Speech Completion Detection Using Dual AI Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic apparatuses face errors in speech recognition due to incomplete user utterances, leading to inconvenient repeated inputs, as they rely on button press or timeout methods to determine speech completion.
Innovation Solution
An electronic apparatus using two models, one trained on sentence structures and another on intents, to determine if a user's speech is completed based on the meaning and context, eliminating the need for external constituents like buttons or timers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If timeout method or button press method is used to determine speech completion, then the electronic apparatus can operate without complex models, but speech recognition errors occur when user utterance is incomplete
Solution Approach 1:
The determination of speech completion is segmented into two independent models: a sentence structure model that analyzes grammatical completeness and an intent model that analyzes semantic completeness. This segmentation allows each model to focus on specific aspects of speech completion, improving overall accuracy while keeping individual model complexity manageable
Solution Approach 2:
The patent introduces two intermediary models (sentence structure model and intent model) that act as mediators between the raw speech input and the final speech recognition processing. These models analyze different dimensions of speech completion and provide intermediate determination results that guide whether to proceed with speech recognition, thereby improving reliability without requiring the main system to handle all complexity
2Ease of operation
If button press method is used, then users can control speech input timing, but users must perform additional physical actions increasing operation complexity
Solution Approach 1:
The system performs self-service by automatically determining speech completion through the sentence structure model and intent model without requiring external user actions like button presses. The models autonomously analyze the speech input and determine whether it is complete, allowing users to simply speak naturally without additional physical operations, thereby improving ease of operation
Data Source
AI summary
In the present disclosure, provided are an electronic apparatus and a controlling method therefor. The electronic apparatus of the present disclosure comprises: a memory in which a first model and a second model are stored; and a processor configured to, when user voice is received, generate a response to the user voice according to completion of the received user voice, the completion being determined on the basis of the first model and the second model. Here, the first model uses, as training data, sentence structures of a plurality of completed sentences, and the second model uses, as training data, a plurality of completed sentences and intents corresponding to the plurality of completed sentences. In particular, the first model and the second model may be artificial intelligence models trained according to at least one of machine learning, a neural network, or a deep learning algorithm.