Twin Inference Model for Data Aggregation Resource Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high computing resource consumption during data aggregation in distributed systems, particularly in resource-constrained environments like edge nodes and autonomous vehicles, due to the transmission of large data quantities, which hinders efficient data collection and processing.
Innovation Solution
Implementing a data reduction process using twin inference models at the data aggregator and collector, where inferences are generated locally and revised over time to improve accuracy, reducing the need for extensive data transmission and conserving computing resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data is transmitted in full from data collector to data aggregator, then data accuracy is maintained, but computing resource consumption increases
Solution Approach 1:
The patent segments data transmission into two parts: (1) transmission of only reconstructed data differences or residuals from the twin inference model, and (2) local storage and processing of original data at the data collector. This segmentation reduces transmission volume while maintaining accuracy by only sending the portions that differ from the reconstructed data.
Solution Approach 2:
The patent creates a twin inference model that generates reconstructed data as a copy of the original data. This copy is then used for comparison with actual data, and only the differences are transmitted. The copying approach allows the system to work with smaller data representations while preserving the ability to reconstruct full accuracy when needed.
2Use of energy by moving object
If data is reduced in size before transmission, then computing resource consumption decreases, but data accuracy deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the data aggregator compares the reconstructed data (generated from reduced-size transmitted data) with the actual received data. The differences or residuals are fed back to the data collector, which then adjusts future data reduction processes to maintain accuracy while continuing to reduce transmission volume.
Solution Approach 2:
The patent employs dynamic data reduction where the level of data reduction is adjusted based on the specific characteristics of the data being transmitted. The twin inference model dynamically determines how much data can be reduced while maintaining acceptable accuracy levels, allowing the system to adapt the reduction ratio based on data importance and variability.
3Quantity of substance
If twin inference models are used for data reconstruction, then data transmission volume decreases, but model complexity increases
Solution Approach 1:
The patent designs the twin inference model to serve multiple functions: (1) generating reconstructed data from reduced-size inputs, (2) enabling comparison with actual data to determine transmission needs, and (3) providing a framework for future model updates and improvements. This multi-functionality justifies the model complexity by deriving multiple benefits from a single computational component.
Solution Approach 2:
The patent performs preliminary training and setup of the twin inference model during system initialization or offline periods. This preliminary action prepares the model to efficiently handle real-time data reconstruction with minimal computational overhead during actual data transmission operations, reducing the perceived complexity during runtime.
Data Source
AI summary
Methods and systems for managing data collection are disclosed. A data aggregator may aggregate data collected by a data collector. To reduce computing resources used for aggregation, the data aggregator and data collector may use inferences provided by a twin inference model in place of data collected by the data collector rather than receiving copies of data from the data collector. Over time, the aggregated data may be revised using revised inference models that are revised using subsequently obtained data from the data collector. The revised inference models may be used to obtain revised inferences that may replace original inferences in the aggregated data. The revised inferences may be of higher accuracy due to differences in the data upon which the inference and revised inference models are based.