|
چکیده
|
The increasing demand for high data rates has led to a significant expansion in communication systems, network size and data processing. This demand has also increased the risk of security threats, as new attacks emerge or existing attacks evolve. To deal with these threats, intrusion detection systems (IDS) are considered as a feasible solution [1). An IDS or intrusion detection system analyzes all network activity and, using the information in its database, determines whether the activity is authorized or uncial, and detect whether this.activity can affect the network. You will be harmed or not and will info you in the end. Regarding such activities in recent years, .the use of artificial intelligence networks [2] to detect intrusion to increase cyber security is an extremely attractive and comprehensive solution [3). Various solutions have been proposed for IDSs. But the current US have two problems: low accuracy and slow speed [4}. Researchers are now investigating IDS solutions that use artificial intelligence to increase detection accuracy and reduce false alarm rates [4]. Deep networks [4) are the best tools among,artificial intelligence methods for intrusion detection. Deep leaning methods themselves consist of several subsets. Clearly, different networks such as RNN-CNN-LSTM-GRU-GAN-... are different neural networks that can be used as artificial intelligence tools to detect network intrusion. A special category of these networks, namely recurrent networks, is commonly used for intrusion detection methods. But ,these networks have a low speed and a high processing load due to the fact that they operate in a recurrent manner. On the other hand, deep neural network CNN is a very fast and accurate network.
|