Nowadays, remote sensing object detection has benefited a lot from the development of convolutional neural networks (CNNs). However, it is still a challenging task due to arbitrary orientation and dense distribution of objects in remote sensing images. To deal with these difficulties, we propose two effective attention mechanisms with parallel groups strategy to enhance feature representations in the detection head, named PGAE-head. Significantly, our designs can achieve competitive performance improvement by only introducing tiny parameters and computations in the model. Firstly, the features received by the PGAE-head are divided into multiple groups, which ensures the independence of each group during subsequent attention enhancement. Then, PGAE-head processes these sub-features with enhanced attention mechanisms based on spatial and channel dimensions in parallel to detect more accurate results. Experiments on DOTA and HRSC datasets show that the proposed PGAE-head achieves comparable performances with other state-of-the-art CNN-based models at minimal optimization costs, demonstrating its effectiveness.

Synthetic Aperture Radar (SAR) ship detection algorithms have achieved extensive development in recent years. In spite of this, the insufficient data and the non-intuitive feature of SAR images still brought certain challenges. This paper proposes a multi-modality feature enhancement (MMFE) method based on feature disentanglement for SAR image target detection. By precisely exploring modality-shared features of optical and SAR images, MMFE can optimize the SAR feature representation capability. First, we propose a feature disentanglement (FD) module to acquire transferable modality-shared knowledge, thereby effectively alleviating the modality shift phenomenon in the subsequent modality alignment. Second, we introduce a multi-granularity modality alignment (MGMA) module that further eliminates inter-modality differences, ultimately achieving effective compensation for the SAR modality. Extensive experimental results convincingly demonstrate the compelling ability of MMFE.

Synthetic aperture radar (SAR) ship detection methods have achieved remarkable progress in recent years. However, unlike RGB images, the characteristics of SAR imaging will result in non-intuitive feature representations. Furthermore, due to the insufficient data of SAR images, existing methods relying on plenty of labeled SAR images may be hard to achieve promising performance. To address the aforementioned issues, a cross-modality feature transfer (CMFT) method is proposed in this article, which enhances feature representations in the SAR modality by transferring rich knowledge in the RGB modality. First, we propose a multilevel modality alignment network (MMAN), which encourages the model to effectively learn modality-invariant features and alleviate the large cross-modality discrepancies by aligning features from multilevels (scene level, local level, global level, and instance level). Second, to address the underperformance of samples with non-intuitive features in the modality alignment, we introduce a hard-sample supervision module (HSM) in the stage of feature extraction, which can thoroughly exploit the feature of hard-to-align samples by giving more optimization energy for them. Third, to enhance the discriminability of instance-level features, a feature complementary module (FCM) is customized to fully explore the potential complementary clues between instance-level features and context information for the instance-level feature alignment. Extensive experimental results demonstrate that the CMFT outperforms the state-of-the-art detectors. Compared to the baseline model, CMFT improves the accuracy by 3.1% mean average precision (mAP) on the SSDD dataset and 3.4% mAP on the HRSID dataset, demonstrating its superior SAR ship detection performance.

Running multiple virtual networks, customized for different performance objectives, is a promising way to support diverse applications over a shared substrate. Despite being simple, a static division of resources between virtual networks can be highly inefficient, while dynamic resource allocation runs the risk of instability. This paper uses optimization theory to show that adaptive resource allocation can be stable and can maximize the aggregate performance across the virtual networks. In the DaVinci architecture, each substrate link periodically reassigns bandwidth shares between its virtual links; while at a smaller timescale, each virtual network runs a distributed protocol that maximizes its own performance objective independently. Numerical experiments with a mix of delay-sensitive and throughput-sensitive traffic show that the bandwidth shares converge quickly to the optimal values. We demonstrate that running several custom protocols in parallel and allocating resource adaptively can be more efficient, more flexible, and easier to manage than a compromise one-size-fits-all design.

In the Internet today, traffic management spans congestion control (at end hosts), routing protocols (on routers), and traffic engineering (by network operators). Historically, this division of functionality evolved organically. In this paper, we perform a top-down redesign of traffic management using recent innovations in optimization theory. First, we propose an objective function that captures the goals of end users and network operators. Using all known optimization decomposition techniques, we generate four distributed algorithms that divide traffic over multiple paths based on feedback from the network links. Combining the best features of the algorithms, we construct TRUMP: a traffic management protocol that is distributed, adaptive, robust, flexible and easy to manage. Further, TRUMP can operate based on implicit feedback about packet loss and delay. We show that using optimization decompositions as a foundation, simulations as a building block, and human intuition as a guide can be a principled approach to protocol design.

No abstract available.

In the Internet today, traffic engineering is performed assuming that the offered traffic is inelastic. In reality, end hosts adapt their sending rates to network congestion, and network operators adapt the routing to the measured traffic. This raises the question of whether the joint system of congestion control (transport layer) and routing (network layer) is stable and optimal. Using the established optimization models for TCP and traffic engineering as a basis, we find the joint system can be stablized and often maximizes aggregate user utility. We prove that both stability and optimality of the joint system can be guaranteed for sufficiently elastic traffic simply by tuning the cost function used for traffic engineering. Then, we present a new algorithm that adapts on a smaller timescale to changes in traffic distribution and is more robust to large traffic bursts. Uniting the network and transport layers in a multi-layer approach, this algorithm, Distributed Adaptive Traffic Engineering (DATE), jointly optimizes the goals of end users and network operators and reacts quickly to avoid bottlenecks. Simulations demonstrate that DATE converges quickly.

In the Internet today, traffic engineering is performed assuming that the offered traffic is inelastic. In reality, end hosts adapt their sending rates to network congestion, and network operators adapt the routing to the measured traffic. This raises the question of whether the joint system of congestion control and routing is stable and optimal. Using established optimization models for TCP and traffic engineering as a basis, we find the joint system is stable and typically maximizes aggregate user utility through simulation. The joint system may deviate from this solution when the topology is not uniform. A modification to the joint system will guarantee stability and optimality for applications that are sufficiently elastic, but at the cost of robustness.

Despite the large body of work studying congestion control and adaptive routing in isolation, much less attention has been paid to whether these two resource-allocation mechanisms work well together to optimize user performance. Most analysis of congestion control assumes static routing, and most studies of adaptive routing assume that the offered traffic is fixed. In this paper, we analyze the interaction between congestion control and adaptive routing, and study the stability and optimality of the joint system. Previous work has shown that the system can be modelled as a joint optimization problem that naturally leads to a primal-dual algorithm with shortest-path routing using congestion prices as the link weights. In practice, the algorithm is commonly unstable. We consider three alternative timescale separations and examine the stability and optimality of each system. Our analytic characterizations and simulation experiments demonstrate how the step size of the congestion-control algorithm affects the stability of the system, and how the timescale of each control loop and homogeneity of link capacities affect system stability and optimality. The stringent conditions imposed for stability suggests that congestion price would be a poor feedback mechanism in practice.