Automated Voice Interface Prompt Generation from Corpus Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current voice user interface (VUI) systems for call routing and similar applications require significant time and expense to manually craft dialog prompts and grammars, leading to high development costs and long time-to-market, as well as inefficiencies in reusing prompts across different applications.
Innovation Solution
Automatically generating speech dialog prompts from an annotated transcription corpus, using data archetypes and keyword extraction to create confirmation, back-off, and disambiguation prompts, reducing the need for manual grammar writing and enabling more accurate and reusable prompts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual crafting of dialog prompts and grammars is used, then accuracy and precision of voice user interface is improved, but development time and cost increase significantly
Solution Approach 1:
The system automatically generates dialog prompts and grammars by processing existing training corpus data, allowing the system to serve itself rather than requiring manual crafting. The automated system extracts patterns from training data and generates appropriate prompts, reducing development time while maintaining quality through data-driven generation rather than manual creation
Solution Approach 2:
The patent replaces the mechanical process of manual prompt crafting with an automated computational system. Instead of humans manually creating prompts, the system uses algorithms to process training data and generate prompts automatically, substituting manual mechanical work with automated computational processes
2Measurement precision
If manual crafting of dialog prompts is used, then precision of user interaction is improved, but development cost increases
Solution Approach 1:
The system automatically generates dialog prompts by processing training corpus data, eliminating the need for manual crafting. The automated system extracts patterns from existing data and generates appropriate prompts, reducing development costs while maintaining interaction precision through data-driven generation
Solution Approach 2:
The patent changes the parameters of prompt generation from manual creation to automated data-driven generation. By transforming the generation process into a computational task that processes training data, the system reduces costs while maintaining quality through systematic extraction of interaction patterns from existing corpus
3Measurement precision
If hand-crafted prompts are used, then accuracy for specific applications is improved, but reusability across applications decreases
Solution Approach 1:
The system generates prompts that are applicable across multiple applications by extracting universal patterns from training corpus data. The automated generation process creates prompts that can be reused across different applications, eliminating the need for application-specific manual crafting while maintaining accuracy through data-driven patterns
Solution Approach 2:
The patent uses copying by extracting patterns from training data and generating prompts that replicate successful interaction patterns. This allows prompts to be copied and reused across applications, maintaining effectiveness without requiring manual recreation for each specific application
Data Source
AI summary
A method for developing a voice user interface for a statistical semantic system is described. A set of semantic meanings is defined that reflect semantic classification of a user input dialog. Then, a set of speech dialog prompts is automatically developed from an annotated transcription corpus for directing user inputs to corresponding final semantic meanings. The statistical semantic system may be a call routing application where the semantic meanings are call routing destinations.