Search

Deep Neural Network (DNN)-based point cloud semantic segmentation has presented significant breakthrough using large-scale labeled aerial laser point cloud datasets. However, annotating such large-scaled point clouds is time-consuming. Self-Supervised Learning (SSL) is a promising approach to this problem by pre-training a DNN model utilizing unlabeled samples followed by a fine-tuned downstream task involving very limited labels. The traditional contrastive learning for point clouds selects the hardest negative samples by solely relying on the distance between the embedded features derived from the learning process, potentially evolving some negative samples from the same classes to reduce the contrastive learning effectiveness. This work proposes a hard-negative sample-aware self-supervised contrastive learning algorithm to pre-train the model for semantic segmentation. We designed a k-means clustering-based Absolute Positive And Negative samples (AbsPAN) strategy to filter the possible false-negative samples. Experiments on two typical ALS benchmark datasets demonstrate that the proposed method is more appealing than supervised training schemes without pre-training. Especially when the labels are severely inadequate (10% of the ISPRS training set), the results obtained by the proposed HAVANA method still exceed 94% of the supervised paradigm performance with full training set. [ABSTRACT FROM AUTHOR]

The study on the Zhangjialiang No. 2 coal of Middle Jurassic Yan'an Formation in the northeastern Ordos Basin has been analyzed by gas chromatography-mass spectrometry (GC-MS) to determine geochemical characteristics and reconstruct palaeovegetation and palaeoclimate changes. The distribution of n -alkanes is mainly unimodal, with a medium and high molecular weight. Pr/Ph ratio (2.03–10.6) is high in most samples. The values of Tm/Ts ratio are between 1.42 and 4.92, and a predominance of C29 over C27 steranes has been found for the analyzed samples. These parameters indicate that all samples were deposited in an oxidizing freshwater environment, and the organic matter was mainly derived from terrestrial higher plants, with a small amount of aquatic organisms. Thermal maturity parameters such as C3122S/(22S + 22R) homohopane, C30αβ/(αβ + βα) hopane, together with the total organic carbon (TOC) contents (43.1% - 58.6% (except floor 2-13)) and the detected of C2717β(H)-22, 29, 30-trisnorhopane (βTm), suggest that most of samples have a relatively low maturity in the early stage. Besides, gymnosperm-derived biomarkers, including labdane, norpimarane, isopimarane, phyllocladane, abietane, rimuane, ent-beyerane, ent-kaurane, suggest that the vegetation during the Middle Jurassic Yan'an Formation was characterized by conifers. [ABSTRACT FROM AUTHOR]

To mismatch various power disturbances of interconnected power grid, an integrated generation‐control and power‐dispatch (IGCPD) for performance‐based frequency regulation market is proposed. Unlike the conventional automatic generation control with an independent controller and a power distributor, IGCPD can significantly improve the dynamic control performance via combining these two parts into an integrated decision agent. The IGCPD eliminates the waste of frequency regulation resources and repeated regulation of the units caused by the mismatch of the controller and power distributor. In addition, by setting the reward function reasonably, IGCPD can save regulatable capacity of fast units for regulation in the next large‐scale disturbance. In order to enhance the decision ability, a swarm intelligence‐based deep deterministic policy gradient (SI‐DDPG) algorithm is proposed to acquire the control knowledge and implement a high‐quality decision for this agent. Moreover, the swarm intelligence‐based training sample can achieve a high learning robustness for SI‐DDPG which solve the problem of low robustness in conventional deep reinforcement learning. The proposed technique is verified on an England two‐area load frequency control model and the west China two‐area LFC model. IGCPD improves the area control error control performance and reduce the frequency regulation mileage payment of the interconnected power grid. [ABSTRACT FROM AUTHOR]

Storage virtualization is an important component of large-scale online services in multi-tenant clouds. It typically shares the physical storage among guest machines and performs transactional operations for high-performance data processing. However, even with the recent mediated pass-through virtualization optimization, the operations of multi-tenant storage I/O meet the bottleneck, and thus degrade the throughput performance of the cloud storage services. We observe that the root cause of the problem is the unawareness of varying and imbalanced workload inefficiency of resource management in the multi-tenant cloud storage setting. In this paper, we present FinNVMe, a new throughput-oriented NVMe storage virtualization management mechanism, that (1) passes-through I/O performance-critical resources and emulates privileged resources to provide high throughput in a workload-aware manner among multi-tenant VMs, (2) enables fine-grained scheduling for I/O resources to achieve promising flexibility and scalability with respective to virtualization, and (3) adopts the queue binding and the queue shuffling to reduce the virtualization and management overhead, and involves active polling for further I/O acceleration. This article subsequently evaluates FinNVMe with micro benchmarks on two typical scenarios (both balanced and imbalanced workload) and the real-world storage workloads to show its high throughput performance, along with the flexibility and scalability of virtualization and resource management. For example, FinNVMe achieves up to 20 percent throughput improvement with more stable latency in the varying and imbalanced workload. [ABSTRACT FROM AUTHOR]