Web-of-Open-Science / QRL

Quantum Review Letters
0 stars 0 forks source link

Implementing Embedded Uniqueness for Naturally One-to-One Monoids in a High Speed Learning Neural Network for Cyber Defense #4

Open utterances-bot opened 4 years ago

utterances-bot commented 4 years ago

Implementing Embedded Uniqueness for Naturally One-to-One Monoids in a High Speed Learning Neural Network for Cyber Defense Software Engineering Review

 

With the latest generation of cybercrime, characterized by the use of machine learning, self-modeling and automation of the spread of attack tools, as well as integration across several sets of tools, the overwhelming number of experts believe that traditional security methods are already poorly productive and are a qualitatively new approach for the implementation of network and information security is needed. The imperfection of modern methods of protection against attacks against external unauthorized traffic leads to the fact that many companies whose resources have access to the Internet face the inaccessibility of their own services providing different services / information. This report is dedicated to the application and experimentation of one specific artificial intelligence method for protecting network servers and hosts on the network. A neural network has been developed with high self-learning speed and fast response to DDoS attacks. A new neural network training method has been developed, based on the uniform processing of messages from each neuron, this mechanism provides fast and efficient attack filtering. A comparison of the developed neural network with other similar solutions was made, which showed that the proposed neural network is more optimized for high loads and is able to detect and neutralize DDoS attacks in the shortest time. Long-term idle neural network testing and protection against DDoS attacks shows relatively low CPU, RAM, and SSD load during massive DDoS attacks.  

https://www.scholarchain.eu/ser/index

nraychev commented 4 years ago

To comment on this article, you must first log in with your GitHub account