Dynamic Speech Recognizer Selection for Chat Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech recognition systems in chatbots and digital personal assistants face challenges in accurately recognizing voice commands, especially those containing names, nicknames, titles, addresses, and geographical locations, due to limited computational resources and reliance on statistic-based techniques, leading to unreliable user interactions.
Innovation Solution
A method is introduced where a processor dynamically selects and combines multiple speech recognizers based on triggers and context within user inputs, utilizing pattern-based, free-dictation, address book, and dynamically created recognizers to enhance recognition accuracy by separating and processing audio inputs into parts, thereby improving the reliability of voice command interpretation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single speech recognition system is used, then device complexity is reduced, but recognition accuracy for diverse inputs deteriorates
Solution Approach 1:
The audio input is divided into multiple parts or segments, and different speech recognition systems are selectively applied to different segments based on their characteristics. This allows each segment to be processed by the most appropriate recognizer, improving overall accuracy without requiring all recognizers to process every input simultaneously.
Solution Approach 2:
The system dynamically selects and switches between different speech recognition systems based on the characteristics of the input audio or recognized patterns. This dynamic adaptation allows the system to optimize recognition accuracy for each specific input without maintaining a static, overly complex architecture.
2Measurement precision
If multiple speech recognizers are used, then recognition accuracy improves, but computational resource consumption increases
Solution Approach 1:
By segmenting the audio input and applying different recognizers to different segments, the system avoids using multiple heavy-weight recognizers for the entire audio stream. Only necessary portions are processed by specialized recognizers, reducing overall computational load.
Solution Approach 2:
The system applies speech recognition processing only to the extent necessary - using specialized recognizers only for specific segments or patterns that require them, rather than applying all available recognition resources to every input uniformly.
3Device complexity
If statistic-based techniques are used, then system simplicity is maintained, but recognition reliability for names and proper nouns deteriorates
Solution Approach 1:
The system introduces intermediary components that bridge between simple statistic-based recognition and more reliable specialized recognition. Triggers or pattern-matching mechanisms act as intermediaries to identify when proper nouns or specific patterns are present, then activate appropriate specialized recognizers for those segments.
Solution Approach 2:
Different recognition approaches are applied to different parts of the speech input based on local characteristics. Statistic-based techniques handle general speech, while specialized recognizers with higher reliability are applied locally to segments containing names, proper nouns, or specific patterns where they are most needed.
Data Source
AI summary
Disclosed are computer-implemented methods and systems for dynamic selection of speech recognition systems for the use in Chat Information Systems (CIS) based on multiple criteria and context of human-machine interaction. Specifically, once a first user audio input is received, it is analyzed so as to locate specific triggers, determine the context of the interaction or predict the subsequent user audio inputs. Based on at least one of these criteria, one of a free-diction recognizer, pattern-based recognizer, address book based recognizer or dynamically created recognizer is selected for recognizing the subsequent user audio input. The methods described herein increase the accuracy of automatic recognition of user voice commands, thereby enhancing overall user experience of using CIS, chat agents and similar digital personal assistant systems.