Deep Learning Data Traffic Analysis for Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The analysis of large and varying data traffic in devices becomes cumbersome, necessitating the need for improved systems and methods to detect anomalies effectively.
Innovation Solution
A method utilizing a deep learning algorithm, operable on a processing unit, including a graphical processing unit, that reconstructs data traffic by accounting for past data to update its model, allowing for anomaly detection in real-time or offline training scenarios, particularly in devices like cars.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional data analysis methods are used on large data traffic, then anomaly detection can be performed, but the analysis becomes cumbersome and slow
Solution Approach 1:
The patent replaces traditional mechanical data analysis methods with a deep learning-based automated system. The neural network automatically learns patterns and anomalies from data traffic without requiring manual analysis rules, thereby increasing processing speed while managing complexity through algorithmic automation.
Solution Approach 2:
The deep learning system performs self-training by automatically updating its own model parameters based on reconstructed data errors. The system serves itself by continuously improving its anomaly detection capability through unsupervised learning from the data traffic patterns it analyzes.
2Productivity
If deep learning algorithms are used for data traffic analysis, then analysis speed increases significantly, but computational resources and processing complexity increase
Solution Approach 1:
The system performs preliminary training offline before deployment, preparing the deep learning model in advance. This preliminary action allows the model to be optimized and ready for fast inference during actual data traffic analysis, reducing real-time computational burden while maintaining high detection speed.
Solution Approach 2:
The system focuses computational resources on reconstructing and analyzing only the most relevant parts of data traffic patterns that contain anomaly information. By selectively processing critical data portions rather than uniformly analyzing all data, the system achieves high detection speed with reduced overall computational resource consumption.
Data Source
AI summary
A method of analyzing data exchange of at least one device includes feeding a plurality of data exchanged by the at least one device to a system for data exchange analysis that includes a deep learning algorithm. The deep learning algorithm includes at least an input layer, an output layer of the same size as the input layer, and hidden layers. Neurons of the hidden layers receive recurrently, at each time t, only a subset of the data exchanged by the at least one device up to time t, the subset of data comprising current data from time t and only a fraction of past data from time tpast to time t, with tpast<t. The method includes attempting to reconstruct, at the output layer, at each time t, data received at the input layer. The reconstructed data is compared with at least part of the plurality of data. In indication is provided on one or more anomalies in the data, based on at least the comparison.