基于进化神经网络的入侵检测模型

当前的入侵检测技术主要有基于规则的误用检测和基于统计的异常检测。本文提出一个基于遗传算法的神经网络入侵检测系统模型，该模型将神经网络与遗传算法结合起来，利用神经网络自学习、自适应的特性，同时克服了神经网络易陷入局部最优，训练速度慢的缺点。该模型具有智能特性，能够较好地识别新的攻击。     

Just Use What You Have: Ethical Fashion Discourse and the Feminisation of Responsibility

ABSTRACT

The rise of fast fashion has meant that young women (even those on relatively low incomes) are able to ‘regularly consume and discard fashionable clothing’ [Buckley, Cheryl, and Hazel Clark. 2012. “Conceptualizing Fashion in Everyday Lives.” Design Issues 28 (4): 18–28. doi:10.1162/DESI_a_00172., 21]. While this development may be aligned with the democratisation of fashion—the fact that the supply chains that deliver fast fashion are not consistent with the principles of global democracy is now also relatively common knowledge in the democratised West. This, along with growing awareness of the ecological harms associated with the fashion industry has contributed to what Elke Gaugele [2014. Aesthetic Politics in Fashion. Vienna: Sternberg Press] has termed the ‘ethical turn’ in fashion. However, despite the fact that young women are often not deemed capable of translating their (ethical) attitudes into (ethical) behaviours [McNeill, Lisa, and Rebecca Moore. 2015. “Sustainable Fashion Consumption and the Fast Fashion Conundrum: Fashionable Consumers and Attitudes to Sustainability in Clothing Choice.” International Journal of Consumer Studies 39 (3): 212–222], nor able to be ‘trusted to consistently make good decisions’ [Brooks, Andrew. 2015. Clothing Poverty: The Hidden World of Fast Fashion and Second-Hand Clothes. London: Zed Books, 241], they are also increasingly being called to recognise their individual role in the politics of global fashion supply chains. Drawing on examples from scholarly and popular discourses as well as online peer to peer communications, this article explores the historical moment of fast fashion as an instance of both the feminisation of consumption and the feminisation of responsibility.     

Terrorist violence and the enrollment of psychology in predicting Muslim extremism: critical terrorism studies meets critical algorithm studies

Discourse on terrorist violence has long facilitated an especially liberal form of securitisation. Originally evoked in reference to anarchists and communists, a rational consideration of terrorist violence, inaugurated by the concept, asks for deferred judgement about the nature of, or reasons behind, violence related to terror on the premise that state and international legal norms governing the legitimate use of violence fail to circumscribe the proper capacities of the state to regulate and explain terrorism. Where sovereign powers along with their military and civilian instruments of coercion are deemed unable to regulate violence effectively, analysts of terrorist violence and their readership are invited to consider and cultivate new sensibilities. Beginning in the 1980s, studies by psychologists found renewed urgency among a growing cadre of interdisciplinary terror experts who found religion, Islam especially, a key variable of analysis. I situate their contributions in a longer history of secular and racialising discourse about terrorist violence. Central to this history are practices of reading, translating, interpreting and archiving texts. Evidence for the argument is based on the analysis of an algorithm that allegedly predicts the likelihood of terrorist strikes by counting words spoken by al-Qa?ida leaders and correlating their frequency with over 30 psychological categories.     

Maximum likelihood estimation of structural VARFIMA models

This paper considers the maximum likelihood estimation of a class of structural vector autoregressive fractionally integrated moving-average (VARFIMA) models. The structural VARFIMA model includes the fractional cointegration model as one of its special cases. We show that the conditional likelihood Durbin–Levinson (CLDL) algorithm of Tsay (2010a) is a fast and reliable approach to estimate the long-run effects as well as the short- and long-term dynamics of a structural VARFIMA process simultaneously. In particular, the computational cost of the CLDL algorithm is much lower than that proposed in Sowell (1989) and Dueker and Startz (1998). We apply the CLDL method to the Congressional approval data of Durr et al. (1997) and find that the long-run effect of economic expectations on Congressional approval is at least 0.5718, which is over twice the estimate of 0.24 found in Table 2 of Box-Steffensmeier and Tomlinson (2000). This paper also tests the divided party government hypothesis with the CLDL algorithm.     

关于一类半无限多目标规划离散型算法的偏差估计

本文将文献[5]中的偏差估计公式推广到有任意多个目标函数和α∈(0, ∞)的情形并给出约束参数集维数可任意的数值例子.     

一种基于Otsu算法的印章二值化算法

本文提出了一种基于Otsu算法的印章二值化法.该算法通过对算法的改进,考虑了印章图象象素的灰度分布值和它们邻域象素的平均灰度值分布情况.并不是用于整幅印章图象,而是用于分块的印章图象.实验证明,该算法适合印章的二值化处理.     

论自动驾驶汽车生产者的刑事责任

自动驾驶汽车是人类研发、制造、使用和管理的智能产品,不是犯罪主体或刑事责任主体。在自动驾驶汽车自主控制状态下发生交通事故的,其生产者、使用者和其他人员难以按照我国现有刑法的罪名定罪处罚。除非道路交通安全法和刑法有专门的规定,驾驶位人员不接管汽车或接管后无力改变交通事故结果的,不构成交通肇事罪或其他管理过失犯罪。驾驶位人员的注意义务是阻止自动驾驶汽车自主控制下发生交通事故,其注意义务不应过高。允许的风险、紧急避险理论不能为自动驾驶汽车紧急路况处理算法的生产与应用提供合法、合理的解决方案,生产者遵守算法安全标准仅可以使生产行为合法化。鉴于现行刑法不适应自动驾驶汽车应用的特性,我国应当建立以生产者全程负责为中心的新刑事责任体系,使之在自动驾驶汽车生产和应用两个阶段承担安全管理责任,生产者拒不履行自动驾驶汽车应用安全管理义务且情节严重的,应当承担刑事责任。     

Objective Analysis of Impressed Chisel Toolmarks

Historical and recent challenges to the practice of comparative forensic examination have created a driving force for the formation of objective methods for toolmark identification. In this study , fifty sequentially manufactured chisels were used to create impression toolmarks in lead (500 toolmarks total). An algorithm previously used to statistically separate known matching and nonmatching striated screwdriver marks and quasi-striated plier marks was used to evaluate the chisel marks. Impression toolmarks, a more complex form of toolmark, pose a more difficult test for the algorithm that was originally designed for striated toolmarks. Results show in this instance that the algorithm can separate matching and nonmatching impression marks, providing further validation of the assumption that toolmarks are identifiably unique.     

An Accelerated Analytical Process for the Development of STR Profiles for Casework Samples

Significant efforts are being devoted to the development of methods enabling rapid generation of short tandem repeat (STR) profiles in order to reduce turnaround times for the delivery of human identification results from biological evidence. Some of the proposed solutions are still costly and low throughput. This study describes the optimization of an analytical process enabling the generation of complete STR profiles (single‐source or mixed profiles) for human identification in approximately 5 h. This accelerated process uses currently available reagents and standard laboratory equipment. It includes a 30‐min lysis step, a 27‐min DNA extraction using the Promega Maxwell^®16 System, DNA quantification in <1 h using the Qiagen Investigator^® Quantiplex HYres kit, fast amplification (<26 min) of the loci included in AmpF?STR^® Identifiler^®, and analysis of the profiles on the 3500‐series Genetic Analyzer. This combination of fast individual steps produces high‐quality profiling results and offers a cost‐effective alternative approach to rapid DNA analysis.     

多产品联合订货的分销系统利润优化研究

作者在文中研究了在价格折扣驱动下多产品分销系统的库存与订货策略问题.建立了分销系统中各成员的利润函数,给出了分销系统利润优化模型,提出了基于多参数级联编码遗传算法的求解策略.实例计算说明了算法的有效性和分销协作管理条件下联合订货策略的应用潜力。