Three groups of dynamic triaxial tests were performed for saturated Nanjing fine sand subjected to uniform cyclic loading. The tested curves of the excess pore water pressure （EPWP） ratio variation with the ratio of...Three groups of dynamic triaxial tests were performed for saturated Nanjing fine sand subjected to uniform cyclic loading. The tested curves of the excess pore water pressure （EPWP） ratio variation with the ratio of the number of cycles are provided. The concept of the EPWP increment ratio is introduced and two new concepts of the effective dynamic shear stress ratio and the log decrement of effective stress are defined. It is found that the development of the EPWP increment ratio can be divided into three stages： descending, stable and ascending. Furthermore, at the stable and ascending stages, a satisfactory linear relationship is obtained between the accumulative EPWP increment ratio and natural logarithm of the effective dynamic shear stress ratio. Accordingly, the EPWP increment ratio at the number of cycles N has been deduced that is proportional to the log decrement of effective stress at the cycle number N-l, but is independent of the cyclic stress amplitude. Based on the analysis, a new EPWP increment model for saturated Nanjing fine sand is developed from tested data fitting, which provides a better prediction of the curves of EPWP generation, the number of cycles required for initial liquefaction and the liquefaction resistance.展开更多
Background： Sperm DNA fragmentation（sDF） has been proved to be an important parameter in order to predict in vitro the potential fertility of a semen sample. Colloid centrifugation could be a suitable technique to ...Background： Sperm DNA fragmentation（sDF） has been proved to be an important parameter in order to predict in vitro the potential fertility of a semen sample. Colloid centrifugation could be a suitable technique to select those donkey sperm more resistant to DNA fragmentation after thawing. Previous studies have shown that to elucidate the latent damage of the DNA molecule, sDF should be assessed dynamically, where the rate of fragmentation between treatments indicates how resistant the DNA is to iatrogenic damage. The rate of fragmentation is calculated using the slope of a linear regression equation. However, it has not been studied if s DF dynamics fit this model. The objectives of this study were to evaluate the effect of different after-thawing centrifugation protocols on sperm DNA fragmentation and elucidate the most accurate mathematical model（linear regression, exponential or polynomial） for DNA fragmentation over time in frozen-thawed donkey semen.Results： After submitting post-thaw semen samples to no centrifugation（UDC）, sperm washing（SW） or single layer centrifugation（SLC） protocols, sD F values after 6 h of incubation were significantly lower in SLC samples than in SW or UDC.Coefficient of determination（R-2） values were significantly higher for a second order polynomial model than for linear or exponential. The highest values for acceleration of fragmentation（aSDF） were obtained for SW, fol owed by SLC and UDC.Conclusion： SLC after thawing seems to preserve longer DNA longevity in comparison to UDC and SW. Moreover,the fine-tuning of models has shown that sDF dynamics in frozen-thawed donkey semen fit a second order polynomial model, which implies that fragmentation rate is not constant and fragmentation acceleration must be taken into account to elucidate hidden damage in the DNA molecule.展开更多
The Fine Structure Constant (α) is a dimensionless value that guides much of quantum physics but with no scientific insight into why this specific number. The number defines the coupling constant for the strength of ...The Fine Structure Constant (α) is a dimensionless value that guides much of quantum physics but with no scientific insight into why this specific number. The number defines the coupling constant for the strength of the electromagnetic force and is precisely tuned to make our universe functional. This study introduces a novel approach to understanding a conceptual model for how this critical number is part of a larger design rather than a random accident of nature. The Fine Structure Constant (FSC) model employs a Python program to calculate n-dimensional property sets for prime number universes where α equals the whole number values 137 and 139, representing twin prime universes without a fractional constant. Each property is defined by theoretical prime number sets that represent focal points of matter and wave energy in their respective universes. This work aims to determine if these prime number sets can reproduce the observed α value, giving it a definable structure. The result of the FSC model produces a α value equal to 137.036, an almost exact match. Furthermore, the model indicates that other twin prime pairs also have a role in our functional universe, providing a hierarchy for atomic orbital energy levels and alignment with the principal and azimuthal quantum numbers. In addition, it construes stable matter as property sets with the highest ratio of twin prime elements. These results provide a new perspective on a mathematical structure that shapes our universe and, if valid, has the structural complexity to guide future research.展开更多
Using our recently published electron’s charge electromagnetic flux manifold fiber model of the electron, described by analytical method and numerical simulations, we show how the fine structure constant is embedded ...Using our recently published electron’s charge electromagnetic flux manifold fiber model of the electron, described by analytical method and numerical simulations, we show how the fine structure constant is embedded as a geometrical proportionality constant in three dimensional space of its charge manifold and how this dictates the first QED term one-loop contribution of its anomalous magnetic moment making for the first time a connection of its intrinsic characteristics with physical geometrical dimensions and therefore demonstrating that the physical electron charge cannot be dimensionless. We show that the fine structure constant (FSC) α, and anomalous magnetic moment α<sub>μ</sub> of the electron is related to the sphericity of its charge distribution which is not perfectly spherical and thus has a shape, and therefore its self-confined charge possesses measurable physical dimensions. We also explain why these are not yet able to be measured by past and current experiments and how possible we could succeed.展开更多
随着钓鱼邮件数量的迅速增加以及对抗技术的不断演进,传统的钓鱼邮件检测方法在效率和准确性方面面临严峻挑战.为此,提出了一种基于大语言模型(large language model,LLM)的钓鱼邮件检测方法,以解决现有系统检测率低、漏报率高及人机交...随着钓鱼邮件数量的迅速增加以及对抗技术的不断演进,传统的钓鱼邮件检测方法在效率和准确性方面面临严峻挑战.为此,提出了一种基于大语言模型(large language model,LLM)的钓鱼邮件检测方法,以解决现有系统检测率低、漏报率高及人机交互性差等问题.通过全面分析钓鱼邮件的关键特征,包括邮件头部字段、正文内容、URL、二维码、附件及HTML页面,利用特征插入算法构建高质量的训练数据集.基于预训练语言模型LLaMA和低秩自适应微调技术(low-rank adaptation,LoRA),在仅更新0.72%模型参数(约50 MB)条件下实现领域知识迁移,获得钓鱼邮件检测大模型.实验结果显示,与传统方法相比,基于大语言模型的检测方法显著提升了检测的准确性与鲁棒性,整体准确率达到94.5%,有效降低了误报率,增强了钓鱼邮件特征的分类与解释能力,提供了更具实用性和可靠性的钓鱼邮件检测方案.展开更多
时间序列预测技术在医疗、金融、交通特别是电力领域被广泛应用。随着深度学习技术的发展,深度神经网络已被广泛应用于时间序列预测任务,其预测性能相比传统的预测方法有了显著提升。作为深度学习下一步发展方向的大规模预训练-微调模型...时间序列预测技术在医疗、金融、交通特别是电力领域被广泛应用。随着深度学习技术的发展,深度神经网络已被广泛应用于时间序列预测任务,其预测性能相比传统的预测方法有了显著提升。作为深度学习下一步发展方向的大规模预训练-微调模型,其在计算机视觉和自然语言处理任务上取得成功,但其在时间序列预测领域上的研究还非常有限。文章基于电力领域时间序列数据的特性及时序预测任务的特点,提出一种基于生成式预训练Transformer 2(generative pre-trained transformer 2,GPT-2)的泛用时间序列大规模预测模型面向时间序列的自回归解码(auto-decoding GPT for time series,ADGPT)。ADGPT在微调过程中引入了可学习的分解模块,将输入序列分解为季节组件和趋势组件,以解耦复杂的时态关联。ADGPT在微调过程中只冻结GPT-2模型中的注意力层权重和前馈层参数,以适应时间序列数据的特性。在3个电力领域真实数据集上的实验表明,ADGPT与最先进的时间序列预测模型相比可提高平均9.4%的预测精度,从而验证其在电力领域时间序列预测任务上的有效性。展开更多
基金Key Research Project of National Natural Science Foundation of China Under Grant No.90715018National Basic Research Program of China Under Grant No.2007CB714200the Special Fund for the Commonweal Industry of China Under Grant No.200808022
文摘Three groups of dynamic triaxial tests were performed for saturated Nanjing fine sand subjected to uniform cyclic loading. The tested curves of the excess pore water pressure （EPWP） ratio variation with the ratio of the number of cycles are provided. The concept of the EPWP increment ratio is introduced and two new concepts of the effective dynamic shear stress ratio and the log decrement of effective stress are defined. It is found that the development of the EPWP increment ratio can be divided into three stages： descending, stable and ascending. Furthermore, at the stable and ascending stages, a satisfactory linear relationship is obtained between the accumulative EPWP increment ratio and natural logarithm of the effective dynamic shear stress ratio. Accordingly, the EPWP increment ratio at the number of cycles N has been deduced that is proportional to the log decrement of effective stress at the cycle number N-l, but is independent of the cyclic stress amplitude. Based on the analysis, a new EPWP increment model for saturated Nanjing fine sand is developed from tested data fitting, which provides a better prediction of the curves of EPWP generation, the number of cycles required for initial liquefaction and the liquefaction resistance.
基金partially supported by grants RZ2009-00006-00-00(Instituto Nacional de Investigacion y Tecnología Agraria y Alimentaria,Ministerio de Ciencia e Innovación,Spain)AGL-2013-42726-R(Secretaria de Estado de Investigacion,Desarrollo e Innovacion,Ministerio de Economia y Competitividad,Spain)+1 种基金supported by a Ph.D.fellowship from the ceiA3(Andalucia,Spain)with funding provided by Banco Santander through its Global Division,Santander Universidadesfunded by the Swedish Foundation for Equine Research,Stockholm,Sweden(H14-47-008)
文摘Background： Sperm DNA fragmentation（sDF） has been proved to be an important parameter in order to predict in vitro the potential fertility of a semen sample. Colloid centrifugation could be a suitable technique to select those donkey sperm more resistant to DNA fragmentation after thawing. Previous studies have shown that to elucidate the latent damage of the DNA molecule, sDF should be assessed dynamically, where the rate of fragmentation between treatments indicates how resistant the DNA is to iatrogenic damage. The rate of fragmentation is calculated using the slope of a linear regression equation. However, it has not been studied if s DF dynamics fit this model. The objectives of this study were to evaluate the effect of different after-thawing centrifugation protocols on sperm DNA fragmentation and elucidate the most accurate mathematical model（linear regression, exponential or polynomial） for DNA fragmentation over time in frozen-thawed donkey semen.Results： After submitting post-thaw semen samples to no centrifugation（UDC）, sperm washing（SW） or single layer centrifugation（SLC） protocols, sD F values after 6 h of incubation were significantly lower in SLC samples than in SW or UDC.Coefficient of determination（R-2） values were significantly higher for a second order polynomial model than for linear or exponential. The highest values for acceleration of fragmentation（aSDF） were obtained for SW, fol owed by SLC and UDC.Conclusion： SLC after thawing seems to preserve longer DNA longevity in comparison to UDC and SW. Moreover,the fine-tuning of models has shown that sDF dynamics in frozen-thawed donkey semen fit a second order polynomial model, which implies that fragmentation rate is not constant and fragmentation acceleration must be taken into account to elucidate hidden damage in the DNA molecule.
文摘The Fine Structure Constant (α) is a dimensionless value that guides much of quantum physics but with no scientific insight into why this specific number. The number defines the coupling constant for the strength of the electromagnetic force and is precisely tuned to make our universe functional. This study introduces a novel approach to understanding a conceptual model for how this critical number is part of a larger design rather than a random accident of nature. The Fine Structure Constant (FSC) model employs a Python program to calculate n-dimensional property sets for prime number universes where α equals the whole number values 137 and 139, representing twin prime universes without a fractional constant. Each property is defined by theoretical prime number sets that represent focal points of matter and wave energy in their respective universes. This work aims to determine if these prime number sets can reproduce the observed α value, giving it a definable structure. The result of the FSC model produces a α value equal to 137.036, an almost exact match. Furthermore, the model indicates that other twin prime pairs also have a role in our functional universe, providing a hierarchy for atomic orbital energy levels and alignment with the principal and azimuthal quantum numbers. In addition, it construes stable matter as property sets with the highest ratio of twin prime elements. These results provide a new perspective on a mathematical structure that shapes our universe and, if valid, has the structural complexity to guide future research.
文摘Using our recently published electron’s charge electromagnetic flux manifold fiber model of the electron, described by analytical method and numerical simulations, we show how the fine structure constant is embedded as a geometrical proportionality constant in three dimensional space of its charge manifold and how this dictates the first QED term one-loop contribution of its anomalous magnetic moment making for the first time a connection of its intrinsic characteristics with physical geometrical dimensions and therefore demonstrating that the physical electron charge cannot be dimensionless. We show that the fine structure constant (FSC) α, and anomalous magnetic moment α<sub>μ</sub> of the electron is related to the sphericity of its charge distribution which is not perfectly spherical and thus has a shape, and therefore its self-confined charge possesses measurable physical dimensions. We also explain why these are not yet able to be measured by past and current experiments and how possible we could succeed.
文摘随着钓鱼邮件数量的迅速增加以及对抗技术的不断演进,传统的钓鱼邮件检测方法在效率和准确性方面面临严峻挑战.为此,提出了一种基于大语言模型(large language model,LLM)的钓鱼邮件检测方法,以解决现有系统检测率低、漏报率高及人机交互性差等问题.通过全面分析钓鱼邮件的关键特征,包括邮件头部字段、正文内容、URL、二维码、附件及HTML页面,利用特征插入算法构建高质量的训练数据集.基于预训练语言模型LLaMA和低秩自适应微调技术(low-rank adaptation,LoRA),在仅更新0.72%模型参数(约50 MB)条件下实现领域知识迁移,获得钓鱼邮件检测大模型.实验结果显示,与传统方法相比,基于大语言模型的检测方法显著提升了检测的准确性与鲁棒性,整体准确率达到94.5%,有效降低了误报率,增强了钓鱼邮件特征的分类与解释能力,提供了更具实用性和可靠性的钓鱼邮件检测方案.
文摘时间序列预测技术在医疗、金融、交通特别是电力领域被广泛应用。随着深度学习技术的发展,深度神经网络已被广泛应用于时间序列预测任务,其预测性能相比传统的预测方法有了显著提升。作为深度学习下一步发展方向的大规模预训练-微调模型,其在计算机视觉和自然语言处理任务上取得成功,但其在时间序列预测领域上的研究还非常有限。文章基于电力领域时间序列数据的特性及时序预测任务的特点,提出一种基于生成式预训练Transformer 2(generative pre-trained transformer 2,GPT-2)的泛用时间序列大规模预测模型面向时间序列的自回归解码(auto-decoding GPT for time series,ADGPT)。ADGPT在微调过程中引入了可学习的分解模块,将输入序列分解为季节组件和趋势组件,以解耦复杂的时态关联。ADGPT在微调过程中只冻结GPT-2模型中的注意力层权重和前馈层参数,以适应时间序列数据的特性。在3个电力领域真实数据集上的实验表明,ADGPT与最先进的时间序列预测模型相比可提高平均9.4%的预测精度,从而验证其在电力领域时间序列预测任务上的有效性。
