Combating discrimination using Bayesian networks

Discrimination in decision making is prohibited on many attributes (religion, gender, etc…), but often present in historical decisions. Use of such discriminatory historical decision making as training data can perpetuate discrimination, even if the protected attributes are not directly present in the data. This work focuses on discovering discrimination in instances and preventing discrimination in classification. First, we propose a discrimination discovery method based on modeling the probability distribution of a class using Bayesian networks. This measures the effect of a protected attribute (e.g., gender) in a subset of the dataset using the estimated probability distribution (via a Bayesian network). Second, we propose a classification method that corrects for the discovered discrimination without using protected attributes in the decision process. We evaluate the discrimination discovery and discrimination prevention approaches on two different datasets. The empirical results show that a substantial amount of discrimination identified in instances is prevented in future decisions.     

Process optimisation and physicochemical characterisation of enzymatic hydrolysates of proteins from co‐products of Atlantic Salmon (Salmo salar) and Yellowtail Kingfish (Seriola lalandi)

The aim of this study was to develop an enzymatic hydrolysis process of protein co‐products for two major commercial fish species in Australia: Atlantic salmon (AS) and Yellowtail kingfish (YTK). The outcomes are to produce high protein recovery of fish protein hydrolysates within controlled molecular weight ranges that display enhanced physicochemical properties of oil binding and emulsification. Three enzymes (Flavourzyme, Neutrase and Alcalase) were applied to processing co‐products. Protein recovery and physicochemical properties were evaluated with increasing hydrolysis time from 30 min to 180 min and ratio of enzyme to substrate (E/S) from 0.5% to 3.0%. In order to achieve a product with optimum emulsifying capacity (50 ± 0.6 m² g^?1), an E/S ratio of 0.6–1.3% Flavourzyme was applied for 30–111 min with a protein recovery of 55%; in order to achieve a product with optimum oil‐binding capacity (8.3 ± 0.3 g oil g hydrolysates^?1), an E/S ratio of 2.3–3.0% Flavourzyme was applied for 25–64 min with a protein recovery of 70%. YTK protein hydrolysates were further membrane‐fractionated into five fractions (>100 kDa, 50–100 kDa, 30–50 kDa, 10–30 kDa and <10 kDa), and of these, the 10–30 kDa exhibited the best properties of oil binding (19 ± 0.3 g oil g hydrolysates^?1) and emulsification (57 ± 0.7 m² g^?1). These results demonstrate the importance of enzymatic hydrolysis of seafood co‐products into high‐value ingredients for food products and processing.     

Analysis of in situ monitored thermal cycling benefits for wireless packaging early reliability evaluation

With the increasing complexity of the power amplifier (PA) module architecture, the probability of a thermally induced stress related failure mechanism increases. To help evaluate the increase in module complexity, a more sophisticated in situ monitored thermal cycle reliability test is available. The module is monitored in real time using a resistance daisy chain methodology designed to provide coverage using resistance feedback throughout the entire hierarchy of the module and carrier board interface. Monitored temperature cycling allows for real time failure feedback and enhanced failure signature information. Further, the testing technique has proven to be a valuable method for capturing the early stages of a module mechanical failure at the temperature extremes. Moreover, statistical evaluation of the failure data (Weibull analysis) is improved and better accuracy of the failures in time (FIT) rate can be determined.     

男装高级成衣

正顺势随心正如James Brown的名曲I Feel Good所唱的那样,本季的男装"再创造"正以一股随心之势铺展开来:以层出不穷的新品和繁多的华丽秀场打头,随之发展成为无法抗拒的流行大潮。令人折服的设计、人性化的考量、与商业的完美融合外加保守内敛的气质共同打造出尖端的个性表达。高超的缝纫技艺凸显风格却不激进,运动和街头风共同打造出舒适的穿着感受。这一切都使得2015年春夏的男装充满了新鲜感和趣味性——对服装构造的全新解码以及蜕变中的年轻品牌和成熟品牌一起共同打造出简约却不平凡的夏日装束。     

男装高级成衣

一组商业化的精选系列,但是风格新颖。紧身造型趋于更加舒适并结合新廓形和有趣的比例变化。明亮的色调,图案与纹理组合预示着又一个有趣而质感十足的季节即将到来。尽管受许多市场区域持续不断的紧缩政策及消费者对未来经济形势的担忧等影响,但仍可以藉由色彩、版型和面料的新颖选择而虏获消费者的注意力和青睐。     

The impact of compulsory cycle helmet legislation on cyclist head injuries in New South Wales,Australia: A rejoinder

This paper challenges the conclusion of a recent paper by Walter et al. (Accident Analysis and Prevention 2011, doi:10.1016/j.aap.2011.05.029) reporting that despite numerous data limitations repealing the helmet legislation in Australia could not be justified. This conclusion is not warranted because of the limited time period used in their analysis and the lack of data beyond a few years before the introduction of legislation, the failure to adequately account for the effect of the phasing in of the legislation, the effect of the marked reduction in child cyclists, and the non-comparability of the pedestrian and cycling injuries and related lack of consideration of the severity of head injuries. The extent to which helmet legislation deters people from cycling is discussed.     

Evaluation of a Brix refractometer to estimate serum immunoglobulin G concentration in neonatal dairy calves

The objective of this study was to evaluate the utility of a digital Brix refractometer for the assessment of success of passive transfer of maternal immunoglobulin compared with the measurement of serum total protein (STP) by refractometry. Blood samples (n = 400) were collected from calves at 3 to 6d of age. Serum IgG concentration was determined by radial immunodiffusion (RID), and STP and percentage Brix (%Brix) were determined using a digital refractometer. The mean IgG concentration was 24.1 g/L [standard deviation (SD) ± 10.0] with a range from 2.1 to 59.1 g/L. The mean STP concentration was 6.0 g/dL (SD ± 0.8) with a range from 4.4 to 8.8 g/dL. The mean %Brix concentration was 9.2% (SD ± 0.9) with a range of 7.3 to 12.4%. Brix percentage was highly correlated with IgG (r = 0.93). Test characteristics were calculated to assess failure of passive transfer (FPT; serum IgG <10 g/L). The sensitivity and specificity of STP at 5.5 g/dL were 76.3 and 94.4%, respectively. A receiver operating characteristic curve was created to plot the true positive rate against the false positive rate for consecutive %Brix values. The optimal combination of sensitivity (88.9%) and specificity (88.9%) was at 8.4% Brix. Serum total protein was also positively correlated with %Brix (r = 1.00) and IgG (r = 0.93). Dairy producers can successfully monitor their colostrum management and the overall success of passive transfer using a digital Brix refractometer to estimate IgG concentration of colostrum and calf serum.     

Spark Plasma Sintering of Nanostructured Aluminum: Influence of Tooling Material on Microstructure

The influence of tooling material, i.e., graphite and WC-Co, on the microstructure of a spark plasma sintering (SPS) consolidated, nanostructured aluminum alloy is studied in this paper. The results show that tooling selection influences microstructure evolution, independent of process parameters. The influence of tooling on microstructure is rationalized on the basis of the following factors: heating rate, electrical current density, localized heating, and imposed pressure. A theoretic framework, based on the physical properties of graphite and WC-Co, is formulated to explain the observed behavior.     

Fast multi-scale detection of overlapping communities using local criteria

Many systems can be described using graphs, or networks. Detecting communities in these networks can provide information about the underlying structure and functioning of the original systems. Yet this detection is a complex task and a large amount of work was dedicated to it in the past decade. One important feature is that communities can be found at several scales, or levels of resolution, indicating several levels of organisation. Therefore a graph may have several community structures. Also networks tend to be large and hence require efficient processing. In this work, we present a new algorithm with linear complexity for the fast detection of communities across scales using a local criterion. We exploit the local aspect of the criterion to enable parallel computation and improve the execution speed further. The algorithm is tested against very large generated multi-scale networks and experiments demonstrate its efficiency and accuracy.