The use of drug detection dogs in Sydney, Australia

Introduction and Aims. At present there is little research into the use of drug detection dogs. The present study sought to explore the use of detection dogs in Sydney, Australia, utilising multiple data sources.

Design and Methods. Data were taken from interviews with 100 regular ecstasy users and 20 key experts as part of the 2006 New South Wales arm of the Ecstasy and Related Drugs Reporting System, and secondary data sources.

Results. The majority of regular ecstasy users reported taking some form of precaution if made aware that dogs would be at an event they were attending. A small proportion of the sample reported consuming their drugs when coming into contact with detection dogs. One group of key experts viewed the use of detection dogs as useful; one group disliked the use of detection dogs though cooperated with law enforcement when dogs were used; and one group considered that detection dogs contribute to greater harm. Secondary data sources further suggested that the use of detection dogs do not significantly assist police in identifying and apprehending drug suppliers.

Discussion and Conclusions. The present study suggests that regular ecstasy users do not see detection dogs as an obstacle to their drug use. Future research is necessary to explore in greater depth the experiences that drug users have with detection dogs; the effect detection dogs may have on deterring drug consumption; whether encounters with detection dogs contribute to drug-related harm; and the cost–benefit analysis of this law enforcement exercise.

Fast one-class support vector machine for novelty detection

Novelty detection arises as an important learning task in several applications. Kernel-based approach to novelty detection has been widely used due to its theoretical rigor and elegance of geometric interpretation. However, computational complexity is a major obstacle in this approach. In this paper, leveraging on the cutting-plane framework with the well-known One-Class Support Vector Machine, we present a new solution that can scale up seamlessly with data. The first solution is exact and linear when viewed through the cutting-plane; the second employed a sampling strategy that remarkably has a constant computational complexity defined relatively to the probability of approximation accuracy. Several datasets are benchmarked to demonstrate the credibility of our framework.

Estimation of the diagnostic accuracy of the glyco-C and US9 gene-based polymerase chain reaction technique for the detection of bovine Herpesvirus type 5 DNA in decomposed brain suspension from a slaughter house using Bayesian analysis, Brazil

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Real-time quantitative broad-range PCR assay for detection of the 16S rRNA gene followed by sequencing for species identification

Here we determined the analytical sensitivities of broad-range real-time PCR-based assays employing one of three different genomic DNA extraction protocols in combination with one of three different primer pairs targeting the 16S rRNA gene to detect a panel of 22 bacterial species. DNA extraction protocol III, using lysozyme, lysostaphin, and proteinase K, followed by PCR with the primer pair Bak11W/Bak2, giving amplicons of 796 bp in length, showed the best overall sensitivity, detecting DNA of 82% of the strains investigated at concentrations of < or =10(2) CFU in water per reaction. DNA extraction protocols I and II, using less enzyme treatment, combined with other primer pairs giving shorter amplicons of 466 bp and 342 or 346 bp, respectively, were slightly more sensitive for the detection of gram-negative but less sensitive for the detection of gram-positive bacteria. The obstacle of detecting background DNA in blood samples spiked with bacteria was circumvented by introducing a broad-range hybridization probe, and this preserved the minimal detection limits observed in samples devoid of blood. Finally, sequencing of the amplicons generated using the primer pair Bak11W/Bak2 allowed species identification of the detected bacterial DNA. Thus, broad-spectrum PCR targeting the 16S rRNA gene in the quantitative real-time format can achieve an analytical sensitivity of 1 to 10 CFU per reaction in water, avoid detection of background DNA with the introduction of a broad-range probe, and generate amplicons that allow species identification of the detected bacterial DNA by sequencing. These prerequisites are important for its application to blood-containing patient samples.

High-dimensional and large-scale anomaly detection using a linear one-class SVM with deep learning

High-dimensional problem domains pose significant challenges for anomaly detection. The presence of irrelevant features can conceal the presence of anomalies. This problem, known as the '. curse of dimensionality', is an obstacle for many anomaly detection techniques. Building a robust anomaly detection model for use in high-dimensional spaces requires the combination of an unsupervised feature extractor and an anomaly detector. While one-class support vector machines are effective at producing decision surfaces from well-behaved feature vectors, they can be inefficient at modelling the variation in large, high-dimensional datasets. Architectures such as deep belief networks (DBNs) are a promising technique for learning robust features. We present a hybrid model where an unsupervised DBN is trained to extract generic underlying features, and a one-class SVM is trained from the features learned by the DBN. Since a linear kernel can be substituted for nonlinear ones in our hybrid model without loss of accuracy, our model is scalable and computationally efficient. The experimental results show that our proposed model yields comparable anomaly detection performance with a deep autoencoder, while reducing its training and testing time by a factor of 3 and 1000, respectively.