Non-Public Network ML Training With Mobile Data Collection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning proves challenging in non-public communication networks due to limited training data, which hampers effective management and optimization of these networks.
Innovation Solution
The use of automated or autonomous mobile devices to collect additional training data by determining beneficial locations and revising their routes to include these locations, iteratively enriching the training dataset to validate and improve the machine learning model's performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning is applied in a non-public communication network, then network management capability is improved, but training data sufficiency deteriorates due to limited network scope and device count
Solution Approach 1:
Automated or autonomous mobile devices are made multi-functional by enabling them to both perform their primary operational tasks and collect training data for machine learning models. The same devices that execute industrial tasks also gather network performance data, location information, and operational metrics, thereby serving dual purposes without requiring additional dedicated hardware.
Solution Approach 2:
The system proactively identifies locations and conditions where additional training data would be beneficial before actual model training occurs. By analyzing current model performance and data distribution, the system determines in advance what data is needed and directs mobile devices to collect it during their normal operations, ensuring data availability before training requirements arise.
2Adaptability or versatility
If automated or autonomous mobile devices are re-routed to collect additional training data, then training data diversity is improved, but device operational efficiency deteriorates due to route revisions
Solution Approach 1:
Instead of completely re-routing mobile devices, the system implements partial route adjustments by adding strategic waypoints to existing routes. Devices follow their primary operational paths but deviate temporarily to visit selected locations for data collection, then return to their original trajectories. This partial modification maintains operational efficiency while still achieving data diversity goals.
Solution Approach 2:
The system merges data collection tasks with the existing operational routes of mobile devices. Rather than creating separate dedicated data collection routes, the approach combines both objectives by integrating data collection waypoints into the devices' primary task paths, thereby eliminating redundant travel and maintaining operational productivity.
3Measurement precision
If iterative training and evaluation of the machine learning model is performed, then model accuracy is improved, but training time and computational resources deteriorate
Solution Approach 1:
The system implements a feedback-driven iterative training process where model performance is continuously evaluated and used to guide subsequent data collection and re-training decisions. After each training iteration, the system assesses model accuracy on validation data and determines whether additional data collection is necessary, creating a closed-loop process that adapts to actual model needs rather than following a fixed training schedule.
Solution Approach 2:
The machine learning model training process becomes self-service by automatically determining when additional training data is needed based on validation performance. The system autonomously evaluates model adequacy, identifies data gaps, directs mobile devices to collect necessary data, and triggers re-training only when beneficial, eliminating the need for manual intervention or predetermined training schedules.
Data Source
AI summary
Equipment that supports a non-public communication network trains a machine learning model with a training dataset to make a prediction or decision in the network. The equipment determines whether the trained model is valid or invalid based on whether predictions or decisions that the trained model makes from a validation dataset satisfy performance requirements. Based on the trained model being invalid, the equipment analyzes the training dataset and/or the trained model to determine what additional training data to add to the training dataset. The equipment transmits signaling for configuring one or more autonomous or automated mobile devices served by the network to help collect the additional training data. The equipment then re-trains the model with the training dataset as supplemented with the additional training data.