Dialogue Device Knowledge Transition Reduces Scenario Description
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Scenario-type dialogue systems require extensive and costly creation of dialogue scenarios and response rules to handle user utterances outside their predefined scenarios, limiting their flexibility and efficiency in long-term dialogue.
Innovation Solution
A dialogue device and method that utilize a knowledge base with relation labels to generate system utterances based on user inputs, incorporating a basic scenario and user scenario to determine transition-destination knowledge, reducing the need for extensive scenario description and enabling flexible dialogue.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a scenario-type dialogue system is used to provide high-quality dialogue, then dialogue quality is improved, but the amount of cost and description required increases significantly
Solution Approach 1:
The patent segments the dialogue scenario into a basic scenario framework and dynamic knowledge base components. The basic scenario contains only essential transition methods between knowledge types, while specific knowledge content is stored separately in the knowledge base, allowing the system to maintain dialogue quality without requiring exhaustive scenario descriptions for every possible utterance.
Solution Approach 2:
The patent introduces dynamic knowledge bases that can be updated and expanded without modifying the basic scenario structure. The knowledge base adapts to new user utterances by adding new knowledge entries and relations, allowing the system to handle evolving dialogue requirements while maintaining a simple, stable scenario framework.
2Adaptability or versatility
If more response rules are described to handle user utterances outside the scenario, then dialogue flexibility is improved, but the complexity and cost of creating the dialogue scenario increases
Solution Approach 1:
The patent creates a universal basic scenario framework that can handle multiple types of user utterances through a small set of transition rules. The knowledge base serves as a multi-functional repository that adapts to different dialogue contexts, allowing the same basic scenario structure to support diverse dialogue scenarios without requiring separate response rules for each case.
Solution Approach 2:
The patent introduces the knowledge base as an intermediary layer between the user utterance and the scenario processing. The knowledge base translates diverse user inputs into standardized knowledge representations that the basic scenario can process, eliminating the need for numerous specific response rules while maintaining flexibility in handling various utterance types.
3Duration of action of stationary object
If a large amount of dialogue scenario description is created to perform long-time dialogue, then dialogue duration is improved, but the cost and complexity of system creation increases
Solution Approach 1:
The patent prepares a reusable basic scenario framework and knowledge base structure in advance, which can support long-term dialogue without requiring extensive pre-description of all possible dialogue paths. The preliminary structure includes general transition methods and knowledge organization patterns that can be applied repeatedly across different dialogue sessions and contexts.
Solution Approach 2:
The patent enables long-term dialogue by allowing the knowledge base to evolve and change its parameters (knowledge entries, relations, and structures) over time. The basic scenario remains stable while the knowledge base adapts to new information and contexts, providing a mechanism for indefinite dialogue continuation without requiring proportional increases in scenario description.
Data Source
AI summary
Provided are a dialogue device, a dialogue method, a data structure, and a program capable of realizing various dialogues while reducing the amount of description of a dialogue scenario.A knowledge transition unit 120 determines a next type of knowledge based on: a knowledge base 130 in which a relation label indicating each of relations between a plurality of types of knowledge is attached to knowledge about each of utterances to express the knowledge about the utterance; a user utterance; current knowledge; and a dialogue scenario including a basic scenario in which a transition method between the plurality of types of knowledge in the knowledge base is determined using the relation label, and an utterance generation unit 150 generates a system utterance based on the next type of knowledge.