Adaptive Dialogue Management Interfaces for Multimodal Intent Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chatbot technologies face challenges in accurately understanding user intents and resolving application domains due to limitations in voice or speech recognition, leading to misinterpretation and inefficiencies in providing services.
Innovation Solution
A dialogue management framework that utilizes a pre-trained machine learning model to analyze user inputs, including utterances, clicks, touches, and gaze tracking, to resolve application domains, intents, and slots, adaptively determining an expression scheme for a hierarchical interface to enhance user interaction and service provision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If voice or speech recognition is used for user input, then interaction convenience is improved, but understanding accuracy deteriorates due to misinterpretation
Solution Approach 1:
The patent combines multiple input modalities (voice, text, gaze, touch) into a unified dialogue management system. The gaze estimation module processes eye movement data alongside voice and text inputs, merging these different data types to achieve more accurate intent recognition while maintaining user interaction convenience.
Solution Approach 2:
The patent introduces a gaze estimation module as an intermediary between the user and the dialogue management system. This module translates gaze movements into input signals that complement voice and text, serving as a mediator that enhances understanding accuracy without requiring direct changes to the core dialogue management architecture.
2Measurement precision
If multiple input modalities are integrated, then understanding accuracy is improved, but system complexity increases
Solution Approach 1:
The patent segments the input processing system into distinct modules: voice processing, text processing, and gaze estimation modules. Each module handles specific input types independently before feeding into the unified dialogue management framework, which reduces overall system complexity by allowing modular development and independent optimization of each component.
Solution Approach 2:
The dialogue management framework is designed with universality to handle multiple input modalities through a common processing architecture. The framework can accommodate voice, text, and gaze inputs using the same intent recognition and slot filling mechanisms, reducing the need for separate specialized systems for each modality and thereby lowering system complexity.
3Measurement precision
If gaze tracking is added to input methods, then intent resolution accuracy is improved, but measurement and detection difficulty increases
Solution Approach 1:
The patent replaces complex mechanical gaze tracking hardware with software-based gaze estimation algorithms that process standard camera images. Instead of using specialized eye-tracking devices with complex mechanical components, the system uses computational methods to estimate gaze direction from conventional camera data, significantly reducing measurement and detection difficulty while maintaining or improving intent resolution accuracy.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is a dialogue management method and apparatus. The dialogue management method includes sequentially resolving an application domain to provide a service to a user, a function appropriate for an intent of a user from among functions of the application domain, and at least one slot to perform the function, adaptively determining an expression scheme of a dialogue management interface depending on a progress of the sequentially resolving, and displaying the dialogue management interface.