Boole函数的线性可分性(Ⅰ)——n-维超立方体的基本理论

Boole函数的线性可分性是前向人工神经网络理论中的一个比较困难的问题之一。目前仅对变量数n≤7的某些问题给予讨论。当n≥8时,尚无判别Boole函数线性可分的一般准则,更无线性可分Boole函数的计数公式。基于此,本文详细地研究了与Boole函数线性可分性有关的n-维超立方体的基本理论,包括n-维超立方体的基本性质、超立方体中的平行线、子超立方体的计数等,并给出了构造n-维超立方体图的一种新方法。     

具有阻力特性材料的σ_f－σ关系及一种新的阻力曲线测定方法

通过对相变增韧陶瓷及一种可切削玻璃－陶瓷动态疲劳（恒应力速率）试验中高应力速率区断裂应力下降现象的理论分析，发现这种现象与材料的阻力特性（Ｒ－ｃｕｒｖｅ）密切相关。确立的σ_ｆ－σ理论关系能够很好地描述整个应力速率区间内的动态疲劳试验结果。高应力速率区σ_ｆ－σ在双对数坐标下为负斜率直线，直线斜率为（ｍ为阻力曲线ＫＲ＝ｋ（△ａ）~ｍ的指数），断裂主要由材料阻力行为控制；低应力速率区，σ_ｆ－σ在双对数坐标下为正斜率直线，直线斜率为　（ｎ为应力腐蚀指数），断裂主要由材料应力腐蚀行为控制。建立了测定材料阻力特性的一种新方法，分别用这种方法及压痕／弯曲方法对一种可切削玻璃－陶瓷的阻力特性进行了实验测定，两种方法所得结果有很好的一致性。     

生物包粒的分类与命名——兼论徐州大北望中下寒武统生物包粒特征

对于碳酸盐岩中由生物及生物化学作用形成的具有同心或放射结构的异化颗粒(生物包粒),已为人们所熟悉,但目前还没有统一的分类与命名。本文在研究徐州大北望中下寒武统生物包粒特征的基础上,结合已有研究成果,对由生物及生物化学作用形成的具有同心或放射状内部结构的碳酸盐颗粒,按其粒度和形成颗粒的生物类型不同进行分类命名,提出分类命名的原则。     

新型低压抽屉式开关柜在冶金系统的应用

介绍新型低压抽屉式开关柜的结构性能和特点，并与早期低压柜作了比较。阐述了新型低压抽屉式开关柜在冶金系统应用中的几点改进情况。     

Diagnosis and correction of multiple logic design errors in digitalcircuits

This paper presents a technique to correct multiple logic design errors in a gate-level netlist. A number of methods have been proposed for correcting single logic design errors. However, the extension of these methods to more than one error is still very limited. We direct our attention to circuits with a low multiplicity of errors. By assuming different error dependency scenarios, multiple errors are corrected by repeatedly applying a single error search and correction algorithm. Experimental results on correcting double-design errors and triple-design errors on ISCAS and MCNC benchmark circuits are included     

Inhibitory effect of ganciclovir on the HSV1-tk positive subcutaneous tumors transplanted with human ovarian cancer in nude mice

OBJECTIVE: To investigate the inhibitory effect in vivo of ganciclovir (GCV) on the growth of human ovarian cancer cells (AO) transducted with the thymidine kinase gene of herpes simplex virus I type (HSV1-tk). METHODS: Tumors were induced in nude mice by subcutaneous injection of AO cells and AO cells carried with HSV1-tk gene from China strain (AO/HSV1-tk cells). When the growing tumors were visible, GCV was injected daily into the peritoneum of the nude mice. RESULTS: The average weights of survived AO/HSV1-tkc tumors and AO tumors treated with GCV were 0.087 +/- 0.036 g and 0.661 +/- 0.260 g respectively. Most of the survived AO/HSV1-tkc cells treated with GCV were characterized by hypertrophy and necrosis, but their nuclear chromatins predominantely took the forms of heterchromatins. CONCLUSIONS: GCV could effectively inhibit the growth of HSV1-tk positive human ovarian cancer cells in vivo, but the nuclei of the survival tumor cells appeared to proliferate actively. As the same results of in vitro experiments, this may suggest that HSV1-tk/GCV gene therapeutic system might be combined with S-phase chemotherapy to increase the long-term effect.     

一种多晶硅高温压力传感器

制作压力传感器时，在二氧化硅层上淀积多晶硅膜，既可利用优良的机械特性，又可保证压敏电阻与衬底间具有良好的绝缘性，由此可大大提高器件的温度特性。介绍了一种多晶硅压力传感器的原理和设计。实验结果表明，这类传感器具有灵敏度好，精度高等特点，电路工作范围为０－２５０℃，且具有良好的温度稳定性。     

Weight smoothing to improve network generalization

A weight smoothing algorithm is proposed in this paper to improve a neural network's generalization capability. The algorithm can be used when the data patterns to be classified are presented on an n-dimensional grid (n>/=1) and there exists some correlations among neighboring data points within a pattern. For a fully-interconnected feedforward net, no such correlation information is embedded into the architecture. Consequently, the correlations can only be extracted through sufficient amount of network training. With the proposed algorithm, a smoothing constraint is incorporated into the objective function of backpropagation to reflect the neighborhood correlations and to seek those solutions that have smooth connection weights. Experiments were performed on problems of waveform classification, multifont alphanumeric character recognition, and handwritten numeral recognition. The results indicate that (1) networks trained with the algorithm do have smooth connection weights, and (2) they generalize better.