On-Device AI Interpretation for Selective Multi-Speaker Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current on-device AI systems struggle to accurately interpret conversations between multiple speakers in real time and identify individual speakers for selective interpretation services.
Innovation Solution
An on-device AI device that preprocesses utterance voices, classifies them by speaker, and selectively interprets the voice of a specific speaker into a target language, with the option to distribute AI interpretation processing to external devices when needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If on-device AI processes all utterance voices from multiple speakers, then interpretation completeness is improved, but processing accuracy and speaker identification capability deteriorate
Solution Approach 1:
The patent segments the audio input stream into separate utterance voices by detecting silence intervals between speakers. The processor identifies when one speaker stops and another starts by analyzing audio signal continuity, thereby separating mixed audio into distinct speaker segments for individual processing
Solution Approach 2:
The patent introduces an intermediary classification process that acts as a mediator between raw audio input and interpretation. The processor classifies each segmented utterance by identifying speaker characteristics and grouping them, creating an intermediate structured representation that enables accurate speaker identification before final interpretation
2Speed
If on-device AI performs full AI interpretation processing locally, then processing speed is improved, but power consumption increases
Solution Approach 1:
The patent applies partial action by performing only the necessary preprocessing steps (segmentation and classification) locally on the device, while potentially delegating full AI interpretation to external servers. This selective local processing maintains speed advantages for critical path operations while reducing overall power consumption
Solution Approach 2:
The patent introduces an intermediary classification layer that prepares data in a structured format suitable for both local and remote processing. This intermediary step enables flexible architecture where certain operations can be performed locally for speed while others can be offloaded to reduce power consumption
3Adaptability or versatility
If on-device AI processes multiple speakers simultaneously, then service capability is improved, but device complexity increases
Solution Approach 1:
The patent simplifies multi-speaker processing by segmenting the audio stream into discrete utterance units separated by silence intervals. This segmentation transforms a complex simultaneous processing problem into a sequence of simpler individual utterance processing tasks
Solution Approach 2:
The patent implements self-service through automatic speaker classification and utterance segmentation without requiring manual intervention. The system automatically identifies speaker boundaries and classifies utterances based on acoustic characteristics, enabling multi-speaker support through automated rather than manually configured processes
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An on-device artificial intelligence (AI) device capable of interpreting a conversation between multiple speakers in real time and a method thereof are provided. The AI device can include an input module where utterance voice of a speaker is input; and a processor configured to perform AI processing to interpret the utterance voice of the speaker into a target language, wherein the processor is configured to when utterance voices are input from a plurality of speakers, preprocess the utterance voices, classify the preprocessed utterance voices by speaker, when a specific speaker is selected from the plurality of speakers, extract utterance voice of the selected specific speaker from the classified utterance voices by speaker, and interpret the utterance voice of the specific speaker into a target language and output it.