Dialogue Processing Device Prioritizing Sentence Units by Dialog Act
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Intelligent automatic response systems face challenges in accurately detecting the main content of a user's query when multiple sentences are involved, leading to reduced user satisfaction due to sequential responses rather than prioritizing sentences based on intended intention.
Innovation Solution
A method and device for processing user speech by identifying and prioritizing sentence units based on dialog act information, using a combination of rule-based and machine learning-based models to determine the sequence of processing, ensuring that sentences reflecting proactive intentions, such as requests or questions, are processed first.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an intelligent agent sequentially responds to each sentence in a user's speech, then it can process all sentences, but the user's satisfaction is considerably reduced due to inability to detect main content priority
Solution Approach 1:
The system performs preliminary analysis of dialog act information for each sentence unit before processing responses. By extracting and analyzing dialog acts (such as requests, questions, statements) in advance, the system determines priority orders of sentence units ahead of time, enabling it to identify the main content intent early and respond more efficiently without sequential processing delays
Solution Approach 2:
The system divides the user's speech into multiple sentence units and analyzes each unit's dialog act information separately. This segmentation allows the system to evaluate the importance and priority of each sentence independently, then process them in the determined priority order rather than strictly sequentially, improving both detection accuracy and response efficiency
2Ease of operation
If the system processes sentences in the order they appear in speech, then processing is simple, but sentences reflecting user intention may be processed later reducing satisfaction
Solution Approach 1:
The system performs preliminary extraction of dialog act information from each sentence unit before processing. By analyzing the dialog acts (request, question, statement, etc.) in advance, the system determines priority orders ahead of time, allowing it to reorder sentence units based on their importance to user intent rather than processing them in the original speech order, thus reducing time loss while maintaining operational efficiency
Solution Approach 2:
The system dynamically adjusts the processing order of sentence units based on their dialog act information and determined priority orders. Instead of a fixed sequential processing approach, the system flexibly reorders sentence units according to their relative importance to the user's main intent, optimizing response time while keeping the processing mechanism adaptable rather than rigid
Data Source
AI summary
A method, performed by a dialogue processing device, of processing dialogue associated with a user based on dialog act information, the method comprises receiving speech information, corresponding to speech of the user, including a plurality of sentence units; identifying a first sentence unit and a second sentence unit, of the plurality of sentence units, based on receiving the speech information; extracting a first dialog act indicative of an intention of the first sentence unit and extracting a second dialog act indicative of an intention of the second sentence unit; extracting a first dialog act indicative of an intention of the first sentence unit and extracting a second dialog act indicative of an intention of the second sentence unit; processing the first sentence unit and the second unit in a sequence according to respective priority orders assigned based on a number of empty slots of dialogue frames of the sentence units.