Deep Growing Neural Gas Network for Dialogue Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dialogue parsing computer systems fail to provide natural interaction experiences for human end users, as they require cumbersome or unnatural memorized commands and struggle to accurately parse natural end-user speech.
Innovation Solution
A method and system that utilize a trained deep growing neural gas neural network to process pre-processed dialogue transcript data, which involves converting words into word embeddings and associating them with concepts, allowing for seamless parsing of human dialogue, particularly suited for applications like quick service restaurant ordering and voice-based customer service.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If current dialogue parsing systems use traditional parsing methods, then they can process dialogue data, but they require cumbersome memorized commands and cannot accurately parse natural speech
Solution Approach 1:
The patent replaces traditional mechanical dialogue parsing methods with a deep growing neural gas neural network, a biological-inspired computational system. This substitution enables the system to naturally learn and parse human speech patterns without requiring rigid predefined commands, thereby improving both ease of operation and parsing accuracy simultaneously
Solution Approach 2:
The patent transforms the dialogue parsing approach by changing the fundamental parameters of the system - transitioning from rule-based parsing to neural network-based parsing with dynamic node growth. This parameter change allows the system to adapt to natural speech variations while maintaining high accuracy in extracting meaningful information
2Ease of operation
If dialogue parsing systems use traditional methods, then they can process basic dialogue data, but they provide cumbersome user experience requiring memorized commands
Solution Approach 1:
The deep growing neural gas neural network is self-organizing and self-adapting. It automatically grows its structure and learns speech patterns without requiring manual programming of dialogue rules, thereby simplifying the user interface while managing system complexity through autonomous learning mechanisms
3Productivity
If current systems integrate dialogue parsing into automated platforms, then they can automate customer service, but they cannot seamlessly extract speech data
Solution Approach 1:
The patent replaces traditional speech extraction mechanisms with a deep growing neural gas neural network that seamlessly integrates into automated platforms. This biological-inspired system continuously learns and adapts to extract speech data with high accuracy while maintaining automated service productivity
Data Source
AI summary
A computer implemented method and associated computer system for dialogue parsing. The method includes receiving dialogue transcript data, pre-processing dialogue transcript data to generate pre-processed dialogue transcript data, providing pre-processed dialogue transcript data as an input to a trained deep growing neural gas neural network; and receiving parsed dialogue transcript data as an output from the trained deep growing neural gas neural network.