Automated AI Model Re-Learning in M2M Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine-to-machine (M2M) systems lack effective mechanisms for automated re-learning of artificial intelligence (AI) models, which is essential for maintaining accuracy and adaptability in dynamic environments.
Innovation Solution
A method and apparatus for generating resources for training AI models, performing initial learning, collecting learning data, and triggering re-learning based on predefined criteria, utilizing a transceiver and processor to manage the AI model's training and re-training processes within the M2M system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If AI models are trained initially and then used in M2M systems, then the system can operate with pre-existing knowledge, but the model cannot adapt to dynamic environmental changes
Solution Approach 1:
The system performs preliminary actions by collecting and storing learning data during normal operation, so that when re-learning is needed, the data is already prepared and can be used immediately. This reduces the time loss associated with re-learning by having the necessary resources ready in advance.
Solution Approach 2:
The system implements feedback mechanisms where learning data is continuously collected from system operation and used to update the AI model. This feedback loop enables the model to adapt to dynamic environmental changes while the collected data serves as the foundation for efficient re-learning without time loss.
2Extent of automation
If re-learning is performed manually, then control and monitoring are easier, but automation capability and system efficiency are reduced
Solution Approach 1:
The system performs self-service by automatically collecting learning data during operation and initiating re-learning processes without manual intervention. The system monitors its own performance, gathers necessary data, and executes re-learning autonomously, thereby increasing automation while managing complexity through integrated design.
Solution Approach 2:
The system merges multiple functions including data collection, performance monitoring, and model re-learning into a unified automated process. By combining these functions that would otherwise be separate manual tasks into an integrated system, automation is achieved without proportionally increasing complexity.
3Measurement precision
If learning data is collected continuously, then model accuracy is maintained, but data storage and processing requirements increase
Solution Approach 1:
The system applies local quality by selectively collecting and storing only the learning data that is relevant and useful for model improvement. Rather than treating all data uniformly, the system identifies and retains data with specific qualities that contribute to accuracy, reducing unnecessary data volume while maintaining precision.
Solution Approach 2:
The system discards redundant or less useful learning data while recovering and retaining only the essential data needed for model re-learning. This selective approach maintains prediction accuracy by preserving valuable learning patterns while reducing the overall data volume that requires storage and processing.
Data Source
AI summary
The present disclosure may support automated re-learning in a machine-to-machine (M2M) system. A method for operating a device may include: generating a resource for training an artificial intelligence (AI) model; controlling to perform initial learning of the AI model; collecting learning data for re-learning for the AI model; and controlling to perform re-learning of the AI model by using the learning data.