基于免疫的动态垃圾邮件过滤研究

本文基于生物免疫系统防御机制,探讨了将人工免疫原理应用于垃圾邮件过滤系统(IDSFS),通过动态生成垃圾邮件疫苗,提高系统的检出率.另外,针对传统邮件过滤系统存在的错误肯定率、错误否定率等不足及用户对邮件个性化处理的要求,引入用户反馈机制,使得系统精确率、正确率进一步提高.实验结果表明,该系统具有较强的自学习、自适应能力.     

基于马尔可夫决策过程的网格资源发现

传统的网格资源发现方法没有考虑节点和资源本身性能的优劣性。针对这一问题,提出了基于马尔可夫决策过程(MDP)的网格资源发现方法,利用其对固定目标的最优搜索理论建立MDP模型实现报酬最优的资源发现,并对该模型仿真分析,验证其性能。     

一种新型的适应于MANET的基于移动代理的对等入侵检测系统

移动自组织网络(Mobile Ad hoc Networks,MANET)是一种开放性的媒介系统,从而很容易受到黑客的入侵.提出了一种有效的适应于MANET的入侵检测系统,叫做MAPIDS(Mobile Agent-based Peer-to-peer Intrusion Detection System).该系统一旦发现网络中的可疑行为,就会引发一次集体投票过程,以确定该可疑行为是否是入侵行为,并根据投票结果作出相应的进一步处理.相比于其他同样基于集体决策机制的入侵检测系统,MAPIDS节省了更多的网络带宽和电池能耗,并且该系统不存在稀疏节点问题.     

石墨烯基导热薄膜的研究进展

随着现代技术的发展,散热和导热已成为制约芯片器件小型化和大功率制造业发展的关键问题之一。由于传统的金属导热材料存在密度大和易氧化等问题,近年来以石墨烯基材料为代表的非金属碳基材料逐渐成为国内外的研究热点。本文综述了近年国内外石墨烯导热薄膜的制备方法及最新研究成果,分析讨论了热处理工艺、晶粒尺寸、薄膜密度及杂质原子和缺陷等对石墨烯导热薄膜性能的影响和物理机理,并对该领域的发展趋势进行了展望。     

Responses of Growth,Oxidative Injury and Chloroplast Ultrastructure in Leaves of Lolium perenne and Festuca arundinacea to Elevated O3 Concentrations

The effects of increasing atmospheric ozone (O₃) concentrations on cool-season plant species have been well studied, but little is known about the physiological responses of cool-season turfgrass species such as Lolium perenne and Festuca arundinacea exposed to short-term acute pollution with elevated O₃ concentrations (80 ppb and 160 ppb, 9 h d⁻¹) for 14 days, which are widely planted in urban areas of Northern China. The current study aimed to investigate and compare O₃ sensitivity and differential changes in growth, oxidative injury, antioxidative enzyme activities, and chloroplast ultrastructure between the two turf-type plant species. The results showed that O₃ decreased significantly biomass regardless of plant species. Under 160 ppb O₃, total biomass of L. perenne and F. arundinacea significantly decreased by 55.3% and 47.8% (p < 0.05), respectively. No significant changes were found in visible injury and photosynthetic pigment contents in leaves of the two grass species exposed to 80 ppb O₃, except for 160 ppb O₃. However, both 80 ppb and 160 ppb O₃ exposure induced heavily oxidative stress by high accumulation of malondialdehyde and reactive oxygen species in leaves and damage in chloroplast ultrastructure regardless of plant species. Elevated O₃ concentration (80 ppb) increased significantly the activities of superoxide dismutase, catalase and peroxidaseby 77.8%, 1.14-foil and 34.3% in L. perenne leaves, and 19.2%, 78.4% and 1.72-fold in F. arundinacea leaves, respectively. These results showed that F. arundinacea showed higher O₃ tolerance than L. perenne. The damage extent by elevated O₃ concentrations could be underestimated only by evaluating foliar injury or chlorophyll content without considering the internal physiological changes, especially in chloroplast ultrastructure and ROS accumulation.     

Silicon-based inorganic-organic hybrid optoelectronic synaptic devices simulating cross-modal learning

With advances in Si-based technology infrastructures and the rapid integration of Si-based op-toelectronics,Si-based optoelectronic synaptic devices have the po...     

铁矿石的烧结基础特性之新概念

长期以来,人们对烧结用铁矿石的评价主要集中在化学成分、粒度组成、矿物特征等常温性能方面,而对铁矿石在烧结过程中的高温行为和作用知之甚少.根据对这一问题的深入思考和实验室的基础研究,提出了\     

A Predictive Instrument for Sensitive and Expedited Measurement of Ultra-Barrier Permeation

The reliable operation of flexible display devices poses a significant engineering challenge regarding the metrology of high barriers against water vapor. No reliable results have been reported in the range of 10^–6 g∙(m²∙d)⁻¹, and there is no standard ultra-barrier for calibration. To detect trace amount of water vapor permeation through an ultra-barrier with extremely high sensitivity and a greatly reduced test period, a predictive instrument was developed by integrating permeation models into high-sensitivity mass spectrometry measurement based on dynamic accumulation, detection, and evacuation of the permeant. Detection reliability was ensured by means of calibration using a standard polymer sample. After calibration, the lower detection limit for water vapor permeation is in the range of 10^–7 g∙(m²∙d)⁻¹, which satisfies the ultra-barrier requirement. Predictive permeation models were developed and evaluated using experimental data so that the steady-state permeation rate can be forecasted from non-steady-state results, thus enabling effective measurement of ultra-barrier permeation within a significantly shorter test period.