Quantum Machine Learning for 6G Communication Networks: State-of-the-Art and Vision for the Future

The upcoming 5th Generation (5G) of wireless networks is expected to lay a foundation of intelligent networks with the provision of some isolated Artificial Intelligence (AI) operations. However, fully-intelligent network orchestration and management for providing innovative services will only be realized in Beyond 5G (B5G) networks. To this end, we envisage that the 6th Generation (6G) of wireless networks will be driven by on-demand self-reconfiguration to ensure a many-fold increase in the network performanceandservicetypes.Theincreasinglystringentperformancerequirementsofemergingnetworks may finally trigger the deployment of some interesting new technologies such as large intelligent surfaces, electromagnetic-orbital angular momentum, visible light communications and cell-free communications – tonameafew.Ourvisionfor6Gis–amassivelyconnectedcomplexnetworkcapableofrapidlyresponding to the users’ service calls through real-time learning of the network state as described by the network-edge (e.g., base-station locations, cache contents, etc.), air interface (e.g., radio spectrum, propagation channel, etc.), and the user-side (e.g., battery-life, locations, etc.). The multi-state, multi-dimensional nature of the network state, requiring real-time knowledge, can be viewed as a quantum uncertainty problem. In this regard, the emerging paradigms of Machine Learning (ML), Quantum Computing (QC), and Quantum ML (QML) and their synergies with communication networks can be considered as core 6G enablers. Considering these potentials, starting with the 5G target services and enabling technologies, we provide a comprehensivereviewoftherelatedstate-of-the-artinthedomainsofML(includingdeeplearning),QCand QML, and identify their potential benefits, issues and use cases for their applications in the B5G networks. Subsequently,weproposeanovelQC-assistedandQML-basedframeworkfor6Gcommunicationnetworks whilearticulatingitschallengesandpotentialenablingtechnologiesatthenetwork-infrastruct