Digital Assistant Session Window for Continuous Voice Dialog
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional digital assistant systems require triggering inputs to initiate and continue interactions, leading to cumbersome and repetitive exchanges between users and devices, lacking seamless and effective continuous dialog capabilities.
Innovation Solution
A system and process for continuous dialog with a digital assistant that allows for the initiation and continuation of interactions without requiring explicit triggering inputs, using a session window mechanism with variable speech thresholds to manage user inputs and outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional digital assistant systems require triggering inputs to initiate and continue interactions, then the system can clearly identify when user input is intended for the assistant, but the interaction becomes cumbersome and repetitive
Solution Approach 1:
The system performs preliminary action by proactively detecting potential user inputs before a trigger is required. The digital assistant continuously monitors audio inputs and uses speech recognition to identify when the user is speaking, allowing the system to prepare for and respond to queries without requiring the user to first issue a trigger command. This eliminates the repetitive trigger-respond pattern and creates seamless continuous dialogue.
2Adaptability or versatility
If the digital assistant continuously listens for user input without triggers, then seamless continuous dialogue is achieved, but the system may incorrectly process non-user-directed speech
Solution Approach 1:
The system applies local quality by differentiating the processing of different audio inputs based on their characteristics and context. Rather than uniformly processing all audio signals the same way, the system analyzes speech patterns, directionality, and contextual cues to determine whether a detected speech input is actually directed at the assistant. This allows continuous listening while maintaining high accuracy in identifying genuine user queries versus ambient speech.
3Adaptability or versatility
If the session window duration is fixed, then the system structure is simple, but it cannot adapt to varying user interaction patterns
Solution Approach 1:
The system implements dynamics by making the session window duration variable rather than fixed. The session management mechanism dynamically adjusts the time window based on real-time analysis of user interaction patterns, speech characteristics, and contextual factors. This allows the system to extend sessions when users are engaged in continuous dialogue while automatically closing them when no relevant input is detected, adapting to varying user behaviors without requiring complex manual configuration.
Data Source
AI summary
Systems and processes for operating an intelligent automated assistant are provided. For example, a first speech input directed to a digital assistant is received from a user. A first response is provided based on the first speech input. A session window is initiated, wherein the session window is associated with a variable speech threshold. A second speech input is received during the session window. In accordance with a determination that the second speech input includes speech directed to the digital assistant, a duration associated with the session window is increased. In accordance with a determination that the variable speech threshold does not exceed a predetermined speech threshold, the session window is ended.