Sensor networks tend to support different traffic patterns since more and more emerging applications have diverse needs. We present MGRP, a Multi-Gradient Routing Protocol for wireless sensor networks, which is fully ...Sensor networks tend to support different traffic patterns since more and more emerging applications have diverse needs. We present MGRP, a Multi-Gradient Routing Protocol for wireless sensor networks, which is fully distributed and efficiently supports endto-end, one-to-many and many-to-one traffic patterns by effectively construct and maintain a gradient vector for each node. We further combine neighbor link estimation with routing information to reduce packet exchange on network dynamics and node failures. We have implemented MGRP on Tiny OS and evaluated its performance on real-world testbeds. The result shows MGRP achieves lower end-to-end packet delay in different traffic patterns compared to the state of the art routing protocols while still remains high packet delivery ratio.展开更多
Multi-gradient drilling is a new offshore drilling method.The accurate calculation of the related wellbore temperature is of great significance for the prediction of the gas hydrate formation area and the precise cont...Multi-gradient drilling is a new offshore drilling method.The accurate calculation of the related wellbore temperature is of great significance for the prediction of the gas hydrate formation area and the precise control of the wellbore pressure.In this study,a new heat transfer model is proposed by which the variable mass flow is properly taken into account.Using this model,the effects of the main factors influencing the wellbore temperature are analyzed.The results indicate that at the position where the separation injection device is installed,the temperature increase of the fluid in the drill pipe is mitigated due to the inflow/outflow of hollow spheres,and the temperature drop of the fluid in the annulus also decreases.In addition,a lower separation efficiency of the device,a shallower installation depth and a smaller circulating displacement tend to increase the temperature near the bottom of the annulus,thereby helping to reduce the hydrate generation area and playing a positive role in the prevention and control of hydrates in deepwater drilling.展开更多
Peripheral nerve defects present complex orthopedic challenges with limited efficacy of clinical interventions.The inadequate proliferation and dysfunction of Schwann cells within the nerve scaffold impede the effecti...Peripheral nerve defects present complex orthopedic challenges with limited efficacy of clinical interventions.The inadequate proliferation and dysfunction of Schwann cells within the nerve scaffold impede the effectiveness of nerve repair.Our previ-ous studies suggested the effectiveness of a magnesium-encapsulated bioactive hydrogel in repairing nerve defects.However,its rapid release of magnesium ions limited its efficacy to long-term nerve regeneration,and its molecular mechanism remains unclear.This study utilized electrospinning technology to fabricate a MgO/MgCO3/polycaprolactone(PCL)multi-gradient nanofiber membrane for peripheral nerve regeneration.Our findings indicated that by carefully adjusting the concentration or proportion of rapidly degradable MgO and slowly degradable MgCO3,as well as the number of electrospun layers,the multi-gradient scaffold effectively sustained the release of Mg2+over a period of 6 weeks.Additionally,this study provided insight into the mechanism of Mg2+-induced nerve regeneration and confirmed that Mg2+effectively promoted Schwann cell proliferation,migration,and transition to a repair phenotype.By employing transcriptome sequencing technology,the study identified the Wingless/integrase-1(Wnt)signaling pathway as a crucial mechanism influencing Schwann cell function during nerve regeneration.After implantation in 10 mm critically sized nerve defects in rats,the MgO/MgCO3/PCL multi-gradient nanofiber combined with a 3D-engineered PCL nerve conduit showed enhanced axonal regeneration,remyelination,and reinnervation of muscle tissue 12 weeks post-surgery.In conclusion,this study successfully developed an innovative multi-gradient long-acting MgO/MgCO3/PCL nanofiber with a tunable Mg2+release property,which underscored the molecular mechanism of magnesium-encapsulated biomaterials in treating nervous system diseases and established a robust theoretical foundation for future clinical translation.展开更多
为解决深度神经网络(Deep Neural Network,DNN)对抗防御方法在提升防御效力的同时牺牲干净样本分类精度的问题,提出一种基于随机多分布噪声与多目标梯度的对抗防御(Random Noise-Defense,RN-Defense)方法。该方法的核心思想是在模型的...为解决深度神经网络(Deep Neural Network,DNN)对抗防御方法在提升防御效力的同时牺牲干净样本分类精度的问题,提出一种基于随机多分布噪声与多目标梯度的对抗防御(Random Noise-Defense,RN-Defense)方法。该方法的核心思想是在模型的每一层引入可训练的多分布噪声,且噪声的强度和分布可通过梯度信息动态调整,以增强模型对多样化输入的适应能力。此外,结合多目标梯度对抗训练,控制每个像素点在梯度方向上进行低阶灰度扰动,以生成对抗样本用于对抗训练。实验结果表明,在CIFAR-10(Canadian Institute for Advanced Research-10)和MNIST(Modified National Institute of Standards and Technology)数据集上,与投影梯度下降(Projected Gradient Descent,PGD)对抗训练相比,RN-Defense方法在PGD攻击下的分类精度分别提升6.4%和7.7%,干净样本的分类精度分别提高6.4%和6.9%,有效提升了模型对抗样本的防御能力,同时保证了干净样本的分类精度。展开更多
为解决多无人机博弈对抗中现有决策方法在复杂场景下奖励设计烦琐,提升算法训练稳定性与样本效率。将多智能体强化学习与元强化学习结合,提出评论家与元策略联合增强的多智能体深度确定性策略梯度(Critic and Meta enhanced multi-agent...为解决多无人机博弈对抗中现有决策方法在复杂场景下奖励设计烦琐,提升算法训练稳定性与样本效率。将多智能体强化学习与元强化学习结合,提出评论家与元策略联合增强的多智能体深度确定性策略梯度(Critic and Meta enhanced multi-agent deep deterministic policy gradient,CM-MADDPG)算法。首先,通过在集中式评论家侧引入适度的结构与学习增强,提升跨机交互的价值估计精度并抑制过估计,从而使训练更稳定,稳态性能更高。其次,在训练期间采用元策略进行自适应调度(如动态调整损失权重与探索强度),提升样本效率与收敛速度,且不增加推理期开销。最后,在相同实验条件下,红方无人机分别采用多智能体深度确定性策略梯度(multi-agent deep deterministic policy gradient,MADDPG)、评论家增强的多智能体深度确定性策略梯度(Critic multi-agent deep deterministic policy gradient,Critic-MADDPG)、元策略增强的多智能体深度确定性策略梯度(Meta enhanced multi-agent deep deterministic policy gradient,Meta-MADDPG),以及CM-MADDPG与随机机动的蓝方无人机进行博弈对抗。仿真结果表明,CM-MADDPG在最终胜率、各无人机奖励,以及收敛速度等指标上均取得显著提升。证明CM-MADDPG在不增加推理计算开销的前提下成功缓解稀疏奖励与训练不稳定问题,提高收敛速度与样本效率。研究成果可为多无人机博弈理论提供一种新思路和途径。展开更多
基金supported by National Key Technologies Research and Development Program of China under Grant No.2014BAH14F01National Science and Technology Major Project of China under Grant No.2012ZX03005007+1 种基金National NSF of China Grant No.61402372Fundamental Research Funds for the Central Universities Grant No.3102014JSJ0003
摘要Sensor networks tend to support different traffic patterns since more and more emerging applications have diverse needs. We present MGRP, a Multi-Gradient Routing Protocol for wireless sensor networks, which is fully distributed and efficiently supports endto-end, one-to-many and many-to-one traffic patterns by effectively construct and maintain a gradient vector for each node. We further combine neighbor link estimation with routing information to reduce packet exchange on network dynamics and node failures. We have implemented MGRP on Tiny OS and evaluated its performance on real-world testbeds. The result shows MGRP achieves lower end-to-end packet delay in different traffic patterns compared to the state of the art routing protocols while still remains high packet delivery ratio.
基金funded by the Key Program of National Natural Science Foundation of China(Grant No.51734010)the Startup Fund of Changzhou University Science Research(Grant No.ZMF22020060).
摘要Multi-gradient drilling is a new offshore drilling method.The accurate calculation of the related wellbore temperature is of great significance for the prediction of the gas hydrate formation area and the precise control of the wellbore pressure.In this study,a new heat transfer model is proposed by which the variable mass flow is properly taken into account.Using this model,the effects of the main factors influencing the wellbore temperature are analyzed.The results indicate that at the position where the separation injection device is installed,the temperature increase of the fluid in the drill pipe is mitigated due to the inflow/outflow of hollow spheres,and the temperature drop of the fluid in the annulus also decreases.In addition,a lower separation efficiency of the device,a shallower installation depth and a smaller circulating displacement tend to increase the temperature near the bottom of the annulus,thereby helping to reduce the hydrate generation area and playing a positive role in the prevention and control of hydrates in deepwater drilling.
基金supported by the National Key R&D Program of China(2022YFB3808000)the National Natural Science Foundation of China(82402802,82404113,82302713,U23A20490)+6 种基金the China Postdoctoral Science Foundation(2023M742390)the Guangdong Basic and Applied Basic Research Foundation(2022A1515012663,2023A1515220250,2023A1515111068)the Shenzhen Science and Technology Innovation Program(RCBS20231211090537061,JCYJ20230807095203007,JCYJ20230807095121041)the Shenzhen Key Medical Discipline Construction Fund(SZXK023)the Sanming Project of Medicine in Shenzhen(SZSM202211038)the Shenzhen High-level Hospital Construction Fund,and the Scientific Research Foundation of Peking University Shenzhen Hospital(LCYJZD2021005,KYQD2023244,KYQD2023245)The authors also gratefully acknowledged the kindly financial support provided by the Youth Talent Support Program of the China Association for Science and Technology,and the Top Young Talents of Foal Eagle Program of Fujian Province to Jin Zhang.
摘要Peripheral nerve defects present complex orthopedic challenges with limited efficacy of clinical interventions.The inadequate proliferation and dysfunction of Schwann cells within the nerve scaffold impede the effectiveness of nerve repair.Our previ-ous studies suggested the effectiveness of a magnesium-encapsulated bioactive hydrogel in repairing nerve defects.However,its rapid release of magnesium ions limited its efficacy to long-term nerve regeneration,and its molecular mechanism remains unclear.This study utilized electrospinning technology to fabricate a MgO/MgCO3/polycaprolactone(PCL)multi-gradient nanofiber membrane for peripheral nerve regeneration.Our findings indicated that by carefully adjusting the concentration or proportion of rapidly degradable MgO and slowly degradable MgCO3,as well as the number of electrospun layers,the multi-gradient scaffold effectively sustained the release of Mg2+over a period of 6 weeks.Additionally,this study provided insight into the mechanism of Mg2+-induced nerve regeneration and confirmed that Mg2+effectively promoted Schwann cell proliferation,migration,and transition to a repair phenotype.By employing transcriptome sequencing technology,the study identified the Wingless/integrase-1(Wnt)signaling pathway as a crucial mechanism influencing Schwann cell function during nerve regeneration.After implantation in 10 mm critically sized nerve defects in rats,the MgO/MgCO3/PCL multi-gradient nanofiber combined with a 3D-engineered PCL nerve conduit showed enhanced axonal regeneration,remyelination,and reinnervation of muscle tissue 12 weeks post-surgery.In conclusion,this study successfully developed an innovative multi-gradient long-acting MgO/MgCO3/PCL nanofiber with a tunable Mg2+release property,which underscored the molecular mechanism of magnesium-encapsulated biomaterials in treating nervous system diseases and established a robust theoretical foundation for future clinical translation.
摘要为解决深度神经网络(Deep Neural Network,DNN)对抗防御方法在提升防御效力的同时牺牲干净样本分类精度的问题,提出一种基于随机多分布噪声与多目标梯度的对抗防御(Random Noise-Defense,RN-Defense)方法。该方法的核心思想是在模型的每一层引入可训练的多分布噪声,且噪声的强度和分布可通过梯度信息动态调整,以增强模型对多样化输入的适应能力。此外,结合多目标梯度对抗训练,控制每个像素点在梯度方向上进行低阶灰度扰动,以生成对抗样本用于对抗训练。实验结果表明,在CIFAR-10(Canadian Institute for Advanced Research-10)和MNIST(Modified National Institute of Standards and Technology)数据集上,与投影梯度下降(Projected Gradient Descent,PGD)对抗训练相比,RN-Defense方法在PGD攻击下的分类精度分别提升6.4%和7.7%,干净样本的分类精度分别提高6.4%和6.9%,有效提升了模型对抗样本的防御能力,同时保证了干净样本的分类精度。
摘要为解决多无人机博弈对抗中现有决策方法在复杂场景下奖励设计烦琐,提升算法训练稳定性与样本效率。将多智能体强化学习与元强化学习结合,提出评论家与元策略联合增强的多智能体深度确定性策略梯度(Critic and Meta enhanced multi-agent deep deterministic policy gradient,CM-MADDPG)算法。首先,通过在集中式评论家侧引入适度的结构与学习增强,提升跨机交互的价值估计精度并抑制过估计,从而使训练更稳定,稳态性能更高。其次,在训练期间采用元策略进行自适应调度(如动态调整损失权重与探索强度),提升样本效率与收敛速度,且不增加推理期开销。最后,在相同实验条件下,红方无人机分别采用多智能体深度确定性策略梯度(multi-agent deep deterministic policy gradient,MADDPG)、评论家增强的多智能体深度确定性策略梯度(Critic multi-agent deep deterministic policy gradient,Critic-MADDPG)、元策略增强的多智能体深度确定性策略梯度(Meta enhanced multi-agent deep deterministic policy gradient,Meta-MADDPG),以及CM-MADDPG与随机机动的蓝方无人机进行博弈对抗。仿真结果表明,CM-MADDPG在最终胜率、各无人机奖励,以及收敛速度等指标上均取得显著提升。证明CM-MADDPG在不增加推理计算开销的前提下成功缓解稀疏奖励与训练不稳定问题,提高收敛速度与样本效率。研究成果可为多无人机博弈理论提供一种新思路和途径。
基金新能源汽车动力总成技术北京市重点实验室开放课题哈尔滨工程大学博士研究生校长创新基金项目+2 种基金国家留学基金项目(No.202406680033)海洋负排放国际大科学计划Agency for Science,Technology and Research(A*STAR)Project Low-carbon Energy Research(No.U2102d2005)。