Wireless Flow Characteristic Adjustment for Rare Resource-Demand Samples
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing supervised machine learning models for resource demand prediction in wireless networks struggle with insufficient and biased data, leading to poor generalization and underfitting, particularly in rare or infrequently occurring scenarios.
Innovation Solution
A system and method to adjust flow characteristics in wireless networks to create training samples for underrepresented scenarios by mimicking the treatment of rare applications, using a QoS-aware scheduler to temporarily override actual QoS requirements, thereby obtaining samples without artificial traffic generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If supervised machine learning models use existing training data for resource demand prediction, then the model can be trained with available data, but the data is insufficient and biased leading to poor generalization and underfitting in rare scenarios
Solution Approach 1:
The system performs preliminary analysis of training data to identify underrepresented scenarios before model training. By proactively detecting data gaps and generating synthetic samples for rare scenarios in advance, the system ensures comprehensive training data coverage without waiting for natural occurrence of rare events during model operation.
Solution Approach 2:
The system creates synthetic copies of training samples by adjusting flow characteristics of existing data flows to mimic rare scenarios. Instead of requiring actual rare events to occur, the system copies and modifies existing flow data to generate representative samples for underrepresented conditions, enabling the model to learn from synthesized rare scenario data.
2Ease of manufacture
If supervised machine learning models are trained with limited diverse data, then training can proceed with available resources, but the model fails to capture essential characteristics and misses relevant relations between variables
Solution Approach 1:
The system changes flow characteristics parameters (such as data rate, latency requirements, QoS parameters) of existing data flows to generate synthetic samples representing rare scenarios. By systematically varying these parameters, the system creates diverse training data that maintains physical realism while covering underrepresented conditions, thereby improving model adaptability without requiring new data collection infrastructure.
3Quantity of substance
If the system generates training data by adjusting flow characteristics of existing data flows, then diverse training samples can be obtained without artificial traffic generation, but the system complexity increases due to data analysis and flow characteristic adjustment mechanisms
Solution Approach 1:
The system uses existing data flows and their inherent characteristics to generate training samples, rather than requiring separate artificial traffic generation infrastructure. By leveraging real network flows and adjusting their characteristics, the system makes the existing traffic serve dual purposes: actual data transmission and training data generation, thereby reducing overall system complexity.
Solution Approach 2:
The system implements a multi-functional mechanism where the same data flow serves both as actual network traffic and as source material for generating training samples. The flow characteristic adjustment mechanism enables a single data flow to be used for multiple purposes: maintaining network operation and providing diverse training data, thus reducing the need for separate specialized systems.
Data Source
AI summary
A system (1, 21) for obtaining samples for a dynamic resource management system in a wireless network is configured to analyze (101) data for a dynamic resource management system to determine one or more combinations of input values which are not present or underrepresented in the data. The dynamic resource management system uses supervised learning to estimate resource demand. The system is further configured to adjust (103) one or more flow characteristics of a data flow between a user device and a base station to obtain one of the combinations of input values, determine a quantity of resources used for the data flow and actual state and/or cell characteristics relevant to the data flow, create a new sample based on the used quantity of resources, the adjusted flow characteristics, and the actual state and/or cell characteristics, and store the new sample in the data.