Summary Virtual machine (VM) consolidation is to assign a set of VMs requested by operators to the physical machines (PMs) in the data centers so that certain cost, profit or performance objective is optimized, subject to the PMs resource capacity constraints. It is an important mean to decrease the total power consumption by reducing the number of active physical machines (PM) in a cloud. In this context, most of the existing solutions rely on highly frequent live migration to reduce the number of active physical machines. However, live migration is a high resource consumption and time‐consuming operation, thus, frequent use of live migration not only increases energy consumption but also affects the stability of the physical machine, which in turn affects the services on the virtual machine. Reducing the number of active physical machines while reducing the number of live migrations is a major challenge in the face of massive fluctuating virtual machine loads. In order to solve this problem, in this paper, we present a VM consolidation algorithm for predictable loads (VCPL) to reduce the live migration operations. First, we present a cyclic usage prediction (CUP) method to predict the load in a whole cycle (a day) of a VM. Then, we separate the VMs with stable and cyclic load out from others and consolidate them to PMs by using VCPL to make sure each PM has a stable load. Thus, energy can be reduced by avoiding most of live migration operations, and the stability of the data center can be observably improved. We evaluate our solution through simulations on real‐world workloads, the results show that, 66% of long‐term VMs have stable and cyclic loads and are predictable, by using VCPL, the live migration operations occurring on the PMs which accommodate those VMs can be reduced significantly than other solutions.

Most matting research resorts to advanced semantics to achieve high-quality alpha mattes, and a direct low-level features combination is usually explored to complement alpha details. However, we argue that appearance-agnostic integration can only provide biased foreground (FG) details and that alpha mattes require different-level feature aggregation for better pixel-wise opacity perception. In this article, we propose an end-to-end hierarchical and progressive attention matting network (HAttMatting++), which can better predict the opacity of the FG from single RGB images without additional input. Specifically, we utilize channel-wise attention (CA) to distill pyramidal features and employ spatial attention (SA) at different levels to filter appearance cues. This progressive attention mechanism can estimate alpha mattes from adaptive semantics and semantics-indicated boundaries. We also introduce a hybrid loss function fusing structural similarity, mean square error, adversarial loss, and sentry supervision to guide the network to further improve the overall FG structure. In addition, we construct a large-scale and challenging image matting dataset comprised of 59,000 training images and 1,000 test images (a total of 646 distinct FG alpha mattes), which can further improve the robustness of our hierarchical and progressive aggregation model. Extensive experiments demonstrate that the proposed HAttMatting++ can capture sophisticated FG structures and achieve state-of-the-art performance with single RGB images as input.

Acquiring the most representative examples via active learning (AL) can benefit many data-dependent computer vision tasks by minimizing efforts of image-level or pixel-wise annotations. In this paper, we propose a novel Collaborative Panoptic-Regional Active Learning framework (CPRAL) to address the semantic segmentation task. For a small batch of images initially sampled with pixel-wise annotations, we employ panoptic information to initially select unlabeled samples. Considering the class imbalance in the segmentation dataset, we import a Regional Gaussian Attention module (RGA) to achieve semantics-biased selection. The subset is highlighted by vote entropy and then attended by Gaussian kernels to maximize the biased regions. We also propose a Contextual Labels Extension (CLE) to boost regional annotations with contextual attention guidance. With the collaboration of semantics-agnostic panoptic matching and region-biased selection and extension, our CPRAL can strike a balance between labeling efforts and performance and compromise the semantics distribution. We perform extensive experiments on Cityscapes and BDD10K datasets and show that CPRAL outperforms the cutting-edge methods with impressive results and less labeling proportion.

Recently, deep convolutional neural networks (CNNs) have been widely explored in single image super-resolution (SISR) and contribute remarkable progress. However, most of the existing CNNs-based SISR methods do not adequately explore contextual information in the feature extraction stage and pay little attention to the final high-resolution (HR) image reconstruction step, hence hindering the desired SR performance. To address the above two issues, in this paper, we propose a two-stage attentive network (TSAN) for accurate SISR in a coarse-to-fine manner. Specifically, we design a novel multi-context attentive block (MCAB) to make the network focus on more informative contextual features. Moreover, we present an essential refined attention block (RAB) which could explore useful cues in HR space for reconstructing fine-detailed HR image. Extensive evaluations on four benchmark datasets demonstrate the efficacy of our proposed TSAN in terms of quantitative metrics and visual effects. Code is available athttps://github.com/Jee-King/TSAN.

Glass is very common in the real world. Influenced by the uncertainty about the glass region and the varying complex scenes behind the glass, the existence of glass poses severe challenges to many computer vision tasks, making glass segmentation as an important computer vision task. Glass does not have its own visual appearances but only transmit/reflect the appearances of its surroundings, making it fundamentally different from other common objects. To address such a challenging task, existing methods typically explore and combine useful cues from different levels of features in the deep network. As there exists a characteristic gap between level-different features, i.e., deep layer features embed more high-level semantics and are better at locating the target objects while shallow layer features have larger spatial sizes and keep richer and more detailed low-level information, fusing these features naively thus would lead to a sub-optimal solution. In this paper, we approach the effective features fusion towards accurate glass segmentation in two steps. First, we attempt to bridge the characteristic gap between different levels of features by developing a Discriminability Enhancement (DE) module which enables level-specific features to be a more discriminative representation, alleviating the features incompatibility for fusion. Second, we design a Focus-and-Exploration Based Fusion (FEBF) module to richly excavate useful information in the fusion process by highlighting the common and exploring the difference between level-different features. Combining these two steps, we construct a Progressive Glass Segmentation Network (PGSNet) which uses multiple DE and FEBF modules to progressively aggregate features from high-level to low-level, implementing a coarse-to-fine glass segmentation. In addition, we build the first home-scene-oriented glass segmentation dataset for advancing household robot applications and in-depth research on this topic. Extensive experiments demonstrate that our method outperforms 26 cutting-edge models on three challenging datasets under four standard metrics. The code and dataset will be made publicly available.

Multi-agent reinforcement learning (MARL) based methods for adaptive traffic signal control (ATSC) have shown promising potentials to solve the heavy traffic problems. The existing MARL methods adopt centralized or distributed strategies. The former only models the environment as an agent and suffers from the exponential growth of action and state space. The latter extends the independent reinforcement learning methods, such as DQN, to multiple interactions directly or propagates information, such as state and policy, without taking their qualities into account. In this paper, we propose a multi-agent transfer reinforcement learning method to enhance the performance of MARL for ATSC, which is termed as multi-agent transfer soft actor-critic with the multi-view encoder (MT-SAC). The MT-SAC combines centralized and distributed strategies. In MT-SAC, we propose a multi-view state encoder and a transfer learning paradigm with guidance. The encoder processes input states from multiple perspectives and uses an attention mechanism to weigh the neighborhood information. While the paradigm enables the agents to handle different conditions for improving generalization abilities by transfer learning. Experimental studies on different scale road networks show that the MT-SAC outperforms the state-of-the-art algorithms and makes the traffic signal controllers more collaborative and robust.

Deep reinforcement learning has achieved great success in laser-based collision avoidance work because the laser can sense accurate depth information without too much redundant data, which can maintain the robustness of the algorithm when it is migrated from the simulation environment to the real world. However, high-cost laser devices are not only difficult to apply on a large scale but also have poor robustness to irregular objects, e.g., tables, chairs, shelves, etc. In this paper, we propose a vision-based collision avoidance framework to solve the challenging problem. Our method attempts to estimate the depth and incorporate the semantic information from RGB data to obtain a new form of data, pseudo-laser data, which combines the advantages of visual information and laser information. Compared to traditional laser data that only contains the one-dimensional distance information captured at a certain height, our proposed pseudo-laser data encodes the depth information and semantic information within the image, which makes our method more effective for irregular obstacles. Besides, we adaptively add noise to the laser data during the training stage to increase the robustness of our model in the real world, due to the estimated depth information is not accurate. Experimental results show that our framework achieves state-of-the-art performance in several unseen virtual and real-world scenarios.

Nowadays, more and more sensors, devices and applications are connected in Industrial Internet of Things (IIoT), producing massive real-time flows which need to be scheduled for Quality-of-Service provision. To realize application-aware and adaptive flow scheduling, the problem of traffic classification must be addressed at first. When edge computing paradigm is introduced into IIoT, the traffic classification service can be deployed on edge node in the near-end. Recently, deep-learning-based IIoT traffic classification methods show better performance, but the computational cost of deep learning model is too high to be deployed on edge node. Moreover, increasingly unknown flows generated by new devices and emerging industrial APPs lead to frequent training of traffic classifiers. It is difficult to migrate the complex process of classifier training from cloud server to edge nodes with limited resources. To address these issues, we take the benefits of hash mechanism and consensus mechanism in blockchain to design a lightweight IIoT traffic classification service, which is more applicable for edge computing paradigm. First, inspired by the hash mechanism in blockchain and the learning to hash for big data, we propose a new learning-to-hash method named extension hashing. By this method, we can build the set of binary coding tress (BCT set), then generating hash table for more efficient k-nearest neighbor-based classification without complex classifier training. Then, we design a new voting-based consensus algorithm to synchronize the BCT sets and the hash tables across edge nodes, thereby providing the traffic classification service. Finally, we conduct data-driven simulations to evaluate the proposed service. By comparing traffic classification results on public data set, we can see that the proposed service achieves the highest classification accuracy with the minimal time cost and memory usage.

Most existing single image super-resolution (SISR) methods continually increase the depth or width of networks, without adequately exploring contextual features which are essential for reconstruction. Moreover, such existing methods pay little attention to the final high-resolution(HR) image reconstruction step and therefore hinder the desired SR performance. In this paper, we propose a multi-context and enhanced reconstruction network (MCERN) for SISR. Specifically, a novel model named Multi-Context Block (MCB) which extracts more image contextual features with multibranch dilated convolution. Applying multiple MCBs with residual and dense connections, we can effectively extract contextual and hierarchical features for obtaining the coarse super-resolution result. Then an enhanced reconstruction block (ERB) is followed to extract essential spatial features on the high-resolution image to refine the coarse result to a better result. Extensive benchmark evaluations demonstrate the efficacy of our proposed MCERN in terms of metric accuracy and visual effects.

Abstract Image matting is a long‐standing problem in computer graphics and vision, mostly identified as the accurate estimation of the foreground in input images. We argue that the foreground objects can be represented by different‐level information, including the central bodies, large‐grained boundaries, refined details, etc. Based on this observation, in this paper, we propose a multi‐scale information assembly framework (MSIA‐matte) to pull out high‐quality alpha mattes from single RGB images. Technically speaking, given an input image, we extract advanced semantics as our subject content and retain initial CNN features to encode different‐level foreground expression, then combine them by our well‐designed information assembly strategy. Extensive experiments can prove the effectiveness of the proposed MSIA‐matte, and we can achieve state‐of‐the‐art performance compared to most existing matting networks.

Abstract N‐Queens problem derives three variants: obtaining a specific solution, obtaining a set of solutions and obtaining all solutions. The purpose of the variant I is to find a constructive solution, which has been solved. Variant III is aiming to find all solutions and the largest number of queens currently being resolved is 26. Variant II whose purpose is to obtain a set of solutions for larger‐scale problems relies on various intelligent algorithms. In this paper, we use a master‐slave model genetic algorithm that combines the idea of the evolutionary algorithm and simulated annealing algorithm to solve Variant III, and use a parallel fitness function based on compute unified device architecture. Experimental results show that our scheme achieved a maximum 60‐fold speedup over the single‐CPU counterpart. On this basis, a two‐level parallel genetic algorithm based on the island model and master‐slave model is implemented on the GPU cluster by using message passing interface technology. Using two‐node and three‐node GPU cluster, speedup of 1.46 and 2.01 are obtained on average over single‐node, respectively. Compared with the sequential genetic algorithm, the two‐level parallel genetic algorithm makes full use of the parallel computing power of GPU cluster in solving N‐Queen variant II and improves the performance by 99.19 times in the best case.

Abstract Point cloud semantic segmentation, a crucial research area in the 3D computer vision, lies at the core of many vision and robotics applications. Due to the irregular and disordered of the point cloud, however, the application of convolution on point clouds is challenging. In this article, we propose the “coordinate convolution,” which can effectively extract local structural information of the point cloud, to solve the inapplicability of conventional convolution neural network (CNN) structures on the 3D point cloud. The “coordinate convolution” is a projection operation of three planes based on the local coordinate system of each point. Specifically, we project the point cloud on three planes in the local coordinate system with a joint 2D convolution operation to extract its features. Additionally, we leverage a self‐encoding network based on image semantic segmentation U‐Net structure as the overall architecture of the point cloud semantic segmentation algorithm. The results demonstrate that the proposed method exhibited excellent performances for point cloud data sets corresponding to various scenes.

GPU L1 data cache contention, caused by a huge amount of concurrent threads, leads to insufficient cache utilization and poor performance, especially for cache unfriendly applications. Cache bypassing is a widely- used method to alleviate this problem, and Decoupled L1D (D-L1D) is a preventive bypassing scheme, which achieves performance improvement for cache unfriendly applications by considering the data locality of memory access streams. However, our experiments and analyses show that limited performance gain by D-L1D is attained due to the pre-defined locality threshold. To address this issue, we propose a novel bypassing scheme named as Dynamic D-L1D (DD-L1D) that directs the L1 data cache to the less contention by dynamically updating the locality threshold during runtime. We evaluate four metrics in DD-L1D to indicate the L1 cache bypassing state, and choose bypassing miss rate in our final configuration. The experimental results demonstrate that DD-L1D improves the baseline performance by 1.45X on average for cache unfriendly benchmarks. It also outperforms D-L1D and the state-of-the-art GPU cache bypassing schemes with lower hardware overhead and memory traffic.

Abstract We design dynamic k‐d (DKD) tree based on classical k‐d tree for animated scene rendering. Our method can inherit the benefit of efficient traversal of k‐d tree and minimize time cost to update DKD tree, making it well suited for animated geometry. DKD employs primitive reset, redistribution to reflect the updated positions of geometry, and leaf node incremental growing to avoid the deterioration of hierarchy quality due to refitting. Our experiments show that DKD has a significant rendering performance improvement than selected existing methods. Copyright © 2016 John Wiley & Sons, Ltd.

General-Purpose Graphic Processing Units (GPGPU) have been widely used in high performance computing as application accelerators due to their massive parallelism and high throughput. A GPGPU generally contains two layers of schedulers, a cooperative-thread-array (CTA) scheduler and a warp scheduler, which administer the thread level parallelism (TLP). Previous research shows the maximized TLP does not always deliver the optimal performance. Unfortunately, existing warp scheduling schemes do not optimize TLP at runtime, which is impossible to fit various access patterns for diverse applications. Dynamic TLP optimization in the warp scheduler remains a challenge to exploit the GPGPU highly-parallel compute power.

Concurrent applications in virtualized environments (VE) encounter synchronization problems such as Lock Holder Preemption (LHP). Hybrid co-scheduling is an effective approach to address such problems. However, the contention and exclusiveness between multiple concurrent domains in hybrid co-scheduling cause a serious performance degradation and unfairness. To keep the benefits brought by hybrid co-scheduling for multiple concurrent domains in VE, we propose two optimization schemes named Partial Co-Scheduling (PCS) and Boost Co-Scheduling (BCS) using finer space granularity. Instead of raising co-scheduling signals for all online CPUs, PCS scheme raises the co-scheduling signals only for the indispensable CPUs, while the rest CPUs are untouched. BCS scheme boosts the priorities for co-scheduled virtual CPUs (VCPUs) to induce the scheduler to pick the appropriate VCPUs. We implement both PCS and BCS into Credit Scheduler in Xen 4.0.1 and evaluate their performance compared with original hybrid co-scheduling and co-descheduling under different scenarios. The experimental results show that our proposals effectively alleviate the CPU run-time contention and achieve better performance and fairness compared to traditional hybrid co-scheduling.

Through the development and research in management information system for a cement manufactory, we built a general management information system architecture (GMISA) model. This approach discusses view selection, function assignment consideration, reaction to changes in execution environment or user requirements by switching algorithms at runtime, and some significant methods and properties in the system implementation. The model is divided into several modules that are reusable and adaptive for the end-user or developer.