GAN and DRL Based Intent Translation and Deep Fake Configuration Generation for Optimization
Talha Ahmed Khan, Khizar Abbas, Muhammad Afaq, Adeel Rafiq, Wang‐Cheol Song · 2020
The requirements of next-generation of networks have imposed a multi-dimensional/ directional complexity over network management. As a result, the only way forward is through the automation of administrative and management procedures of networks. However, the best solution for having an automation system is through Machine Learning. The next-generation network services have complex requirements that can be determined using service graphs. However, autonomous translation of high-level user requirements to a service graph is itself a complex problem and requires an intelligent solution. Hence in this manuscript, a GAN (generative adversarial network) based fake service graph generation and DRL (Deep reinforcement learning) based optimized graph selection mechanism is considered for automatic intent translation.