对经典隐马尔可夫模型学习算法的改进

改进经典隐马尔可夫模型(HMM)的状态转移和输出观测值的假设条件,并在经典隐马尔可夫模型的基础上导出新模型的学习算法.新算法避免了经典隐马尔可夫模型中状态转移概率和输出观测值概率计算时只考虑当前状态而不考虑历史的简单做法.     

新型氨基酸尾式卟啉及莫锌配合物的合成与表征

金属卟啉配合物作为主体分子对氨基酸酯及其衍生物的分子识别研究是当今卟啉仿生化学的重要课题[1～6].这一领域的研究成果在生物传感器的研制,癌症的诊断和治疗,模拟含卟啉生物大分子的功能以及在生物大分子构象和类型的识别研究中具有潜在的应用价值[7].     

活用情境，巧设问题，生成发展

二次函数的概念教学,以生成性教学理论为指导,依据学生原有的知识基础和学习经验确定学习目标,通过创设问题情境,实行课堂灵动对话,让学生自然获得研究二次函数概念要素之间关系的方法,自然生成二次函数概念,形成研究数学概念的思维能力,发展数学直观、数学抽象等素养.     

Preparation,Characterization,and in vitro Release of Biodegradable Erythromycin-gelatin Microspheres

Blank and erythromycin-loaded gelatin microspheres were successfully fabricated via emulsion chemical- crosslinking technique. The surface morphology of the microspheres was characterized by scanning electron microscope（SEM） and optical microscope. The results show that the microspheres were spherical and smooth. The particle average size of erythromycin-loaded microspheres was found to be 20.6 μm, with a high purity of more than 90% and with a good dispersibility. The microspheres could be obtained in a high yield. Erythromycin released from the microspheres was monitored in buffer and artificial body fluid at 37 ℃. Average drug content was 27.2%, and erythromycin-loaded gelatin microspheres showed good release profiles with a nearly constant release during 4-8 h in artificial body fluid in vitro degradation studies. These gelatin microspheres are useful for studying and developing various drug-delivery systems.     

Synthesis, Structure Characterization, and Cu^2＋ Recognition of 3-{ [3- （Phenylsulfonamido） benzoyl ] methylidene } -3,4- dihydroquinoxaline-2 （1H） -one

IntroductionSubstituted diketo acid derivatives possess a varie-ty of biological properties, such as analgesic[1—3], anti-bacterial[4]as well as anti-inflammatory[3]. Very re-cently, the results of random screening of millions ofcompounds have shown that…     

Zn(Thr)Ac2·2H2O固态配合物的制备及标准生成焓

在水-丙酮混合溶剂中合成了未见文献报导的Zn(Thr)Ac2@2H2O固态配合物,通过化学分析、元素分析、IR、XRD和TG-DTG等对其组成、结构及热稳定性进行了研究.用微量热法测定了配合物在298.15 K时在纯水中的溶解焓,计算了Zn(Thr)2+(aq,∞)和Zn(Thr)Ac2@2H2O(s)的标准摩尔生成焓分别为(955.24±5.70)kJ@mol-1和(-570.92±5.71)kJ@mol-1.     

医学影像处理的深度学习可解释性研究进展

随着医学影像数据的迅速增长,传统的影像分析方法给医生带来巨大挑战。利用计算机视觉技术提供自动或半自动辅助诊断,可大大缓解人工阅片压力,提高诊断的准确性,促进医疗流程的标准化建设等。目前,深度学习卷积神经网络在医学影像处理中已取得不俗表现,但深度学习“黑匣子”的不可解释性阻碍了智能医疗诊断的发展。为增强对医学影像数据处理的深度学习可解释性的了解,对近几年相关研究进展进行了综述。首先,综述了深度学习在医学领域的应用现状及面临的问题,对神经网络的可解释性内涵进行了讨论;然后,从现有深度学习可解释性的常见方法出发,重点讨论了医学影像处理的深度学习可解释性研究进展;最后,探讨了医学影像处理的深度学习可解释性的发展趋势。     

四种新的水杨醛缩氨基酞菁Schiff碱的合成及光学性质

以四氨基金属酞菁为原料，合成了四种新的Schiff碱C60H36N12O4M(M=Cu，Ni，Zn，Co)，并用质谱、元素分析、红外、紫外对其结构进行了分析和表征．通过荧光光谱对其光学性质进行了研究，四个Schiff碱官能团及中心金属对其光学性质及稳定性有一定的影响．     

铜的喹哪啶配合物[Cu(Me-quin)2H2O]的合成、表征及晶体结构

本文合成了铜的喹哪啶配合物[Cu(Me-quin)     

Preparation and Characterization of Polyborosiloxanesand Their Blends with Phenolic Resin as ShapeableCeramic Precursors简

Two polyborosiloxanes(PBSis) with char yield up to 74.13% at 800 °C were synthesized by the direct polycondensation of boric acid with phenyltrimethoxysilane in diglyme. The PBSis were characterized by gel permeation chromatography, IR spectroscopy as well as1H-,29Si- and11B-NMR. PBSi modified phenol-formaldehyde resins(PBSi/PFs) were prepared at different PBSi/PF mass ratios and were cured at 150 °C. The PBSi/PFs were characterized by IR spectroscopy, scanning electron microscopy, thermogravimetric analysis and tensile test. The results revealed that the cured PBSi/PFs had sea-island morphology and higher char yield than the common PF. PBSi/PF blend with PBSi/PF mass ratio of 0.4:1 had char yield up to 70.83% at 800 °C. The PBSi/PFs had tensile strength similar to PF. The ceramization of PBSi/PFs was also studied. The silicon boron oxycarbide(SiBOC) ceramics formed were characterized by IR spectroscopy and elemental analysis. This method provided a valuable way to prepare easily shapeable polymer blends as ceramic precursors.