Inference Result Display with Reliability Highlighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional inference apparatuses face challenges in maintaining accurate inference results for unknown data attributes, especially in medical diagnosis, due to insufficient known data sets and changing characteristics over time, leading to unreliable results.
Innovation Solution
An information processing apparatus that displays inference results for multiple attributes and conditions, allowing users to select parameters for specifying inference conditions, highlighting the reliability of each result based on factors like the number of learning data sets or inference accuracy, and enabling users to weight and composite these results for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If supervised learning is performed with available known data sets, then inference accuracy can be improved, but when the number of known data sets is insufficient, the inference accuracy cannot be maintained at a satisfactory level
Solution Approach 1:
The patent changes the parameter of inference conditions by allowing users to select different parameters (number of learning data sets, inference accuracy, etc.) to specify inference conditions. This enables the system to adapt to situations with limited data by adjusting inference conditions rather than being constrained by fixed data requirements
Solution Approach 2:
The patent introduces dynamic inference conditions that can be adjusted based on available data and user needs. Instead of static supervised learning with fixed data requirements, the system dynamically adapts inference conditions to maintain reliability even when known data sets are insufficient
2Productivity
If inference apparatuses are used to provide diagnosis support, then productivity can be improved, but users cannot determine the reliability of inference results, leading to distrust and ineffective use
Solution Approach 1:
The patent implements feedback by displaying reliability information about inference results to users. The system provides information on the number of learning data sets used, inference accuracy, and other parameters that help users assess the reliability of each inference result, enabling informed decision-making
Solution Approach 2:
The patent introduces an intermediary layer that translates complex inference processes into understandable reliability metrics for users. By displaying inference conditions and reliability information in an accessible format, it bridges the gap between the inference apparatus and the end user, building trust and effective utilization
3Measurement precision
If multiple inference apparatuses with different inference methods are used to improve accuracy through weighting and integration, then device complexity increases
Solution Approach 1:
The patent makes a single inference apparatus universal by enabling it to perform multiple inference operations under different conditions. Instead of requiring separate apparatuses for different inference methods, the system can switch between different inference conditions using the same underlying apparatus, reducing complexity while maintaining accuracy
Data Source
AI summary
An information processing apparatus includes a display control unit adapted to, with respect to each of a plurality of predetermined attributes, display on a display unit, for each of a plurality of inference conditions, inference results obtained by inferring a probability that input data belongs to the attribute under the inference condition; and a receiving unit adapted to receive selection of one of a plurality of parameters for specifying an inference condition from a user, wherein the display control unit highlights, with respect to each of the plurality of inference conditions, the inference results according to values of the selected parameter of the inference condition.