AI Conversation Model Switching for Knowledge and Empathic Responses
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI-based conversation systems face challenges in providing complete and appropriate responses to user queries, especially when emotion-based conversations are required, and require significant time and resources to develop, train, and update conversation models for different domains.
Innovation Solution
A method and system that dynamically selects between a knowledge-based and empathic conversation model based on the presence or absence of related knowledge, using a first model trained with user utterance and knowledge, and a second model trained solely on user utterance, to enhance response appropriateness and reduce development costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple domain-specific conversation models are developed and operated, then response accuracy for specific domains is improved, but development cost and time increase significantly
Solution Approach 1:
The patent implements a universal conversation model that can handle multiple domains through dynamic knowledge base switching rather than creating separate models for each domain. The system determines the user's intent and switches between different knowledge bases (general knowledge, domain-specific knowledge, emotional support knowledge) to provide accurate responses across various domains, eliminating the need for multiple specialized models
Solution Approach 2:
The patent introduces a knowledge base determination module as an intermediary between the user input and the conversation model. This module acts as a mediator that analyzes the user's utterance, determines the appropriate knowledge base to use, and routes the query accordingly, thereby achieving domain-specific accuracy without requiring separate models for each domain
2Measurement precision
If domain-specific conversation models are developed and operated, then response accuracy for specific domains is improved, but operational cost increases
Solution Approach 1:
The system employs a single universal conversation model that can operate across multiple domains by dynamically selecting appropriate knowledge bases. This approach reduces operational costs by maintaining one model infrastructure rather than running multiple domain-specific models simultaneously, while still achieving high accuracy through context-appropriate knowledge selection
Solution Approach 2:
The patent applies local quality by providing different knowledge base resources to different parts of the conversation flow based on user needs. Instead of uniformly applying all knowledge bases to all queries, the system selectively activates specific knowledge bases (general, domain-specific, or emotional support) based on the local context of each user interaction, optimizing resource utilization and reducing operational costs
3Measurement precision
If knowledge-based conversation model is used, then informational accuracy is improved, but emotional appropriateness may deteriorate
Solution Approach 1:
The patent implements a dynamic conversation model that can adapt its behavior based on the situation. The system dynamically switches between knowledge-based responses (for informational accuracy) and empathic responses (for emotional appropriateness) by determining the user's emotional state and the context of the conversation. This dynamic adaptation allows the model to prioritize emotional support when needed while maintaining informational accuracy when appropriate
Solution Approach 2:
The system changes the operational parameters of the conversation model based on the situation. When emotional support is needed, the model shifts from a fact-retrieval parameter set to an empathic response parameter set. The knowledge base determination module detects emotional keywords and context, then adjusts the model's behavior parameters to generate emotionally appropriate responses rather than strictly factual ones
4Measurement precision
If conversation models are classified by knowledge domains, then domain-specific accuracy is improved, but model complexity increases
Solution Approach 1:
The patent segments the knowledge bases into distinct categories (general knowledge, domain-specific knowledge, emotional support knowledge) while maintaining a single unified conversation model. This segmentation allows the system to load and use only the relevant knowledge base for each interaction, achieving domain-specific accuracy without the complexity of maintaining multiple separate models. The knowledge base determination module manages this segmentation by selecting appropriate segments based on user input
Data Source
AI summary
There is provided a conversation method and system for operating a conversation model according to the presence or absence of related knowledge. A conversation method according to an embodiment may receive an input of a user utterance, may determine a conversation model based on an inputted user utterance content, may create a conversation content by using the determined conversation model, and may convert the created conversation content into a voice, and may output the voice. Accordingly, different conversation models may be operated according to the presence or absence of related knowledge, so that an empathic conversation may be generated instead of generating an inappropriate response in a knowledge-based conversation, and user's satisfaction on a conversation may be enhanced.