Solution Approach 2:
The patent introduces a wearable device as an intermediary between the user and the health care system. This intermediary automatically captures health data through sensors and transmits it to the server, removing the burden of manual input from the user while ensuring comprehensive health information collection.
4Measurement precision
If conventional health care methods collect health information, then they can analyze user health status, but they cannot quickly determine dangerous situations or provide timely external help
Solution Approach 1:
The patent implements a feedback mechanism where the server continuously receives health data from wearable devices, compares it against normal ranges, and immediately triggers alerts when dangerous situations are detected. This real-time feedback loop enables quick determination of dangerous states and timely external assistance.
Solution Approach 2:
The patent pre-sets normal ranges for various biometric parameters and prepares alert protocols in advance. When health data is received, the system immediately compares it against these pre-defined thresholds, enabling rapid detection of dangerous situations without requiring complex real-time analysis.
Data Source
AI summary
A user care system accuses a chatbot. According to an embodiment of the present disclosure, not only a current state of a user can be quickly determined on the basis of a content of a conversation between a chatbot and the user, but also the user can quickly receive help from the outside when it is determined that the user is in a dangerous situation. Further, not only the chatbot can quickly determine a current state of the user on the basis of a change in biometric information of the user, but also the user can quickly receive help from the outside when it is determined that the user is in a dangerous situation.