Wireless AI Data Collection Using KPI-Based Conditional Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for wireless data traffic in 5G/NR communication systems requires efficient data collection mechanisms to support various applications and deployments, while existing data collection methods often result in excessive traffic and storage demands, especially when utilizing AI/ML modules, without ensuring the collection of meaningful data.
Innovation Solution
An AI-aided data collection framework that includes a conditional data collection unit and model training unit, utilizing a data collection module, data sample evaluation module, and AI model to selectively gather data based on predefined conditions, followed by on-demand model fine-tuning and retraining to optimize AI module performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional data collection methods are used to support AI/ML modules in wireless systems, then more data can be collected for model training, but wireless traffic and storage demands increase excessively
Solution Approach 1:
The system performs preliminary actions by having the network device generate collection condition configurations before actual data collection occurs. These configurations include predefined conditions that determine what data should be collected, allowing the terminal device to filter data locally before transmission, thus reducing unnecessary wireless traffic and storage demands while still gathering sufficient training data for AI/ML models
Solution Approach 2:
The invention extracts only the necessary and meaningful data from the overall data stream by implementing collection condition configurations that specify particular criteria for data selection. This selective extraction approach ensures that only data meeting predefined conditions is collected and transmitted, significantly reducing the volume of wireless traffic and storage requirements compared to collecting all available data
2Quantity of substance
If comprehensive data collection is performed without selection criteria, then more training data is available for AI models, but the collected data may include unnecessary information reducing training efficiency
Solution Approach 1:
The system applies local quality by implementing collection condition configurations that define specific criteria for different types of data collection. Each condition configuration tailors the data collection process to local requirements, ensuring that only data with relevant characteristics for specific AI/ML training purposes is collected, thereby maintaining high data quality and meaningfulness
Solution Approach 2:
The network device performs preliminary actions by generating collection condition configurations before data collection begins. These configurations establish quality filters in advance, ensuring that only meaningful data meeting predetermined criteria is collected, thus preventing the collection of unnecessary information that would reduce training efficiency
3Reliability
If AI models are trained with insufficient or irrelevant data, then training resources are wasted, but model performance does not improve
Solution Approach 1:
The system implements feedback mechanisms where the network device receives data collection capability reports from terminal devices and generates appropriate collection condition configurations based on this feedback. This closed-loop approach ensures that data collection is continuously optimized to match actual AI/ML training needs, improving model training effectiveness while avoiding waste of training resources on irrelevant data
Solution Approach 2:
The network device performs preliminary actions by generating collection condition configurations before data collection and training begin. These configurations are based on feedback about terminal capabilities and training requirements, ensuring that data collection is optimized in advance to provide high-quality training data, thereby improving training effectiveness while reducing resource consumption
Data Source
AI summary
A method includes: transmitting, by a first electronic device, a data collection capability report to a second electronic device in response to a request, the data collection capability report including identifications of associated artificial intelligence (AI) models and key performance indicators (KPIs) that each of the AI models is capable of evaluating; receiving, by the first electronic device, a data collection configuration message from the second electronic device, the data collection configuration message including an enablement status for each of the KPIs; receiving, by the first electronic device, a data collection request from the second electronic device, the data collection request including a collection condition configuration associated with each of the KPIs; and collecting, by the first electronic device, data samples based on the collection condition configuration to generate and transfer a data package including collected data samples satisfying the collection condition configuration.