Voice Recognizing Apparatus Sequential Start Language

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Voice recognition systems struggle to accurately identify start languages when they are uttered seamlessly with utterance languages, as the vocalization differences in these contexts lead to misrecognition, especially in IoT devices that require precise user enrollment.

Innovation Solution

A voice recognizing apparatus that intelligently recognizes voices by acquiring a sequential start language uttered with an utterance language, determining its recognition score using a start language recognition model, and updating the model to set the sequential start language as an additional start language, allowing for seamless voice recognition even when start and utterance languages are spoken together.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a voice recognition system uses only a basic start language for enrollment, then the system structure remains simple, but it cannot recognize sequential start languages uttered with utterance languages

Engineering Contradiction:

Improverecognition of sequential start languageVSAvoidvoice recognition model structure

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The voice recognition model is designed to dynamically adapt by sequentially updating start language information. The model transitions from a static basic start language to a dynamic structure that incorporates sequential start languages through continuous learning and updates based on recognition scores and authentication attempts.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs preliminary enrollment by acquiring basic start language information before actual use. This preliminary action establishes a foundation that can later be expanded with sequential start languages, allowing the system to prepare for seamless speech recognition in advance.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If the system enrolls users using only basic start language, then the enrollment process is simple, but recognition accuracy decreases when users utter sequential start languages

Engineering Contradiction:

Improvestart language recognition accuracyVSAvoiduser enrollment procedure

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The system implements feedback mechanisms by evaluating recognition scores of sequential start languages and comparing them against authentication attempt thresholds. This feedback loop enables the system to automatically update and refine its start language recognition capabilities without requiring complex manual re-enrollment procedures.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The voice recognition model performs self-updates by automatically incorporating sequential start language information when recognition scores meet authentication criteria. This self-service capability eliminates the need for manual re-enrollment, maintaining ease of operation while improving recognition accuracy.

Inventive Principle:

Principle #25Self-service

3Measurement precision

If the voice recognition model is updated frequently with sequential start languages, then recognition accuracy improves, but system stability decreases

Engineering Contradiction:

Improvesequential start language recognitionVSAvoidrecognition model consistency

Core Design Contradiction:

Measurement precisionVSStability of the object's composition

Solution Approach 1:

The system maintains continuity by performing gradual, incremental updates to the voice recognition model rather than abrupt changes. The continuous learning process updates start language information sequentially based on authentication attempts, ensuring stable and consistent model evolution while improving recognition accuracy over time.

Inventive Principle:

Principle #20Continuity of useful action

4Reliability

If the system requires authentication attempts before updating, then false updates are prevented, but response time increases

Engineering Contradiction:

Improveupdate accuracyVSAvoidmodel update delay

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system applies preliminary anti-action by establishing authentication attempt thresholds before updates occur. This preventive measure blocks potential false updates in advance, ensuring reliability. The threshold mechanism allows legitimate sequential start languages to pass while preventing unauthorized or erroneous updates.

Inventive Principle:

Principle #9Preliminary anti-action

Data Source

PatentUS11423878B2Intelligent voice recognizing method, apparatus, and intelligent computing device

Publication Date: 2022.08.23 LG ELECTRONICS INC

AI summary

Disclosed are an intelligent voice recognizing method, a voice recognizing apparatus, and an intelligent computing device. The an intelligent voice recognizing method according to an embodiment of the present disclosure receives a voice, acquires a sequential start language uttered sequentially with a utterance language from the voice, and sets the sequential start language as an additional start language other than a basic start language when the sequential start language is recognized as a start language of a voice recognizing apparatus, thereby being able to authenticate a user and recognize a voice even through a seamless scheme voice that is uttered in an actual situation. According to the present disclosure, one or more of the voice recognizing device, intelligent computing device, and server may be related to artificial intelligence (AI) modules, unmanned aerial vehicles (UAVs), robots, augmented reality (AR) devices, virtual reality (VR) devices, and 5G service-related devices.