Digital Assistant Multi-Party Voice Query Tailoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic digital assistants are limited to processing and responding to single voice queries, failing to effectively manage complex multi-party voice dialogues in dynamic situations, such as those encountered by first responders.
Innovation Solution
An improved system and method for an electronic digital assistant that processes multi-party voice dialog by identifying roles within the conversation, accessing a database of multi-party interaction templates, and generating queries to retrieve relevant information, allowing for responsive and supplemental information delivery to involved parties.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional electronic digital assistants process only single voice queries, then device complexity and processing resources are reduced, but the ability to manage complex multi-party voice dialogues is insufficient
Solution Approach 1:
The patent segments multi-party voice dialogues by identifying and separating individual speakers and their roles within the conversation. The system divides the complex multi-party interaction into distinct speaker segments, each with assigned roles (e.g., first speaker, second speaker, dispatcher), allowing the digital assistant to process each segment independently while maintaining context awareness. This segmentation approach enables handling of complex dialogues without proportionally increasing overall system complexity.
Solution Approach 2:
The patent applies preliminary action by pre-defining interaction templates that contain predetermined response patterns and information retrieval protocols for various speaker role combinations. Before processing the actual dialogue, the system loads and prepares these templates based on the identified speaker roles. This preliminary preparation allows the digital assistant to quickly match observed interactions against pre-configured templates, reducing real-time processing complexity while maintaining high adaptability to different dialogue scenarios.
2Loss of information
If electronic digital assistants provide comprehensive responses in multi-party conversations, then information completeness is improved, but memory and processing resources increase
Solution Approach 1:
The patent applies local quality by providing different levels of information completeness to different parties based on their roles and needs. Rather than delivering identical comprehensive information to all participants, the system tailors responses to each speaker's specific context, authority level, and information requirements. For example, dispatchers receive different information than field officers, and supervisors receive different summaries than operational personnel. This role-based differentiation maintains information completeness for each user while reducing overall memory requirements by avoiding redundant information storage.
Solution Approach 2:
The patent uses copying by creating and storing reusable interaction templates that capture common dialogue patterns and response structures. Instead of generating entirely new comprehensive responses for each multi-party interaction, the system copies and adapts pre-defined templates based on the specific speaker roles and interaction type. This templating approach allows the system to provide information-complete responses by referencing stored templates rather than generating all information from scratch, significantly reducing memory and processing resource requirements.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An input audio signal is monitored at an electronic digital assistant of an electronic computing device. The assistant determines that the audio signal includes audio from multiple speaking parties, identifies first and second parties of the multiple parties and roles of the parties, accesses a database of multi-party interaction templates and selects, as a function of the roles of the parties, a particular multi-party interaction template that matches the roles, and then generates, based on a detected content of the audio signal and the particular multi-party interaction template, a database query to retrieve information responsive to or supplemental to content of the audio signal. The assistant then causes the query to be performed results of the query provided to the first party, the second party, a dispatcher or administrator associated with the parties, or a computer-aided-dispatch system associated with the parties.