Search

The advancing applications based on machine learning and deep learning in communication networks have been exponentially increasing in the system architectures of enabled software-defined networking, network functions virtualization, and other wired/wireless networks. With data exposure capabilities of graph-structured network topologies and underlying data plane information, the state-of-the-art deep learning approach, graph neural networks (GNN), has been applied to understand multi-scale deep correlations, offer generalization capability, improve the accuracy metrics of prediction modelling, and empower state representation for deep reinforcement learning (DRL) agents in future intelligent network management and orchestration. This paper contributes a taxonomy of recent studies using GNN-based approaches to optimize the control policies, including offloading strategies, routing optimization, virtual network function orchestration, and resource allocation. The algorithm designs of converged DRL and GNN are reviewed throughout the selected studies by presenting the state generalization, GNN-assisted action selection, and reward valuation cooperating with GNN outputs. We also survey the GNN-empowered application deployment in the autonomous control of optical networks, Internet of Healthcare Things, Internet of Vehicles, Industrial Internet of Things, and other smart city applications. Finally, we provide a potential discussion on research challenges and future directions.

International audience; Recent studies have shown that Plasmodium falciparum malaria parasites in Pailin province, along the border between Thailand and Cambodia, have become resistant to artemisinin derivatives. To better define the epidemiology of P. falciparum populations and to assess the risk of the possible spread of these parasites outside Pailin, a new epidemiological tool named "Focused Screening and Treatment" (FSAT), based on active molecular detection of asymptomatic parasite carriers was introduced in 2010. Cross-sectional malariometric surveys using PCR were carried out in 20 out of 109 villages in Pailin province. Individuals detected as P. falciparum carriers were treated with atovaquone-proguanil combination plus a single dose of primaquine if the patient was non-G6PD deficient. Interviews were conducted to elicit history of cross-border travel that might contribute to the spread of artemisinin-resistant parasites. After directly observed treatment, patients were followed up and re-examined on day 7 and day 28. Among 6931 individuals screened, prevalence of P. falciparum carriers was less than 1%, of whom 96% were asymptomatic. Only 1.6% of the individuals had a travel history or plans to go outside Cambodia, with none of those tested being positive for P. falciparum. Retrospective analysis, using 2010 routine surveillance data, showed significant differences in the prevalence of asymptomatic carriers discovered by FSAT between villages classified as "high risk" and "low risk" based on malaria incidence data. All positive individuals treated and followed-up until day 28 were cured. No mutant-type allele related to atovaquone resistance was found. FSAT is a potentially useful tool to detect, treat and track clusters of asymptomatic carriers of P. falciparum along with providing valuable epidemiological information regarding cross-border movements of potential malaria parasite carriers and parasite gene flow.