Next-generation wireless networks are expected to support extremely high data rates and radically new applications, which require a new wireless radio technology paradigm. The challenge is that of assisting the radio in intelligent adaptive learning and decision making, so that the diverse requirements of next-generation wireless networks can be satisfied. Machine learning is one of the most promising artificial intelligence tools, conceived to support smart radio terminals. Future smart 5G mobile terminals are expected to autonomously access the most meritorious spectral bands with the aid of sophisticated spectral efficiency learning and inference, in order to control the transmission power, while relying on energy efficiency learning/inference and simultaneously adjusting the transmission protocols with the aid of quality of service learning/inference. Hence we briefly review the rudimentary concepts of machine learning and propose their employment in the compelling applications of 5G networks, including cognitive radios, massive MIMOs, femto/small cells, heterogeneous networks, smart grid, energy harvesting, device-todevice communications, and so on. Our goal is to assist the readers in refining the motivation, problem formulation, and methodology of powerful machine learning algorithms in the context of future networks in order to tap into hitherto unexplored applications and services.

Future wireless networks have a substantial potential in terms of supporting a broad range of complex compelling applications both in military and civilian fields, where the users are able to enjoy high-rate, low-latency, low-cost and reliable information services. Achieving this ambitious goal requires new radio techniques for adaptive learning and intelligent decision making because of the complex heterogeneous nature of the network structures and wireless services. Machine learning (ML) algorithms have great success in supporting big data analytics, efficient parameter estimation and interactive decision making. Hence, in this article, we review the thirty-year history of ML by elaborating on supervised learning, unsupervised learning, reinforcement learning and deep learning. Furthermore, we investigate their employment in the compelling applications of wireless networks, including heterogeneous networks (Het-Nets), cognitive radios (CR), Internet of things (IoT), machine to machine networks (M2M), and so on. This article aims for assisting the readers in clarifying the motivation and methodology of the various ML algorithms, so as to invoke them for hitherto unexplored services as well as scenarios of future wireless networks.

Abstract-Unmanned Aerial Vehicles (UAV) have been widely used both in military and in civilian applications. However, the cooperation of small and mini drones in a network is capable of further improving both the performance and the coverage area of UAVs. Naturally, there are numerous new challenges to be solved before the wide-spread introduction of multi-UAV based heterogeneous Flying Ad Hoc Networks (FANET), including the formulation of a stable network structure. Meanwhile, an efficient gateway selection algorithm and management mechanism is required as well. On the other hand, the stability control of the hierarchical UAV network guarantees the efficient collaboration of the drones. In this article, we commence with surveying the FANET structure and its protocol architecture. Then, a variety of distributed gateway selection algorithms and cloudbased stability control mechanisms are addressed, complemented by a range of open challenges.Index Terms-Multi-UAV network system of small and mini drones, FANET network structure, distributed gateway selection algorithms, cloud-based stability control for collaboration and cooperation.

The recently proposed self-ensembling methods have achieved promising results in deep semi-supervised learning, which penalize inconsistent predictions of unlabeled data under different perturbations. However, they only consider adding perturbations to each single data point, while ignoring the connections between data samples. In this paper, we propose a novel method, called Smooth Neighbors on Teacher Graphs (SNTG). In SNTG, a graph is constructed based on the predictions of the teacher model, i.e., the implicit self-ensemble of models. Then the graph serves as a similarity measure with respect to which the representations of "similar" neighboring points are learned to be smooth on the low-dimensional manifold. We achieve state-of-theart results on semi-supervised learning benchmarks. The error rates are 9.89%, 3.99% for CIFAR-10 with 4000 labels, SVHN with 500 labels, respectively. In particular, the improvements are significant when the labels are fewer. For the non-augmented MNIST with only 20 labels, the error rate is reduced from previous 4.81% to 1.36%. Our method also shows robustness to noisy labels.

Dynamic spectrum access in cognitive radio networks can greatly improve the spectrum utilization efficiency. Nevertheless, interference may be introduced to the Primary User (PU) when the Secondary Users (SUs) dynamically utilize the PU's licensed channels. If the SUs can be synchronous with the PU's time slots, the interference is mainly due to their imperfect spectrum sensing of the primary channel.However, if the SUs have no knowledge about the PU's exact communication mechanism, additional interference may occur. In this paper, we propose a dynamic spectrum access protocol for the SUs confronting with unknown primary behavior and study the interference caused by their dynamic access.Through analyzing the SUs' dynamic behavior in the primary channel which is modeled as an ON-OFF process, we prove that the SUs' communication behavior is a renewal process. Based on the Renewal Theory, we quantify the interference caused by the SUs and derive the corresponding close-form expressions. With the interference analysis, we study how to optimize the SUs' performance under the constraints of the PU's communication quality of service (QoS) and the secondary network's stability.Finally, simulation results are shown to verify the effectiveness of our analysis.

Characterized by their ease of deployment and bird's-eye view, unmanned aerial vehicles (UAVs) may be widely deployed both in surveillance and traffic management. However, the moderate computational capability and the short battery life restrict the local data processing at the UAV side. Fortunately, this impediment may be mitigated by employing the mobile-edge computing (MEC) paradigm for offloading demanding computational tasks from the UAV through a wireless transmission link. However, the offloaded information may become compromised by eavesdroppers. To address this issue, we conceive an energyefficient computation offloading technique for UAV-MEC systems, with an emphasis on physical-layer security. We formulate a number of energy-efficiency problems for secure UAV-MEC systems, which are then transformed to convex problems. Finally, their optimal solutions are found for both active and passive eavesdroppers. Furthermore, the conditions of zero, partial and full offloading are analyzed from a physical perspective. The numerical results highlight the specific conditions of activating the above three offloading options and quantify the performance of our proposed offloading strategy in various scenarios.

Unmanned aerial vehicles (UAVs) have been widely used in both military and civilian applications. Equipped with diverse communication payloads, UAVs cooperating with satellites and base stations (BSs) constitute a space-air-ground three-tier heterogeneous network, which are beneficial in terms of both providing the seamless coverage as well as of improving the capacity for increasingly prosperous Internet of Things (IoT) networks. However, cross-tier interference may be inevitable among these tightly embraced heterogeneous networks when sharing the same spectrum. The power association problem in satellite, UAV and macrocell three-tier networks becomes a critical issue. In our paper, we propose a two-stage joint hovering altitude and power control solution for the resource allocation problem in UAV networks considering the inevitable cross-tier interference from space-air-ground heterogeneous networks. Furthermore, Lagrange dual decomposition and concaveconvex procedure (CCP) method are used to solve this problem, followed by a low-complexity greedy search algorithm. Finally, simulation results show the effectiveness of our proposed twostage joint optimization algorithm in terms of UAV network's total throughput.

No abstract