Speech Interaction Demand Prediction Using Seq2Seq Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional speech interaction systems often fail to accurately understand user demands, leading to poor user experience and low interaction efficiency, as they either return a 'cannot understand' response or prompt users to simplify their instructions, requiring repeated input changes.
Innovation Solution
A speech interaction method that performs demand prediction using a pre-trained Seq2Seq model to map user input to potential demand expressions, returning these expressions to the user for confirmation, rather than simply stating the inability to understand, thereby improving interaction efficiency and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional speech recognition and demand analysis are performed, then the system can understand clear user instructions, but it fails to accurately understand complex or ambiguous user demands, leading to poor user experience
Solution Approach 1:
The system performs demand prediction before final demand analysis, generating predicted demand expressions in advance. This preliminary action allows the system to prepare multiple potential interpretations of ambiguous user input, improving the accuracy of demand understanding while maintaining good user interaction experience by presenting predictions for user confirmation.
2Productivity
If the system returns 'cannot understand' or prompts users to simplify instructions, then it maintains simple response logic, but it requires repeated input changes from users, reducing interaction efficiency
Solution Approach 1:
The system performs demand prediction before final demand analysis, generating predicted demand expressions in advance. This preliminary action allows the system to prepare multiple potential interpretations of ambiguous user input, improving the accuracy of demand understanding while maintaining good user interaction experience by presenting predictions for user confirmation.
Solution Approach 2:
The system presents predicted demand expressions to users for confirmation, creating a feedback loop. Users can confirm or correct the predicted demands, and the system uses this feedback to improve future predictions. This feedback mechanism significantly improves interaction efficiency by reducing repeated input changes, while the feedback data also helps refine the prediction model over time.
3Measurement precision
If the system uses a pre-trained Seq2Seq model for demand prediction, then it can generate accurate predicted demand expressions, but it increases the complexity of the speech interaction system
Solution Approach 1:
The patent introduces a demand prediction model as an intermediary component between speech recognition and demand analysis. This Seq2Seq model acts as a mediator that transforms recognized speech into predicted demand expressions, which then guide the subsequent demand analysis process. The intermediary model improves prediction accuracy while keeping the overall system architecture modular and manageable.
Data Source
AI summary
The present disclosure provides a speech interaction method, apparatus, device and computer storage medium and relates to the field of artificial intelligence. A specific implementation solution is as follows: performing speech recognition and demand analysis for a first speech instruction input by a user; performing demand prediction for the first speech instruction if the demand analysis fails, to obtain at least one demand expression; returning at least one of the demand expression to the user in a form of a question; performing a service response with a demand analysis result corresponding to the demand expression confirmed by the user, if a second speech instruction confirming at least one of the demand expression is received from the user. The present disclosure can efficiently improve the user's interaction efficiency and enhance the user's experience.