Hierarchical Training Data Construction for Multi-Device Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine-learning systems face challenges in constructing hierarchical training data sets that are applicable to specific controlled devices, leading to inefficiencies due to data anomalies from differing free inputs, design variables, and condition variables.
Innovation Solution
A system comprising a central computer system and database constructs hierarchical training data sets by aggregating data from multiple controlled devices, prioritizing real data over simulated values, and using machine-learning to determine optimal control inputs, thereby minimizing data anomalies and improving training accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If training data is aggregated from multiple controlled devices with differing free inputs, design variables, and condition variables, then the quantity and diversity of training data increases, but data anomalies increase and training accuracy decreases
Solution Approach 1:
The patent segments training data into hierarchical groups based on shared characteristics (free inputs, design variables, condition variables). Devices with similar characteristics are grouped together, allowing the system to aggregate data across multiple devices while maintaining data quality within each segment. This segmentation resolves the contradiction by enabling quantity increase through aggregation while preserving accuracy through characteristic-based grouping.
Solution Approach 2:
The patent applies local quality by ensuring that training data within each hierarchical group maintains consistent characteristics (same free inputs, design variables, or condition variables). This allows each segment to have high local data quality while the overall system benefits from aggregated quantity across multiple segments with different characteristics.
2Reliability
If real operational data is prioritized over simulated values, then training relevance to specific conditions improves, but data coverage and completeness may be reduced
Solution Approach 1:
The patent uses simulated data as preliminary training data before real operational data becomes available. This preliminary action allows the machine learning model to be initially trained with simulated values that cover various scenarios, and then subsequently refined using real operational data as it accumulates. This resolves the contradiction by ensuring data coverage through simulated preliminary data while improving reliability when real data is available.
3Measurement precision
If hierarchical training data sets are constructed by grouping devices with shared characteristics, then training effectiveness for specific device types improves, but system complexity increases
Solution Approach 1:
The patent creates a universal hierarchical data construction system that can handle multiple device types by identifying shared characteristics. The same data grouping and training methodology is applied universally across different device types (e.g., water heaters, furnaces, air conditioners) by categorizing them based on their free inputs, design variables, and condition variables. This resolves the contradiction by improving training effectiveness through targeted grouping while managing complexity through a universal classification framework.
Data Source
AI summary
Some embodiments include a system operable to construct hierarchical training data sets for use with machine-learning for multiple controlled devices. Other embodiments of related systems and methods are also provided.