Prelearned Recognition Model Selection by Training Attribute Diversity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques struggle to select a model with high recognition performance that matches the attributes of a recognition target from a plurality of models, particularly when the models are learned using different data sets.
Innovation Solution
An information processing apparatus that learns multiple models from subsets of data sets with specific attributes and stores these models along with their attribute information, allowing for the selection of a model that best matches the attributes of the recognition target data by comparing attribute sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large number of applications are managed in the background, then system functionality is improved, but power consumption increases
Solution Approach 1:
The patent extracts and separates the power management function from the application management system. A dedicated power management module monitors and controls power consumption independently, allowing the system to maintain full application functionality while separately optimizing power usage through selective suspension and prioritization of background applications.
Solution Approach 2:
The system dynamically changes the operational state of background applications based on power conditions. Applications are transitioned between active, suspended, and terminated states depending on available power resources and system priorities, allowing functionality to be maintained when power is abundant while conserving energy when power is limited.
2Use of energy by moving object
If background applications are suspended or terminated, then power consumption is reduced, but user convenience deteriorates
Solution Approach 1:
The system performs preliminary actions by proactively managing background application states before power depletion occurs. It pre-suspends or terminates lower-priority applications in advance, and maintains higher-priority applications ready for immediate resumption, thereby reducing power consumption while minimizing disruption to user experience when applications need to be accessed.
Solution Approach 2:
The power management module implements feedback mechanisms that monitor user interaction patterns and system state. When users access suspended applications, the system learns from this behavior and adjusts its suspension/termination decisions, ensuring that frequently used applications are preserved while less-used ones are suspended or terminated to save power.
3Ease of operation
If all applications are kept active, then user convenience is improved, but battery life is reduced
Solution Approach 1:
The patent segments the application portfolio into different priority levels and operational states. Background applications are divided into high-priority (maintained active), medium-priority (suspended when power is limited), and low-priority (terminated when power is scarce). This segmentation allows the system to extend battery life by selectively managing application states while ensuring that critical applications remain accessible for user convenience.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
An information processing apparatus comprising: first acquiring means for acquiring information about learning data used in learning of each of a plurality of prelearned models for recognizing input data; second acquiring means for acquiring information indicating an attribute of recognition target data; and model selecting means for selecting a model to be used in recognition of the recognition target data from the plurality of models, based on a degree of matching between the attribute of the recognition target data and an attribute of the learning data used in learning of each of the plurality of models, and on diversity of the attribute of the learning data used in learning of each of the plurality of models.