A Review Study of AI Enabled Computer Network

With the advent of pandemic the role of digital communication and in particular computer networks in day to day life has become very predominant and crucial, this has in turn increased the network usage and traffic leading to many different challenges faced by network architects and engineers, on the other hand there had been major growth and development within the field of artificial intelligence (AD and machine learning(ML) in past decade, hence artificial intelligence and in particular, deep learning models can be adapted to decrease the manual interventions and increase the quality of services(QoS)for the computer networks, network security management and many more challenges arising in computer network technology. This paper aims to explore some of the deep learning models and algorithms designed and adopted to effectively deal with different problems such as network routing automation and optimization, classification of network traffic, intrusion detection, monitoring of various network dynamics, checking for network resources availability, detecting anomalies, analyzation of network data, optimization of tracking control so on.

I Introduction

During the past few years the world has seen a major growth and usage of the computer networks in all the sectors be it for business, education, healthcare, transportation, logistics etc., this had a major increase on the data and network traffic and challenges raised by them increased twofold. Deep learning architectures can be easily deployed to mitigate some of the challenges faced by the computer network industry such as the reinforcement learning can be used to assist the traditional routing strategies, recurrent neural network (RNN) based smart agents can be used for routing optimization in traffic engineering to minimize the human errors and interventions on the other hand a multilayer perceptron model (MLP) with Markov Decision process(MDP) control architectures has been proven effective to get the desired policy for knowledge transfer process that can be easily used for deep transfer learning in networking.