Learned Model Control for Sorting IoT Training Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing systems for IoT devices struggle to efficiently sort essential from non-essential data, leading to excessive data handling and processing burdens on servers and communication equipment.
Innovation Solution
A control system comprising a first processing device that generates a learned model and a second processing device that operates using this model, where the second device transmits input information to the first device for updating the model when it belongs to a different area, allowing for data sorting and reduction in data handling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data from a plurality of on-board devices is stored in a continuous manner, then the learning operation can be performed with sufficient data, but the amount of data to be handled increases significantly
Solution Approach 1:
The patent extracts only the essential data features needed for learning operations by introducing a data sorting mechanism that identifies and removes redundant information. The sorting unit selectively extracts important data characteristics while discarding unnecessary data, thereby maintaining learning quality with reduced data volume.
Solution Approach 2:
The patent changes the parameter of data representation by transforming raw continuous data into sorted data with reduced dimensions. The data sorting mechanism alters the data structure to retain only essential features, effectively changing the data parameters from comprehensive to selective representation.
2Reliability
If all data is processed and stored continuously, then comprehensive learning can be achieved, but the processing burden on servers and communication equipment increases
Solution Approach 1:
The patent extracts only the essential data features needed for learning operations by introducing a data sorting mechanism that identifies and removes redundant information. The sorting unit selectively extracts important data characteristics while discarding unnecessary data, thereby maintaining learning quality with reduced data volume.
Solution Approach 2:
The patent segments the data processing function into two parts: data collection at the source and selective sorting/processing at the server. The sorting unit divides the data stream into essential and non-essential components, processing only the necessary portions through the system.
3Reliability
If data sorting is not performed, then all data can be used for learning, but the amount of data to be handled increases unnecessarily
Solution Approach 1:
The patent extracts only the essential data features needed for learning operations by introducing a data sorting mechanism that identifies and removes redundant information. The sorting unit selectively extracts important data characteristics while discarding unnecessary data, thereby maintaining learning quality with reduced data volume.
Solution Approach 2:
The patent changes the parameter of data representation by transforming raw continuous data into sorted data with reduced dimensions. The data sorting mechanism alters the data structure to retain only essential features, effectively changing the data parameters from comprehensive to selective representation.
Data Source
AI summary
In view of the relevant background art, the present invention is to provide a system or apparatus which enables data in use for learning to be sorted out, thereby reducing the amount of data to be handled. Presenting one example of the control system according to the present invention, the control system including a first processing device to generate a learnt model; and a second processing device to operate employing the generated learnt model and provided with input means, wherein the second processing device transmits information input from the input means to the first processing device; and where the input information as transmitted is information belonging to an area different from an area designated with the learnt model, the first processing device generates an updated learnt model based on the information belonging to the different area as received and the learnt model.