Vocabulary Generation System for Conversational AI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conversational systems lack variability in prompts and responses, leading to user disengagement due to repetitive interactions, as they rely on pre-defined static databases that are not scalable or efficient.
Innovation Solution
A vocabulary generation system that learns and adapts by analyzing user utterances to assign utility values and update responses, dynamically generating prompts and responses based on user interactions, effectively expanding its database with user-specific language and maintaining user interest.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a static database of pre-defined words and phrases is used, then the system structure is simple and easy to implement, but the system becomes repetitive and loses user interest over time
Solution Approach 1:
The patent transforms the static database into a dynamic system that automatically learns and adapts from user interactions. The system evolves its vocabulary database by capturing new words, phrases, and expressions from user utterances, making the interaction pattern dynamic rather than static. This resolves the contradiction by maintaining implementation simplicity while achieving adaptability through automated learning mechanisms.
Solution Approach 2:
The system performs self-updating by automatically learning from user interactions without requiring manual database curation. The automated vocabulary capture and utility value assignment mechanisms allow the system to serve itself, continuously improving its adaptability while maintaining the simplicity of a database-driven architecture.
2Ease of repair
If the same static words and phrases are reused, then the system is easy to maintain, but user interest decreases and abandonment increases
Solution Approach 1:
The system implements continuous learning and adaptation through automated vocabulary capture from every user interaction. This continuous update process ensures the database remains fresh and engaging without requiring manual intervention, thereby extending user retention time while maintaining ease of maintenance through automation.
Solution Approach 2:
The system uses feedback from user interactions to automatically update its vocabulary database. By analyzing user utterances and assigning utility values to new vocabulary items, the system creates a feedback loop that continuously improves user engagement while maintaining simple database-based maintenance.
3Adaptability or versatility
If a large predefined database is used to provide variety, then user interest may be maintained, but the system complexity and computational requirements increase
Solution Approach 1:
The system implements a practical approach by capturing and storing only the most useful vocabulary items from user interactions, rather than attempting to store every possible variation. The utility value mechanism filters vocabulary to retain only those items with sufficient usage frequency and relevance, achieving response variety without excessive database growth or system complexity.
Solution Approach 2:
The system dynamically adjusts database parameters such as utility values, minimum occurrence thresholds, and vocabulary selection criteria based on usage patterns. This allows the system to maintain an optimized balance between response variety and system complexity by adapting database parameters rather than simply increasing database size.
4Adaptability or versatility
If the system adapts dynamically to user language, then user engagement increases, but the computational processing required increases
Solution Approach 1:
The system performs preliminary processing by capturing and generalizing vocabulary patterns during idle periods or in batches, rather than processing every utterance in real-time. The utility value assignment and vocabulary generalization occur systematically, reducing peak computational energy requirements while maintaining language adaptation capabilities.
Data Source
AI summary
Increasingly, conversational systems are used in coaching or supportive contexts, either in an embodied form (e.g., as an avatar in an app or website) or just in a speech-driven for (e.g. Siri). There is a need to keep such systems interesting and appealing over time in order to prevent the user from reducing use of the system or abandoning the system all together. The present system is configured to learn new expressions from user utterances and use them based on their predicted utility during interactions with the user. The present system includes components configured for learning new vocabulary and selecting vocabulary for generating new utterances from the system. This way, the system continually expands its vocabulary database with expressions familiar to and/or used by the user and will be able to engage the user with new utterances so that the user does not lose interest in the system.