深度检测DDoS攻击

为了有效地防御DDoS攻击,需要在检测环节尽可能多地获取攻击信息,而现有的方法大多仅注重检测攻击的存在,很难同时给出攻击协议、攻击强度和攻击方式等信息.提出使用多分类的方式,将攻击分为24个不同的种类,并用快速分类器HSMC-SVM作为分类工具,来完成DDoS攻击的多种信息的获取.实验表明,这种方法可以快速完成训练和测试工作,并以较高的识别率识别出不同种类的攻击,为防御环节提供攻击协议、攻击强度和攻击方式等信息.在实际网络中,能满足准确性和实时性的要求,有较强的可行性.     

遥感图像非监督计算机分类方法的研究

通过对几种常用的非监督计算机遥感图像分类方法,如k-means、层次聚类和神经网络的分析研究发现,由于这些方法不能克服数据噪声点的影响,输出结果对输入参数依赖性较大,使其对图像的分类效果受到影响。为了提高图像的非监督分类效果。本文提出了一种基于密度和自适应密度可达聚类算法。实验分析表明,与常用的分类方法相比,该算法具有良好的分类效果。     

Cell Development Enhanced Bionic Silk Hydrogel on Remodeling Immune Pathogenesis of Spinal Cord Injury via M2 Polarization of Microglial

Due to the complex spatial-temporal pathophysiology of spinal cord injury (SCI), effective modulation of SCI-specific inflammatory pathogenesis to achieve desirable therapeutic effects on functional recovery still remains challenging. Herein, cell-enhanced photocrosslinked silk fibroin hydrogels with extracellular matrix-mimicking cues of mechanical properties and RGD (Arg-Gly-Asp) signals are gelled in situ to fill the lesion site to modulate injury-induced neuroinflammation and promote neurite regrowth after SCI. The bionic hydrogel system provides biomimetic mechanical cues to promote neuronal differentiation of neural stem/progenitor cells (NPCs) and neurite growth by activating YAP nuclear expression. Importantly, favored by the strong capacity of silk fibroin hydrogels on macrophage/microglia recruitment, NPCs encapsulated hydrogel (NPCs@SFRGD0.1) effectively promotes recruited macrophages/microglia to M2 polarization in the lesion site by releasing S100A4 and thereby remodels the inflammatory microenvironment after SCI. Moreover, NPCs@SFRGD0.1 successfully reduces glial scar formation and accelerates corticospinal tract axon regrowth to improve locomotor recovery. Overall, this work contributes to illustrating the therapeutic mechanism of NPCs development based biomaterial therapies on modulating inflammatory microenvironment and this NPCs enhanced silk fibroin hydrogel provides a promising therapeutic strategy for SCI.     

Plasma Enhanced Lithium Coupled with Cobalt Fibers Arrays for Advanced Energy Storage

Construction of high efficiency and stable Li metal anodes is extremely vital to the breakthrough of Li metal batteries. In this study, for the first time, groundbreaking in situ plasma interphase engineering is reported to construct high-quality lithium halides-dominated solid electrolyte interphase layer on Li metal to stabilize & protect the anode. Typically, SF₆ plasma-induced sulfured and fluorinated interphase (SFI) is composed of LiF and Li₂S, interwoven with each other to form a consecutive solid electrolyte interphase. Simultaneously, brand-new vertical Co fibers (diameter: ≈5 µm) scaffold is designed via a facile magnetic-field-assisted hydrothermal method to collaborate with plasma-enhanced Li metal anodes (SFI@Li/Co). The Co fibers scaffold accommodates active Li with mechanical integrity and decreases local current density with good lithiophilicity and low geometric tortuosity, supported by DFT calculations and COMSOL Multiphysics simulation. Consequently, the assembled symmetric cells with SFI@Li/Co anodes exhibit superior stability over 525 h with a small voltage hysteresis (125 mV at 5 mA cm⁻²) and improved Coulombic efficiency (99.7%), much better than the counterparts. Enhanced electrochemical performance is also demonstrated in full cells with commercial cathodes and SFI@Li/Co anode. The research offers a new route to develop advanced alkali metal anodes for energy storage.     

GaP/GaNP Heterojunctions for Efficient Solar‐Driven Water Oxidation

The growth and characterization of an n‐GaP/i‐GaNP/p⁺‐GaP thin film heterojunction synthesized using a gas‐source molecular beam epitaxy (MBE) method, and its application for efficient solar‐driven water oxidation is reported. The TiO₂/Ni passivated n‐GaP/i‐GaNP/p⁺‐GaP thin film heterojunction provides much higher photoanodic performance in 1 m KOH solution than the TiO₂/Ni‐coated n‐GaP substrate, leading to much lower onset potential and much higher photocurrent. There is a significant photoanodic potential shift of 764 mV at a photocurrent of 0.34 mA cm^?2, leading to an onset potential of ≈0.4 V versus reversible hydrogen electrode (RHE) at 0.34 mA cm^?2 for the heterojunction. The photocurrent at the water oxidation potential (1.23 V vs RHE) is 1.46 and 7.26 mA cm^?2 for the coated n‐GaP and n‐GaP/i‐GaNP/p⁺‐GaP photoanodes, respectively. The passivated heterojunction offers a maximum applied bias photon‐to‐current efficiency (ABPE) of 1.9% while the ABPE of the coated n‐GaP sample is almost zero. Furthermore, the coated n‐GaP/i‐GaNP/p⁺‐GaP heterojunction photoanode provides a broad absorption spectrum up to ≈620 nm with incident photon‐to‐current efficiencies (IPCEs) of over 40% from ≈400 to ≈560 nm. The high low‐bias performance and broad absorption of the wide‐bandgap GaP/GaNP heterojunctions render them as a promising photoanode material for tandem photoelectrochemical (PEC) cells to carry out overall solar water splitting.     

Intrusion Detection Based on Bidirectional Long Short-Term Memory with Attention Mechanism

With the recent developments in the Internet of Things (IoT), the amount of data collected has expanded tremendously, resulting in a higher demand for data storage, computational capacity, and real-time processing capabilities. Cloud computing has traditionally played an important role in establishing IoT. However, fog computing has recently emerged as a new field complementing cloud computing due to its enhanced mobility, location awareness, heterogeneity, scalability, low latency, and geographic distribution. However, IoT networks are vulnerable to unwanted assaults because of their open and shared nature. As a result, various fog computing-based security models that protect IoT networks have been developed. A distributed architecture based on an intrusion detection system (IDS) ensures that a dynamic, scalable IoT environment with the ability to disperse centralized tasks to local fog nodes and which successfully detects advanced malicious threats is available. In this study, we examined the time-related aspects of network traffic data. We presented an intrusion detection model based on a two-layered bidirectional long short-term memory (Bi-LSTM) with an attention mechanism for traffic data classification verified on the UNSW-NB15 benchmark dataset. We showed that the suggested model outperformed numerous leading-edge Network IDS that used machine learning models in terms of accuracy, precision, recall and F1 score.     

Growth and fatigue properties of pulsed laser deposited Pb1 − x La x (Zr y Ti z )O3 thin films with [0 0 /] preferred orientation

Growth of [0 0 /] preferentially oriented Pb_{1 - x} La _x (Zr _y Ti _z ) O₃ (PLZT) thin films was carried out by using targets of either tetragonal or rhombohedral structure. The tetragonal films grew in a similar manner to the rhombohedral films. Both the substrate temperature (500 or 550 C) and oxygen pressure (0.1 mbar, 10 Pa) required stringent control in order to deposit [0 0 /]-textured PLZT thin films. The ferroelectric and fatigue properties were examined.The films deposited on YBa₂Cu₃O_{7 - x} and CeO₂ coated silicon (YBCO/CeO₂/Si) substrates possessed substantially lower remanent polarization than those grown on YBCO coated SrTiO₃ (YBCO/STO) substrates; this is ascribed to inferior crystallinity of the PLZT/YBCO/CeO₂/Si films. The remanent polarization of tetragonal PLZT films was degraded insignificantly up to 10⁸ polarization switching cycles, whereas that of rhombohedral PLZT films was already reduced to 80% of the initial value after 10⁸ cycles. Low endurance of rhombohedral films was ascribed to the periodic stress induced when the inclined spontaneous polarization vector (P = [111]) switched. On the other hand, high endurance of tetragonal films was explained by the fact that the spontaneous polarization vector (P = [001]) lies along the film's normal such that switch cycles cause no lateral stress.