Temporal Data Image Encoding for Efficient CNN Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data processing systems face challenges in accurately analyzing temporal changes in data due to resource-intensive computations and loss of information through data compression, leading to inaccurate predictions and inefficient use of computational resources.
Innovation Solution
Transforming historical data into images using convolutional neural networks (CNNs) to maintain the temporal sequence of data, allowing for efficient analysis of temporal patterns and improved prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If traditional data compression methods are used to reduce computational resources, then computational resource consumption decreases, but information loss increases and prediction accuracy deteriorates
Solution Approach 1:
The patent transforms temporal data from a one-dimensional time series into a two-dimensional image representation. This dimensional transformation allows the data to be processed by CNNs which can capture temporal patterns and relationships without requiring aggressive compression, thus reducing information loss while maintaining computational efficiency through the power of visual pattern recognition algorithms.
2Speed
If data compression is applied to reduce processing load, then processing speed increases, but measurement precision of temporal changes decreases
Solution Approach 1:
The patent replaces traditional mechanical data processing methods (averaging, sampling) with a neural network-based approach. The CNN automatically learns and extracts temporal patterns from the image representation of data, eliminating the need for manual compression techniques that sacrifice precision. This substitution enables both high processing speed through parallel computation and high measurement precision through automated feature extraction.
3Use of energy by moving object
If traditional prediction techniques are used, then computational resource usage is reduced, but prediction accuracy deteriorates
Solution Approach 1:
The patent creates a visual copy or representation of temporal data in the form of an image, which can then be processed by CNNs. This copying approach allows the system to leverage the computational efficiency of image processing algorithms while maintaining the full detail and resolution of the original temporal data, achieving both low computational resource usage and high prediction accuracy.
Data Source
AI summary
Techniques are disclosed for transforming user data into images for training a machine learning model based on temporal changes in the users' variables. A system receives a request from a device and retrieves a historical user data that includes variables of a user of the device. The system separates, based on a particular time interval, the historical data into subsets that include historical data for the particular time interval at different times. The system generates, based on the subsets, an image that includes rows of pixels corresponding to the variables included in the historical data and columns of pixels corresponding to the subsets placed in temporal order according to the different times at which their particular time interval occurs. Based on the image, the system determines whether to authorize the request by inputting the image into a machine learning model trained on images of historical data for different users.