The paper targets the challenging task of Ship License Plate (SLP) recognition. Existing methods for SLP recognition are hampered by the scarcity of large and publicly available datasets, leading to evaluations on small and non-representative datasets. To alleviate it, we have built a large dataset, called SLP34K, which consists of 34,385 images collected by an intelligent traffic surveillance system. The dataset is carefully manually annotated with text labels and attributes, and presents high data diversity by multiple installation locations and long capturing period of the cameras. Additionally, we propose a simple yet effective SLP recognition baseline method. The baseline is equipped with a strong visual encoder that benefits from initial pre-training via self-supervised learning, followed by further refinement through our devised semantic enhancement module. Extensive experiments on SLP34K verify the effectiveness of our proposed baseline. Moreover, while our baseline is designed for SLP recognition, it can also be used for common scene text recognition and achieve state-of-the-art performance on seven mainstream scene text recognition datasets.

Ship Re-Identification (ReID) aims to accurately identify ships with the same identity across different times and camera views, playing a crucial role in intelligent waterway transportation. However, compared to the widely researched pedestrian and vehicle ReID, Ship ReID has received much less attention, primarily due to the scarcity of large-scale and high-quality ship ReID datasets available for public access. Moreover, several unique challenges make ship ReID particularly difficult: ships are large objects that are hard to capture fully, and the visible area of ships vary significantly due to changes in cargo loading or water surface conditions. These challenges make it difficult to achieve ideal results by directly applying existing ReID methods. To address these challenges, in this paper, we introduce ShipReID-2400, a dataset for ship ReID compiled from a real-world intelligent waterway traffic monitoring system. It comprises 17,241 images of 2,400 distinct ship identities collected over 53 months, ensuring diversity and representativeness. Furthermore, we propose the Disentangle-to-Interact Network ( D2InterNet ), a simple but strong baseline for ship ReID designed to extract discriminative local features despite significant scale variations. Extensive experimental results show that D2InterNet achieves state-of-the-art performance on both the ShipReID-2400 and VesselReID datasets. In addition, despite being designed for ship ReID, D2InterNet also achieves competitive results on the MSMT17 pedestrian ReID dataset, showcasing its good generalization capability. Our dataset and code are publicly available at https://github.com/HuiGuanLab/ShipReID-2400.

Federated learning (FL) in resource-constrained wireless networks faces the challenge of long training delays. In this work, we explore delay-efficient FL by leveraging device-to-device (D2D) communications to accelerate the uploading of local models. We formulate a joint problem of upload mode selection and bandwidth allocation, which is a mixed-integer nonlinear programming (MINLP) problem and difficult to solve directly. To address this, we propose a low-complexity two-step algorithm: the first step determines the upload modes for edge devices, while the second step optimally allocates the bandwidth. Simulation results show that our algorithm outperforms baseline schemes, with delay reduction becoming more pronounced as the number of edge devices increases.

The computing power network (CPN) offers exceptional computational capabilities and reliable network services, with significant potential for future applications. To achieve ubiquitous coverage and efficient computational resource allocation, CPN can be seamlessly coordinated with low-cost autonomous aerial vehicle (AAV)-based mobile computing platforms. This article investigates an efficient low-altitude AAV-assisted computing power and resource allocation mechanism tailored for urban environments. The aim is to ensure seamless scheduling and efficient processing of computational tasks across various computing devices at different layers of the CPN, while minimizing AAV energy consumption and ensuring flight safety. First, this article proposes an Urban AAV-assisted CPN task-allocation and AAV trajectory-management decision-making problem. The AAV works until it safely lands, aiming to minimize overall task processing delay and AAV energy consumption while ensuring fairness in task allocation. Then, a novel AAV-protection-based multiagent deep deterministic policy gradient (UP-MADDPG) algorithm is introduced. It offers dynamic management of secure computing and communication flight paths when facing building blockages. Finally, we compared the proposed algorithm with baseline algorithms across various metrics. Experimental results demonstrate that the proposed algorithm achieves lower and more balanced task execution delay and AAV energy consumption while also improving fairness.

In the contemporary landscape of computationally intensive applications, Computing Power Network (CPN) offers a solution to enhance computational efficiency and cost-effectiveness by integrating and sharing computing resources. However, with the surge in task volume within multi-user environments, effectively scheduling these tasks to optimize system profit and delay presents a significant challenge. This paper introduces an optimization approach leveraging Deep Deterministic Policy Gradient (DDPG) to enhance CPN performance through task decomposition and computing path optimization. We initially construct a multi-layer CPN system model encompassing cloud computing, edge computing, and terminal device layers. Subsequently, we integrate a novel mechanism for convex optimization-based task decomposition, enabling intelligent subdivision of tasks into sub-tasks and dynamic allocation to suitable nodes within the network. Furthermore, we devise a Convex Optimization Task Decomposition-based Multi-Agent Deep Deterministic Policy Gradient (CO-MADDPG) algorithm, empowering multiple computing tasks as independent agents to learn and identify optimal offloading paths and computing nodes, thereby minimizing delay and maximizing system profit. A series of simulation experiments validate the effectiveness of the CO-MADDPG algorithm in handling concurrent tasks, demonstrating its capability to reduce task completion times, enhance system revenue, and maintain adaptability and stability across varying task demands.

We consider the problem of reducing the learning latency of layerwise federated learning through joint device selection and bandwidth allocation. Specifically, we examine practical scenarios with heterogeneous devices with varying system parameters (e.g., CPU frequency, transmit power, etc.) and energy budgets. We formulate a long-term optimization problem, which is difficult to solve even with perfect channel state information. To address the issue, we employ Lyapunov theory to transform the problem into a series of online optimization problems, each of which can be efficiently solved using an alternating optimization-based method. Simulation results show that our scheduling scheme surpasses baseline schemes not only in terms of reducing the learning latency but also in reducing the energy deficit.

We address the minimum-delay beam scheduling problem leveraging reflections for switched beamforming systems. The objective is to efficiently disseminate a data file from a transmitter to a set of nodes via multiple predetermined beams with arbitrary overlapping patterns. The problem is formulated as a challenging mixed integer nonlinear programming (MINLP) and then decomposed into a set of subproblems. The subproblems are still difficult to solve due to their NP-hardness. We propose a heuristic algorithm for the subproblems, based on which two heuristic algorithms with different computational complexities are developed for the original problem. Simulation results high-light the significant reduction in dissemination delay achieved by the proposed algorithms compared to baseline approaches without leveraging reflections.

We jointly consider full-duplex operation and network coding in two-hop relay networks to enhance the throughput of the block transmission of packets over erasure channels. Two coded transmission schemes, termed Fewest Broadcast Packet First (FBPF) and Buffer Contents-based Coded Transmission (BCCT), are proposed, where random linear network coding is employed at the Base Station (BS) and the Relay Station (RS), respectively. Both schemes do not rely on users’ Channel State Information (CSI), buffer status, channel parameters, etc., and hence are practically viable. We derive closed-form upper bounds on the throughput of both schemes. We prove that both schemes achieve the optimal throughput when the BS-to-RS channel is perfect. Through extensive simulations, we demonstrate that both schemes incur substantially higher throughput than the traditional uncoded Automatic Repeat-reQuest (ARQ) scheme and perform close to a general upper bound on the system throughput. Furthermore, even with imperfect Self-Interference Cancellation (SIC) at the full-duplex RS, our schemes are shown to be superior to state-of-the-art coded transmission schemes designed for half-duplex relay networks, given that the impact of imperfect SIC on the BS-to-RS channel quality is not high.

We study efficient federated learning (FL) using random pruning in resource-constrained edge intelligence networks. We propose an edge device selection strategy to identify appropriate edge devices for participating in FL at the beginning of each training iteration. We then formulate an optimization problem that jointly optimizes the pruning ratio, CPU frequency, uplink power, and bandwidth allocation for the selected edge devices. Since the optimization problem is non-convex and challenging to solve directly, we decompose it into three subproblems and propose efficient algorithms or closed-form solutions for each subproblem. Based on the solutions to the subproblems, an alternating optimization algorithm is constructed to solve the original problem. Simulation results demonstrate that our scheme outperforms baseline schemes in terms of both learning accuracy and energy consumption.

Graph neural networks (GNNs) have been widely applied in software-defined network (SDN) for better network modeling and performance prediction. However, the black-box characteristic of deep learning makes the GNNs hard to interpret, such interpretability issue hinders the wide use of GNNs. In this paper, we propose Shapley Explainer, that provides fair importance scores to the input nodes of a GNN within an appropriate computation cost, thereby providing a valid and reasonable interpretation of graph neural network on software defined network. The proposed method derives the importance ranking of topological nodes by combining shapley values with a soft discrete mask matrix. We apply Shapley Explainer to RouteNet model, a GNN model that provides intelligent predictions of SDN network performance metrics. The experimental results show that Shapley Explainer can provide effective interpretations for RouteNet. It also verifies that the RouteNet model can correctly learn the relationship between features, which can provide a better understanding of the prediction process of RouteNet, promoting the application of GNN-based SDN systems in engineering practice.

We consider the minimum-delay multicast scheduling problem for switched beamforming systems. A salient characteristic of mmWave links, reflection, is considered, which enables opportunistic reduction of data dissemination delay. We formulate the problem as a mixed integer nonlinear programming, which is difficult to solve directly. Instead, we decompose the problem into a set of subproblems, by allocating a fixed path to each receiver for data reception. The optimal solution to each subproblem has a contiguous structure, and hence can be computed using a dynamic programming-based approach. We propose an optimal algorithm for the original problem based on the solutions to the subproblems. By simulation we show the outperformance of our algorithm over an optimal multicast scheduling policy without leveraging reflections and a broadcast baseline scheme.

We consider the channel-aware scheduling (CAS) problem for block transmission of packets in two-hop full-duplex relay networks with multiple users. At each time slot, the full-duplex relay station (RS) can fetch a network-coded packet from the macro base station (BS), and schedule a previously received packet for broadcasting to the users over time-varying channels. Our goal is to maximize the broadcast throughput. Since the associated Markov decision programming problem turns out to be intractable as the size of the problem increases, we propose a CAS scheme which is simple to implement and also achieves near-optimal performance. We provide a closed-form expression of the throughput of our scheme when the BS-to-RS channel is perfect, and prove that our scheme is optimal for one-user systems. Finally, numerical results demonstrate that our scheme performs close to an upper bound of the system and outperforms other transmission schemes.

We consider the full-duplex operation with network coding in two-hop relay networks to enhance the throughput for block transmission of packets. We propose a low-complexity transmission scheme, which does not rely on channel state information (CSI), and hence can be easily implemented in practical systems. We derive a closed-form upper bound on the asymptotic throughput of the proposed scheme, and show that the derived upper bound is tighter than a general upper bound on the throughput of any transmission scheme even with perfect CSI. Simulation results show that, the proposed scheme actually performs close to the general upper bound, and in most cases it substantially outperforms the traditional uncoded Automatic Repeat-reQuest scheme which relies heavily on the ACK/NAK feedback for packet retransmission.

Distributed Denial of Service (DDoS) attacks grow rapidly and become one of the fatal threats to the Internet. Automatically detecting DDoS attack packets is one of the main defense mechanisms. Conventional solutions monitor network traffic and identify attack activities from legitimate network traffic based on statistical divergence. Machine learning is another method to improve identifying performance based on statistical features. However, conventional machine learning techniques are limited by the shallow representation models. In this paper, we propose a deep learning based DDoS attack detection approach (DeepDefense). Deep learning approach can automatically extract high-level features from low-level ones and gain powerful representation and inference. We design a recurrent deep neural network to learn patterns from sequences of network traffic and trace network attack activities. The experimental results demonstrate a better performance of our model compared with conventional machine learning models. We reduce the error rate from 7.517% to 2.103% compared with conventional machine learning method in the larger data set.

Extensible software architecture is needed to adapt to the development of router. As an important part of ForCES router software, an extensible LFB management and development model based on ForCES router software was illustrated, which enhances extensibility of the software. The ForCES router software structure was first discussed. Based on that, the architecture of this model was described. In order to testify the feasibility, an implementation of the model was presented and an ipv6 forwarding service based it was brought out. The impact on traffic flows and applications was evaluated.