Wearable Spoken Utterance Completion With Selective Content Rendering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for obtaining specific information, such as the color of an HVAC fan wire, through search engines or automated assistants are latency-prone and computationally wasteful, requiring significant user interaction and resource utilization.
Innovation Solution
Implementing a system that automatically generates and renders continuing content, like the word 'green', to complete a partial spoken utterance, using earbuds, based on audio processing and machine learning models, without explicit user invocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a search engine or automated assistant is used to obtain specific information, then the information can be resolved, but significant latency and computational resource usage occur
Solution Approach 1:
The system pre-loads and caches frequently accessed information (such as HVAC wire colors) into a local knowledge base before it is needed. When a user asks a question, the system first checks the cached knowledge base for an immediate answer, avoiding the latency of real-time search engine queries or automated assistant processing.
Solution Approach 2:
The patent introduces an intermediate cached knowledge base between the user query and the external information sources. This intermediary layer provides rapid access to commonly requested information, reducing the need to repeatedly query external search engines or automated assistants for the same data.
2Reliability
If a search engine or automated assistant is used to obtain specific information, then the information can be resolved, but significant client device resources are utilized
Solution Approach 1:
The system pre-loads and caches frequently accessed information (such as HVAC wire colors) into a local knowledge base before it is needed. When a user asks a question, the system first checks the cached knowledge base for an immediate answer, avoiding the latency of real-time search engine queries or automated assistant processing.
Solution Approach 2:
The patent introduces an intermediate cached knowledge base between the user query and the external information sources. This intermediary layer provides rapid access to commonly requested information, reducing the need to repeatedly query external search engines or automated assistants for the same data.
3Reliability
If an automated assistant is used to obtain information, then the query can be processed, but the query must be longer and more syntactically complex
Solution Approach 1:
The patent introduces an intermediate cached knowledge base between the user query and the external information sources. This intermediary layer provides rapid access to commonly requested information, reducing the need to repeatedly query external search engines or automated assistants for the same data.
4Loss of information
If continuing content is rendered in its entirety, then complete information is provided, but unnecessary content is rendered increasing latency and resource usage
Solution Approach 1:
The system implements selective rendering that provides only the essential portion of the answer needed to satisfy the user's information request. For example, when the cached knowledge base contains information about HVAC wire colors, the system renders only the specific color information rather than entire paragraphs or articles, thereby reducing rendering time and resource consumption while still completing the user's information need.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
Implementations set forth herein relate to generating and rendering continuing content (e.g., natural language content) that can be used, by a human user, in completing a partial spoken utterance of the user. The continuing content can be rendered via a wearable device (e.g., earbuds or eyeglasses). In various implementations, generating and/or rendering of the continuing content can be performed automatically (at least when certain condition(s) are satisfied) responsive to the partial spoken utterance and independent of any user input(s) that explicitly invokes the generating and/or the rendering. Furthermore, in many implementations, at least rendering of the continuing content is selectively performed. For example, automatic rendering of the continuing content can be selectively performed when certain condition(s) are satisfied, such as detection of a disfluency following the partial spoken utterance.