BP网络应用于黄河水质的预测研究

利用黄河兰州段水质指标浓度时间序列作为学习样本，选取了DO、COD、氨氮、重金属离子等l0项指标作为输出参数，运用人工神经网络BP模型中的Levenberg—Marguardt优化算法对学习样本优化建模，将优化好的网络的预测结果与实测结果进行了比较，表明BP网络可以很好地用于黄河水质指标值及水质类别的预测。     

基于改进的卷积神经网络的钢号识别

为获取样本的多样性特征,提出了一种改进的卷积神经网络结构。该网络中引入多层递归神经网络,利用卷积神经网络提取输入图像的浅层特征,同时利用卷积神经网络和递归神经网络并行提取高层特征,最后将两种网络学习到的特征进行融合输入到分类器中分类。利用迁移学习理论解决小样本集数据训练不足的问题,并将这种卷积神经网络结构应用于石油物资管线钢号识别中。实验探究了递归神经网络个数与卷积核个数对网络性能的影响,实验结果表明,改进的网络结构与其它网络进行对比,错误率降低了 3% 。     

Ubiquitin-proteasome pathway as a primary defender against TRAIL-mediated cell death

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) induces apoptotic cell death as well as expression of proinflammatory genes such as CXCL8 in malignant human astrocytoma cells. However, the molecular mechanisms that determine the fate of cells are not yet understood. The ubiquitin (Ub)-proteasome pathway regulates a wide range of cellular functions through degradation of various regulatory proteins; given this, we hypothesized that this pathway may play a central role in TRAIL-mediated signaling. We demonstrate here that inhibition of the Ub-proteasome pathway enhanced TRAIL-mediated cell death of human astrocytoma CRT-MG cells within hours by blocking degradation of active caspase-8 and -3. Proteasome inhibitors suppressed TRAIL-mediated activation of NF-B; however, inhibition of the NF-B pathway alone was not sufficient to enhance TRAIL-mediated cell death. Collectively, these results suggest that the Ub-proteasome pathway may play an important role as an antiapoptotic surveillance system by eliminating activated caspases as well as mediating NF-B-dependent signals.Received 30 December 2003; received after revision 9 February 2004; accepted 13 February 2004     

油田管道风险及完整性评价体系构建与应用

为提升油田管道安全隐患的预警和分析能力，根据 GB32167-2015《油气输送管道完整性管理规范》的要 求和管道密集型油田的实际特点，对Ⅲ类管道进一步精细划分; 同时结合地理和周边居住区情况，对管道的高 后果区进行有效识别，从而对管道进行风险等级分类，完善管道完整性的评价体系。通过互联网 + 数据驱动的 模式建立油田管道完整性数据管理平台，以可视化方式全面评价油田管道运行风险。该平台可提高油田管道 风险评估的准确性，确保及时准确预测风险。     

A molecular solution to the riddle of the giant panda's phylogeny

Although it is generally agreed that the giant panda (Ailuropoda melanoleuca) is a member of the order Carnivora, there has long been disagreement over whether it should be classified with bears, raccoons or as a single member of its own family. Four independent molecular and genetic measures lead to a consensus phylogeny for the giant and lesser pandas. The lesser panda diverged from New World procyonids at approximately the same time as their departure from ursids, while ancestors of the giant panda split from the ursid lineage much later, just before the radiation which led to modern bears. The giant panda's divergence was accompanied by a chromosomal reorganization which can be partially reconstructed from the ursid karyotype, but not from that of procyonids or the lesser panda. The apparently dramatic, but actually limited, distinctions between the giant panda and the bears in chromosomal and anatomical morphology provide a graphic mammalian example of the discordance of molecular and morphological (and chromosomal) evolutionary change.     

无人机航迹规划常用算法综述

为促进航迹规划技术的发展,对航迹规划常用算法进行综述。首先对航迹规划的规划思想和构成进行分析; 其次将航迹规划算法分为传统经典算法和现代智能算法两大类,对其中几种常用算法进行分析总结; 最后阐述现代智能算法在航迹规划应用中的改进、多重算法的融合改进以及多无人机四维航迹规划算法研究3 个研究热点及未来发展趋势。     

火灾报警系统的研究

介绍了火灾报警系统的硬件设置与软件设计原理，讨论了该系统安装调试过程中的有关技术问题，阐述了对该系统进行的模拟联动试验，证明该系统的技术是先进的，工作是可靠的，每个工作环节都准确无误。     

Macrophage origin of platelet activating factor]

Platelet-activating factor (P.A.F.) is a mediator of anaphylaxis released from human and Rabbit basophils which causes aggregation of platelets and release of their vasoactive amines. We have induced the release of P.A.F. from Rat peritoneal cells (P.C.) with ionophore A 23187. After fractionation of P.C. on 5-15% Ficoll gradients, P.A.F. was obtained from macrophage-rich but not from mastocyte-rich fractions and from adherent cells but not from non adherent cells. These data suggest an important new function for the macrophage: aggregation of platelets and release of their vasoactive amines and others mediators of inflammation.