The concept of intelligibility in modern physics (1948)

This is an English translation of Paul Feyerabend's earliest extant essay “Der Begriff der Verständlichkeit in der modernen Physik” (1948). In it, Feyerabend defends positivism as a progressive framework for scientific research in certain stages of scientific development. He argues that in physics visualizability (Anschaulichkeit) and intelligibility (Verständlichkeit) are time-conditioned concepts: what is deemed visualizable in the development of physical theories is relative to a specific historical context and changes over time. He concludes that from time to time the abandonment of visualizability is crucial for progress in physics, as it is conducive to major theory change, illustrating the point on the basis of advances in atomic theory.     

Ostracod fauna of salt Lake Acıgöl (Acı Tuz) (Turkey)

Lakes and wetlands in Mediterranean regions are overexploited and contaminated without sufficient local governmental control. The biodiversity of Salt Lake Ac?göl (Turkey), an internationally recognized ecosystem, is presently endangered by management plans. This study aims to analyse the ostracod community to establish its ecological status as a base for future paleoecological studies using ostracods to reconstruct the recent evolution of this wetland. Samples were collected seasonally and 13 species were recorded. Multivariate classification and ordination statistical methods revealed a major difference between assemblages related to the relative abundance of species tolerant of desiccation vs. species preferring permanent waters. Cyprideis torosa, mainly found in saline, low diversity sites, is a euryhaline species regarded as an indicator of future lake conditions if inappropriate water resource management continues.     

H.264码率控制算法研究

码率控制作为H．264中的关键技术,带宽受限或存储容量受限的条件下,是影响H．246视频传输的重要因素。码率控制原理就是根据现有的传输条件在GOP层、Frame层和基本单元层分配目标比特数并因此改变量化参数,从而达到控制输出码流的目的。探讨了基于Lagrangian优化算法的编码控制模型的H．264编码器。     

基于Cholesky分解的K2DPCA人脸识别研究

K2DPCA(kernel 2D principal component analysis)是基于非线性特征提取的重要人脸识别方法,具有成功的应用.但对大规模训练数据库,其因核矩阵K规模过大、计算代价高而不能有效实现.采用选主元Cholesky,分解方法,仅需计算核矩阵的对角线上元素和部分精选列,得到迹范数意义下核矩阵K的最优Nystr(o|¨)m型低秩近似LL~T来解决该问题.并只需计算小规模矩阵L~TL的特征值和特征向量,实现大规模K2DPCA/KPCA(kernel principal component anialysis)的非线性特征提取.在加噪ORL人脸数据库上的实验结果表明,较K2DPCA/KPCA方法,新方法显著提高了识别率,并可以很大程度上克服噪声的影响;在Extended YaleB大型人脸数据库上的实验结果表明,此算法解决了K2DPCA核矩阵过大而不能有效实现的缺点.     

一类超线性分数阶Schr?dinger方程解的多重性

考虑了一类超线性分数阶Schr?dinger方程,当非线性项f满足广义次临界条件及其它条件时,利用对称山路引理和变分方法,得到了该类方程无穷多个大解的存在性,推广了已有的研究结果.     

带有一般非线性项的分数阶Kirchhoff-Schr?dinger-Poisson系统的变号解

研究了R~3上分数阶Kirchhoff-Schr?dinger-Poisson系统,当非线性项满足广义次临界以及某种单调性条件时,利用变号Nehari流形和定量形变引理,获得了该系统有基态变号解且仅变号一次.     

一类Choquard型拟线性Schrödinger方程解的存在性

考虑一类Choquard型拟线性Schrödinger方程解的存在性。通过变量替换方法将该方程转化为半线性方程,利用山路定理和变分法证明了该方程在适当条件下非平凡解的存在性。     

变分方法及其在非线性偏微分方程应用方面的进展和未决问题

先介绍变分法发展的简单历史以及将来的发展趋势. 然后综述变分法应用于非线性偏微分方程的基本思想和最新成果. 通俗介绍环绕理论、变号临界点理论及应用,其中包括对称扰动方程和Rabinowitz公开问题、Brezis-Nirenberg临界指数方程、Li-Lin公开问题、Bose-Einstein凝聚、Berestycki-Caffarelli-Nirenberg猜测和Lane-Emden方程及猜想.     

局部紧Lindelöf空间的序列覆盖映象

利用特定的覆盖性质，借助序列覆盖映射，建立局部紧Lindeloef空间和CS覆盖、sn覆盖、cs^＊覆盖之间的联系，以及局部紧Lindeloef空间的序列覆盖映象特征，得到并深化局部紧Lindeloef空间的一些相应结果，完善了局部紧Lindeloef空间的映射理论。