Latency Masking in Automatic Dialog Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic dialog systems experience compounded latency due to processing delays in ASR, NLU, and NLG processors, leading to user frustration, which is typically masked with unnatural 'earcons' like music.
Innovation Solution
Implementing a filler generator that outputs paralinguistic events and fixed phrases to naturally mask latency while processing user communications, using a speech synthesis system to provide transitional messages until a response is ready.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional earcons (music) are used to mask latency, then the latency is effectively masked, but the user experience deteriorates due to annoyance and unnaturalness
Solution Approach 1:
The system uses transient, short-lived filler words and phrases (e.g., 'um', 'uh', 'let me think about that') instead of prolonged music earcons. These brief vocal fillers naturally mask the processing latency without causing user annoyance, as they mimic normal human speech patterns during thinking pauses.
Solution Approach 2:
The system changes the parameter of latency masking from musical/audio content to linguistic/speech content. By transforming the masking mechanism from music earcons to speech-based fillers, the system maintains latency masking effectiveness while eliminating the harmful effect of user frustration.
2Loss of time
If processing speed is increased to reduce latency, then user experience improves, but system complexity increases due to additional processing requirements
Solution Approach 1:
The system introduces an intermediary filler generation component that sits between the speech recognition and natural language processing modules. This filler generator provides transitional speech content during processing delays, effectively managing the time loss without requiring fundamental changes to the core processing architecture.
Solution Approach 2:
The system performs preliminary action by generating filler speech content in advance of the actual response being ready. These filler phrases are prepared and delivered during the processing interval, allowing the system to mask latency without accelerating the core processing operations.
3Object-affected harmful factors
If filler words are used to mask latency, then user experience improves by being more natural, but the system complexity increases due to filler generation requirements
Solution Approach 1:
The system uses a simple, lightweight filler generation mechanism that employs pre-defined short phrases and vocal fillers rather than complex real-time speech synthesis. This approach achieves naturalness by using brief, common speech elements that occur naturally in human conversation during pauses.
Solution Approach 2:
The system applies local quality by using different types of filler content for different contexts and durations. Short vocal fillers like 'um' and 'uh' are used for brief delays, while longer phrases like 'let me think about that' are used for more extended processing times, creating a naturally varying speech pattern.
Data Source
AI summary
A technique for masking latency in an automatic dialog system is provided. A communication is received from a user at the automatic dialog system. The communication is processed in the automatic dialog system to provide a response. At least one transitional message is provided to the user from the automatic dialog system while processing the communication. A response is provided to the user from the automatic dialog system in accordance with the received communication from the user.