Using mobile Light Detection and Ranging point clouds to accomplish road scene labeling tasks shows promise for a variety of applications. Most existing methods for semantic labeling of point clouds require a huge number of fully supervised point cloud scenes, where each point needs to be manually annotated with a specific category. Manually annotating each point in point cloud scenes is labor intensive and hinders practical usage of those methods. To alleviate such a huge burden of manual annotation, in this paper, we introduce an active learning method that avoids annotating the whole point cloud scenes by iteratively annotating a small portion of unlabeled supervoxels and creating a minimal manually annotated training set. In order to avoid the biased sampling existing in traditional active learning methods, a neighbor-consistency prior is exploited to select the potentially misclassified samples into the training set to improve the accuracy of the statistical model. Furthermore, lots of methods only consider short-range contextual information to conduct semantic labeling tasks, but ignore the long-range contexts among local variables. In this paper, we use a higher order Markov random field model to take into account more contexts for refining the labeling results, despite of lacking fully supervised scenes. Evaluations on three data sets show that our proposed framework achieves a high accuracy in labeling point clouds although only a small portion of labels is provided. Moreover, comparative experiments demonstrate that our proposed framework is superior to traditional sampling methods and exhibits comparable performance to those fully supervised models.

Traffic Engineering (TE) is an efficient technique to balance network flows and thus improves the performance of a hybrid Software Defined Network (SDN). Previous TE solutions mainly leverage heuristic algorithms to centrally optimize link weight setting or traffic splitting ratios under the static traffic demand. Note that as the network scale becomes larger and network management gains more complexity, it is notably that the centralized TE methods suffer from a high computation overhead and a long reaction time to optimize routing of flows when the network traffic demand dynamically fluctuates or network failures happen. To enable adaptive and efficient routing in distributed TE, we propose a Multi-agent Reinforcement Learning method CMRL that divides the routing optimization of a large network into multiple small-scale routing decision-making problems. To coordinate the multiple agents for achieving a global optimization goal in a hybrid SDN scenario, we construct a reasonable virtual environment to meet different routing constraints brought by legacy routers and SDN switches for training the routing agents. To train the routing agents for determining the local routing policies according to local network observations, we introduce the difference reward assignment mechanism for encouraging agents to cooperatively take optimal routing action. Extensive simulations conducted on the real traffic traces demonstrate the superiority of CMRL in improving TE performance, especially when traffic demands change or network failures happen.

Direct integration methods are widely used for dynamic response computation. However, the performance of their computational accuracy significantly degrades with increasing the time step. Although machine learning methods can address this shortcoming, they require training data for dynamic response computation. This paper proposes a novel computational method to overcome these shortcomings. The proposed approach is a data-free physics-driven estimator, which minimizes the objective function of multi-output least squares support vector machines for regression to model parameters subject to physical constraints introduced by the multi-degree of freedom system's dynamic equilibrium equations and initial conditions in the feature space, bypassing the need for training data (due to the coupled physics) and for satisfying the requirement of the time step due to the built-in optimization procedure. A new efficient step-by-step solver is developed to solve the optimization problem, and the solution is equivalent to a hyperplane satisfying the physical constraints in the feature space. The extension of the proposed approach for nonlinear dynamic response computation is also analyzed theoretically. The numerical results demonstrate that the proposed approach provides the solution with higher accuracy and efficiency and achieves the best performance for large time steps over classical integration methods.

Feature matching is a fundamental problem in feature-based remote sensing image registration. Due to the ground relief variations and imaging viewpoint changes, remote sensing images often involve local distortions, leading to difficulties in high-accuracy image registration. To address this issue, in this article, we propose a robust feature matching method called First Neighbor Relation Guided (FNRG) for remote sensing image registration via guided hyperplane fitting. The key idea of FNRG is to exploit the first neighbor relation of feature points between two images for seeking consistent seeds in a parameter-free manner. To boost more consistent matches based on the consistent seeds, we formulate the feature matching problem into an affine hyperplane fitting problem by imposing the motion consistency, and then we design a hyperplane updating strategy to refine the fitting model. We also introduce a locality preserving structure-based cost function to promote the matching performance of the hyperplane updating strategy. Our method can mine consistent matches from thousands of putative ones within only a few milliseconds, and it also can handle the data with a large-scale change, rotation, or severe nonrigid deformation. Extensive experiments on the remote sensing image data sets with different types of image transformations show that the proposed method achieves significant superiority over several state-of-the-art methods.

To reduce the investment on network infrastructure, many online service providers have begun to adopt the shared inter-DataCenter Wide Area Network (inter-DC WAN) that connects different datacenters and Internet Service Providers (ISPs). The shared inter-DC WAN accommodates two kinds of traffic, i.e. delay-sensitive ISP-facing traffic and high-throughput inter-DC traffic. Traffic Engineering (TE) in the shared inter-DC WAN should determine the routing paths for all traffic to achieve link load balancing, while select lower-latency egress routers for ISP-facing traffic to guarantee the Quality of Service (QoS). Therefore, this paper mainly focuses on jointly optimizing routing paths selection and egress router selection to strike a balance between QoS and link load balancing. Specifically, we first formulate the TE problem in the shared inter-DC WAN as a mixed integer nonlinear programming problem. Then, a TED method is proposed to jointly optimize the egress router selection and routing path selection by learning an intelligent agent with Deep Reinforcement Learning (DRL). The learnt agent can self-adaptively and rapidly select the optimal egress routers by considering the utilization-latency balance when traffic demand changes. Finally, we conduct extensive evaluations on Alibaba WAN with real traffic traces to demonstrate the effectiveness and superiority of the proposed method.