Learning Feedback Analysis Using Attention on Time-Series Movements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing learning assistance technologies fail to provide adequate feedback to users by simply presenting their learning level, making it difficult for them to identify specific areas of improvement and the extent of their inadequacies in mastering movements or actions.
Innovation Solution
An information processing method using supervised-trained machine learning models to determine a user's learning level and identify inadequately learned portions, incorporating an Attention mechanism to highlight differences in time-series media information, and presenting this information visually to the user.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If only learning level is presented to users, then the system is simple to operate, but users cannot identify specific areas of improvement
Solution Approach 1:
The feedback is segmented into multiple components: overall learning level, specific inadequate portions, and detailed improvement suggestions. This segmentation allows users to see both the big picture and specific areas needing attention, resolving the contradiction between simplicity and information completeness.
Solution Approach 2:
The system adds temporal and spatial dimensions to the feedback by showing time-series media information with highlighted inadequate portions. This multi-dimensional presentation provides comprehensive information while maintaining user-friendly visualization through graphical interfaces.
2Loss of information
If detailed feedback on specific portions is provided, then users can identify areas of improvement, but the system complexity increases
Solution Approach 1:
Machine learning models serve as intermediaries that automatically analyze time-series media information and generate detailed feedback. This intermediary processing handles the complexity of detailed analysis while presenting results in a user-friendly format, resolving the contradiction between detailed feedback and system complexity.
Solution Approach 2:
The system replaces manual analysis with automated machine learning-based analysis. This substitution enables detailed feedback generation without requiring complex manual intervention, maintaining system accessibility while providing comprehensive analysis.
3Measurement precision
If machine learning models are used to analyze time-series media information, then accurate learning level determination is achieved, but processing time increases
Solution Approach 1:
The system performs preliminary processing of time-series media information before detailed analysis, preparing data in advance to accelerate the machine learning model's processing. This preliminary action reduces processing time while maintaining accurate learning level determination.
Solution Approach 2:
The system applies partial analysis to commonly occurring patterns using pre-trained models, performing detailed analysis only when necessary. This selective approach reduces overall processing time while maintaining high accuracy for critical assessments.
Data Source
AI summary
Provided is an information processing method that includes an input step of inputting time-series media information representing a movement or conduct of a learning user, a first determination step of determining a learning level of the user on the basis of the time-series media information, and an output step of outputting a portion in the time-series media information at which the movement or action by the user is different from a reference movement or action, on the basis of the learning level of the user determined in the first determination step.