Variational Information Bottleneck for Unsupervised Clustering: Deep Gaussian Mixture Embedding.

In this paper, we develop an unsupervised generative clustering framework that combines the variational information bottleneck and the Gaussian mixture model. Specifically, in our approach, we use the variational information bottleneck method and model the latent space as a mixture of Gaussians. We derive a bound on the cost function of our model that generalizes the Evidence Lower Bound (ELBO) and provide a variational inference type algorithm that allows computing it. In the algorithm, the coders' mappings are parametrized using neural networks, and the bound is approximated by Markov sampling and optimized with stochastic gradient descent. Numerical results on real datasets are provided to support the efficiency of our method.

Status of microbial based cleaning products in statutory regulations and ecolabelling in Europe, the USA, and Canada.

Cleaning products containing living microorganisms as active ingredients are increasingly being used in household, professional and industrial cleaning applications. Microorganisms can degrade soiling associated with dirt, food residues, and grease by enzymatic and metabolic action and out-compete microorganisms associated with odor problems. Their potential for odor control seems to result in a competitive advantage over conventional chemically-based cleaning products. Moreover, producers of microbial-based cleaning products (MBCPs) claim that their products are less harmful to the environment. These promising prospects have triggered interest from consumer and environmental organizations, professional users, and regulators in understanding if there are also possible negative health and environmental impacts which require attention and how the safety of these products is ensured. Unfortunately, there is little information on these issues in the public domain. Moreover, regulatory oversight in Europe is essentially limited to pathogenic properties in the context of worker protection. Canada, in contrast, has a regulatory framework in place to assess risks to human health and the environment from the manufacture, import and/or use of new microorganisms contained in MBCPs. In the absence of mandatory standards, safety assessment and hygienic practices seem to vary considerably across companies. Recently developed ecolabelling standards are - for the time being - the only option for transparent compliance to minimum standards in terms of safety as well as assessments of manufacturer information by third parties. These standards highlight in particular the need for precise taxonomic information for assessing the pathogenic properties and the need to ensure the absence of potentially harmful microorganisms as contaminants. Ecolabelling standards are, however, voluntary and do not cover all relevant safety issues. In order to develop a more comprehensive set of mandatory standards for health and safety, a number of areas would benefit from further research (e.g. the role in plant pathogenicity and other environmental properties of the microorganisms used; the relevance of chronic exposure to dusts and aerosols containing vegetative cells and spores; the relevance of strains which belong to species known to include opportunistic pathogens and possible hazards for particularly sensitive risk groups). Improved knowledge in these areas will contribute to a predictable level of product safety.

Development and use of microbial-based cleaning products (MBCPs): Current issues and knowledge gaps.

Cleaning products containing microbes as active ingredients are becoming increasingly prevalent as an alternative to chemical-based cleaning products. These microbial-based cleaning products (MBCPs) are being used in domestic and commercial settings (i.e., households and businesses) and institutional settings (e.g., hospitals, schools, etc.), in a variety of cleaning activities (hard surface cleaning, odour control, degreasing, septic tank treatments, etc.). They are typically described as "environmentally friendly" and "non-toxic". Publicly available information sources (scientific literature, patent databases, commercial websites) were searched for information on microbial species contained in MBCPs, their mode of action, cleaning applications in which they are used, and their potential impacts on human health and the environment. Although information was found providing a broad indication of microbial genera/species used, information on specific species/strains and quantities produced and sold is generally lacking. This makes it difficult to conduct a meaningful examination of any risks to human health and the environment from the production and use of MBCPs and to determine how effective current policies and regulatory frameworks are in addressing these issues. These and other challenges were addressed at an international workshop in Ottawa, Canada in June 2013 by a number of stakeholders, including industry, government, academic and non-governmental organizations.

Review of current methods for characterizing virulence and pathogenicity potential of industrial Saccharomyces cerevisiae strains towards humans.

Most industrial Saccharomyces cerevisiae strains used in food or biotechnology processes are benign. However, reports of S. cerevisiae infections have emerged and novel strains continue to be developed. In order to develop recommendations for the human health risk assessment of S. cerevisiae strains, we conducted a literature review of current methods used to characterize their pathogenic potential and evaluated their relevance towards risk assessment. These studies revealed that expression of virulence traits in S. cerevisiae is complex and depends on many factors. Given the opportunistic nature of this organism, an approach using multiple lines of evidence is likely necessary for the reasonable prediction of the pathogenic potential of a particular strain. Risk assessment of S. cerevisiae strains would benefit from more research towards the comparison of virulent and non-virulent strains in order to better understand those genotypic and phenotypic traits most likely to be associated with pathogenicity.

Animal models to detect allergenicity to foods and genetically modified products: workshop summary.

Respiratory allergy and allergy to foods continue to be important health issues. There is evidence to indicate that the incidence of food allergy around the world is on the rise. Current estimates indicate that approximately 5% of young children and 1-2% of adults suffer from true food allergy (Kagan 2003). Although a large number of in vivo and in vitro tests exist for the clinical diagnosis of allergy in humans, we lack validated animal models of allergenicity. This deficiency creates serious problems for regulatory agencies and industries that must define the potential allergenicity of foods before marketing. The emergence of several biotechnologically derived foods and industrial proteins, as well as their potential to sensitize genetically predisposed populations to develop allergy, has prompted health officials and regulatory agencies around the world to seek approaches and methodologies to screen novel proteins for allergenicity.