Dynamic Language Model for Speech Recognition Using Display Text
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic speech recognition (ASR) systems face challenges in accurately recognizing speech from recorded voice data in call centers and customer service applications due to vocabulary differences, particularly with words like names, addresses, and product names that are not in the domain-specific grammar model.
Innovation Solution
The system improves speech recognition accuracy by capturing relevant text from a device display associated with the spoken dialog, adding identified words to a dynamic language model, and using this model to recognize the recorded utterance, with features including time stamping, utterance categorization, and personalized models based on user interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a domain-specific grammar model is used for speech recognition, then recognition speed and efficiency are improved, but recognition accuracy deteriorates when encountering vocabulary outside the domain (such as names, addresses, product names)
Solution Approach 1:
The system performs preliminary actions by capturing text from the device display before speech recognition occurs. This captured text is then added to the language model in advance, so when speech recognition happens, the model already contains the relevant vocabulary (names, addresses, product names) needed for accurate recognition, eliminating the need to wait for domain-specific model limitations to be overcome
Solution Approach 2:
The system implements a dynamic language model that can be updated in real-time during customer service interactions. Words captured from the device display are dynamically added to the language model based on temporal relevance, allowing the model to adapt its vocabulary dynamically rather than relying on a static domain-specific grammar model
2Device complexity
If a static language model is used, then system complexity is reduced, but adaptability to different contexts and vocabulary deteriorates
Solution Approach 1:
The system implements self-service by automatically capturing text from the device display and adding relevant words to the language model without requiring manual intervention. The system autonomously identifies, captures, and processes text information, dynamically updating its own language model to adapt to different contexts and vocabulary while maintaining relatively simple system architecture
Solution Approach 2:
The system creates a universal solution that works across different customer service contexts by capturing text from the device display, which contains context-specific information. This approach provides multi-functionality, allowing the same system to adapt to various domains (different products, services, companies) without requiring separate domain-specific models, as the text capture mechanism universally extracts relevant vocabulary from any context
Data Source
AI summary
Disclosed herein are systems, methods, and computer-readable storage media for improving speech recognition accuracy using textual context. The method includes retrieving a recorded utterance, capturing text from a device display associated with the spoken dialog and viewed by one party to the recorded utterance, and identifying words in the captured text that are relevant to the recorded utterance. The method further includes adding the identified words to a dynamic language model, and recognizing the recorded utterance using the dynamic language model. The recorded utterance can be a spoken dialog. A time stamp can be assigned to each identified word. The method can include adding identified words to and/or removing identified words from the dynamic language model based on their respective time stamps. A screen scraper can capture text from the device display associated with the recorded utterance. The device display can contain customer service data.