Dynamic Machine Learning Model Visualization Interface
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for visualizing quality information of machine learning models are inadequate as they cannot dynamically adjust content based on the behavior of the model and user evaluation results, making it difficult for users to understand the correct situation.
Innovation Solution
An information processing apparatus that acquires quantitative and user input information, selects relevant items and visualization methods from predefined display information definitions, and generates display information that includes partial content of both types of information, allowing for dynamic adjustment based on user attributes and developmental phases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all kinds of quality information are displayed, then completeness of information is improved, but ease of understanding deteriorates
Solution Approach 1:
The patent segments quality information into multiple categories (accuracy, robustness, efficiency, interpretability, fairness) and further divides it into fixed items and flexible items. This segmentation allows the system to organize comprehensive information while presenting it in a structured, manageable way that improves user understanding.
Solution Approach 2:
The patent implements dynamic adjustment of displayed information based on user attributes (expertise level, role) and developmental phase (training, testing, operation). The system dynamically selects which flexible items to display and how to visualize them, allowing the information presentation to adapt to different contexts while maintaining completeness.
2Ease of operation
If predetermined fixed items are displayed, then ease of operation is improved, but adaptability deteriorates
Solution Approach 1:
The patent creates a universal display framework that handles both fixed items (always displayed) and flexible items (conditionally displayed) through a unified interface. The same visualization system adapts its content based on user attributes and developmental phase, making the system multi-functional without requiring separate interfaces.
Solution Approach 2:
The system dynamically adjusts the set of displayed flexible items based on user attributes and developmental phase. During training, different items are shown compared to testing or operation phases. This dynamic adaptation allows the system to maintain ease of operation while being highly adaptable to different model behaviors and user needs.
3Measurement precision
If comprehensive quality information is visualized, then measurement precision is improved, but device complexity deteriorates
Solution Approach 1:
The patent segments the visualization system into modular components: fixed item display, flexible item selection, user attribute detection, and developmental phase identification. Each module handles a specific aspect of quality information presentation, reducing overall system complexity while enabling comprehensive measurement precision through their coordinated operation.
Data Source
AI summary
According to one embodiment, an information processing apparatus includes a processor. The processor acquires quantitative information relating to a machine learning model. The processor acquires user input information input from a user. The processor selects, based on the quantitative information and the user input information, an item to be displayed and a visualizing method from display information definitions that include items relating to 10 evaluation of the machine learning model and visualizing methods of the evaluation. The processor generates, based on the selected result, display information that includes at least partial content of the quantitative information and the user input information.