Optimising traceability in trade for live animals and animal products with digital technologies.

Sound animal traceability systems and supply chain management rely on data and information to respond to outcomes that will both protect animal and human health and facilitate trade. Digital technologies present opportunities and new methods for identifying and tracking animals, collecting more data, integrating communication flows, sharing data securely in supply chains, and analysing data to inform decisions and predict outcomes. Together, these technologies drive more efficient, productive and traceable supply chains, which can help to build more effective animal traceability systems. In addition, they can improve monitoring of, and response to, animal disease, food safety risks and food fraud risks; ensure compliance with animal health and food safety standards; simplify border procedures; facilitate trade with less friction; and raise consumer awareness. As the cost of these technologies decline and they become more accessible, the implementation of a digitally enabled animal traceability system will require an increase in supply chain capacity, improvements in digital infrastructure, and the development of a regulatory framework of standards and policies. Ensuring that these requirements are met will require strong commitment from governments, intergovernmental organisations and the wider animal health community.

Les systèmes fiables de traçabilité et de gestion des chaînes d'approvisionnement dépendent des données et de l'information pour élaborer des réponses permettant de protéger la santé publique et la santé animale tout en facilitant le commerce international. Les technologies numériques offrent de nouvelles perspectives et méthodologies pour identifier et tracer les animaux, collecter un plus grand volume de données, intégrer les flux de communication, partager les données de manière sécurisée tout au long des chaînes d'approvisionnement et analyser les données afin de prendre des décisions en connaissance de cause et d'anticiper leurs conséquences. Prises ensemble, ces technologies favorisent une meilleure efficience, productivité et traçabilité des chaînes d'approvisionnement, ce qui à son tour permet de concevoir des systèmes de traçabilité animale plus efficaces. En outre, elles renforcent les capacités de surveillance et de réponse en cas de maladies animales, de risques pour la sécurité sanitaire des aliments ou de fraude alimentaire ; leur utilisation contribue également à garantir la conformité aux normes de santé animale et de sécurité sanitaire des aliments, simplifie les procédures de contrôle aux frontières, facilite un commerce international moins conflictuel et participe à la sensibilisation des consommateurs. Ces technologies étant désormais moins onéreuses et plus faciles d'accès, les seules conditions pour mettre en oeuvre un système de traçabilité animale basé sur le numérique sont d'augmenter les capacités des chaînes d'approvisionnement, d'améliorer l'infrastructure numérique et de mettre en place un cadre réglementaire intégrant les normes et les politiques en la matière. Un engagement fort de la part des gouvernements, des organisations intergouvernementales et plus largement de la communauté de la santé animale sera nécessaire pour que ces conditions soient réunies.

La solidez de los sistemas de trazabilidad de animales y de gestión de las cadenas de suministro pasa por contar con datos e información cuya explotación dé lugar a una respuesta que a la vez proteja la salud animal y humana y facilite el comercio. Las tecnologías digitales abren posibilidades y traen consigo nuevos métodos para identificar y rastrear a los animales, obtener un mayor volumen de datos, integrar los circuitos de comunicación, compartir datos de forma segura en distintos eslabones de las cadenas de suministro y analizar estos datos para fundamentar decisiones y predecir los resultados. En conjunto, estas tecnologías dan lugar a cadenas de suministro más eficientes, productivas y fáciles de rastrear, lo que a su vez puede ayudar a instituir sistemas más eficaces de trazabilidad de animales. Además, pueden servir para: mejorar la vigilancia de las enfermedades animales y los riesgos de inocuidad de los alimentos y de fraude alimentario y la respuesta a tales enfermedades y riesgos; garantizar el cumplimiento de las normas de sanidad animal e inocuidad de los alimentos; simplificar los procedimientos aduaneros; facilitar un comercio más fluido; y sensibilizar a los consumidores. A medida que el costo de estas tecnologías vaya bajando y sea cada vez más fácil acceder a ellas, la implantación de un sistema digital de trazabilidad de animales exigirá una mayor capacidad de la cadena de suministro, una mejor infraestructura digital y la elaboración de un ordenamiento de normas y políticas que regule la cuestión. Para que todas estas condiciones se cumplan hará falta el firme empeño de gobiernos, organizaciones intergubernamentales y el conjunto de profesionales de la sanidad animal.

African swine fever: a global factor affecting agricultural markets over the medium term.

The African swine fever (ASF) outbreak in the People's Republic of China (China) is affecting regional and global meat and feed markets with potential impacts on vegetable oils, biofuels and even pharmaceuticals. Using the Aglink-Cosimo modelling system, the authors adopt three different scenarios to assess the impacts of ASF in China, South-East Asia and the world. The simulation results show a range of possible effects for agricultural commodity markets, notably a large initial protein gap that will be filled by higher production of both eggs and non-pork meats (poultry, beef and sheep/goat) in China and by pork imports from international markets. The results suggest a fast and near complete closure of the protein gap, reflecting China's low responsiveness of meat demand to prices. A sizeable share of the protein gap could remain unfilled if the necessary import infrastructure for meat, with gapless cold chains and efficient and comprehensive sanitary controls, is not set up as assumed in the scenarios. Not filling the protein gap would also leave domestic meat prices at permanently high levels, which could even translate into higher overall inflation rates. The simulations further suggest that an ASF pandemic would drive a lasting wedge between plant protein and animal protein prices, both locally and internationally. Oil meal prices will be particularly adversely affected, whereas pork and poultry will see a significant price rise. Countries that import the former and export the latter are likely to become the main beneficiaries of an ASF pandemic, benefiting from lower input prices and higher output prices for potentially large volumes of exports.

Le foyer de peste porcine africaine (PPA) survenu en République populaire de Chine (Chine) affecte les marchés régionaux et mondiaux de viande et d'aliments fourragers et a également un impact potentiel sur les huiles végétales, les bio-carburants voire les produits pharmaceutiques. En s'appuyant sur le système de modélisation Aglink-Cosimo, les auteurs ont élaboré trois scénarios différents pour évaluer les impacts de la PPA en Chine, en Asie du Sud-Est et dans le monde entier. Les résultats de la simulation font apparaître un large éventail d'effets sur les marchés de produits agricoles, en particulier un déficit protéique initial majeur qui sera compensé par une augmentation de la production d'oeufs et de viande issue d'autres animaux que le porc (volaille, boeuf, mouton et chèvre) en Chine, et par des importations de viande de porc à partir des marchés internationaux. Ces résultats semblent indiquer une élimination rapide et quasiment complète du déficit protéique, traduisant une faible réactivité aux prix de la demande chinoise en viande. Une part non négligeable du déficit protéique pourrait ne pas être comblée si les infrastructures nécessaires pour les importations de viande, avec des chaînes du froid ininterrompues et des inspections sanitaires complètes ne sont pas en place comme le prévoient les scénarios. Le maintien d'un déficit protéique aurait également pour effet de pousser les prix de la viande locale à des niveaux durablement élevés, ce qui pourrait se traduire par une hausse générale de l'inflation. Les simulations indiquent par ailleurs qu'une pandémie de PPA pourrait induire un écart de prix durable entre les protéines végétales et animales, aussi bien localement qu'au niveau international. Les prix des farines de tourteaux seraient particulièrement affectés tandis que les prix de la viande de porc et de volaille connaîtraient une hausse significative. Les pays importateurs de farines de tourteaux et exportateurs de viande de porc et de volaille seront probablement les principaux gagnants en cas de pandémie de PPA, bénéficiant de prix d'intrants bas et de prix d'extrants élevés sur des volumes potentiellement importants d'exportations.

El brote de peste porcina africana en la República Popular de China (China) está afectando a los mercados regionales y mundiales de carne y piensos y puede tener repercusión en los de aceites vegetales, biocombustibles e incluso productos farmacéuticos. Los autores aplican el sistema de modelización Aglink-Cosimo a tres hipótesis distintas para estudiar los efectos de la peste porcina africana a la escala de China, el Sudeste asiático y el mundo entero. Los resultados de la simulación revelan todo un conjunto de posibles efectos sobre los mercados de productos agrícolas, empezando por un gran déficit de proteínas que se cubrirá con una mayor producción tanto de huevos como de carnes no porcinas (avícola, vacuna y ovina/caprina) en China y con la importación de carne porcina a través de los mercados internacionales. Los resultados parecen indicar asimismo una rápida y cuasi completa corrección del déficit de proteínas, lo que pone de relieve que en China la demanda de carne es poco reactiva a los precios. Es posible que una parte considerable del déficit de proteínas quedara por cubrir en caso de que la infraestructura necesaria para las importaciones de carne, con cadenas de frío ininterrumpidas y controles sanitarios eficientes y exhaustivos, no correspondiera a los supuestos contemplados en las hipótesis. El hecho de no cubrir el déficit de proteínas también mantendría en niveles continuamente elevados los precios de la carne en el país, lo que podría incluso hacer subir las tasas de inflación generales. Las simulaciones llevan a pensar además que una pandemia de peste porcina africana induciría un desfase duradero entre los precios de las proteínas vegetales y los de las proteínas animales, tanto en el país como a nivel internacional. El precio de los piensos de aceite, en particular, se vería afectado negativamente, mientras que el de la carne porcina y la avícola sufrirían un sustancial aumento. Los países que importan los primeros y exportan las segundas serían presumiblemente los principales beneficiarios de una pandemia de peste porcina africana, pues saldrían ganando con el menor precio de los insumos y el mayor precio de los productos finales, que podrían exportar en grandes cantidades.

Salvage therapy with everolimus reduces the severity of treatment-refractory chronic GVHD without impairing disease control: a dual center retrospective analysis.

Chronic GVHD (cGVHD) remains the most important cause of late non-relapse mortality post allogeneic hematopoietic SCT (HSCT). Although first-line treatment of cGVHD with steroids is well established, evidence for second-line treatment remains limited. Here, we report a dual center retrospective analysis of the off-label salvage treatment of steroid-refractory cGVHD with everolimus. Out of 80 patients with a median age of 50 (17-70) years, 14 (17%) suffered from mild, 39 (49%) from moderate and 27 (34%) from severe cGVHD. At the final analysis, median follow-up after introduction of everolimus was 724 (14-2205) days. Thirty-four patients (43%) required the addition of further immunosuppression during everolimus-based therapy. Global NIH Severity Score improved in 34 patients (43%), remained stable in 37 patients (46%) and worsened in 9 patients (11%). The total sum of Global NIH Severity Scores in all patients assessable was significantly reduced after treatment with everolimus (P<0.0001). Most frequent grade 3/4 toxicities included infections (n=30) and thrombocytopenia (n=15). There was a single case of relapse. Everolimus-based salvage treatment of refractory cGVHD results in significant improvement of the NIH Severity Score without impairing control of the malignant disease. Finally, these preliminary results demand further verification in prospective trials.

Introductory lecture the epidemiology and determinants of obesity in developed and developing countries.

Obesity is recognized as a serious problem in the industrialized and developed countries of the world. However, little attention is paid to the fact that obesity is becoming an increasing problem in developing countries too, with some countries showing increasing rates of obesity in the midst of the persisting occurrence of childhood malnutrition and stunting. As developing countries embrace the dominant western economic ways of development, industrialization and urbanization they contribute to improvements in living standards, with consequent dramatic changes in diets and lifestyles leading to weight gain and obesity which in turn poses a growing threat to the health. Overweight and obesity is associated with an increased likelihood of non-insulin dependent diabetes mellitus, hypertension, hyper-lipidaemia, and cardiovascular disease. It is also associated with increased rates of breast, colo-rectal and uterine cancer. Obesity is thus an important factor in the increasing morbidity and mortality due to chronic, non-communicable diseases (NCDs) and thereby contributes to premature mortality in the population. Thus, while the problem of undernutrition persists in much of the developing world, overweight and obesity and its related co-morbidities are posing an increasingly important public health problem both in the developed and developing world.

Learning nonregular languages: a comparison of simple recurrent networks and LSTM.

In response to Rodriguez's recent article (2001), we compare the performance of simple recurrent nets and long short-term memory recurrent nets on context-free and context-sensitive languages.

Learning to forget: continual prediction with LSTM.

Long short-term memory (LSTM; Hochreiter & Schmidhuber, 1997) can solve numerous tasks not solvable by previous learning algorithms for recurrent neural networks (RNNs). We identify a weakness of LSTM networks processing continual input streams that are not a priori segmented into subsequences with explicitly marked ends at which the network's internal state could be reset. Without resets, the state may grow indefinitely and eventually cause the network to break down. Our remedy is a novel, adaptive "forget gate" that enables an LSTM cell to learn to reset itself at appropriate times, thus releasing internal resources. We review illustrative benchmark problems on which standard LSTM outperforms other RNN algorithms. All algorithms (including LSTM) fail to solve continual versions of these problems. LSTM with forget gates, however, easily solves them, and in an elegant way.

Processing images by semi-linear predictability minimization.

In the predictability minimization approach, input patterns are fed into a system consisting of adaptive, initially unstructured feature detectors. There are also adaptive predictors constantly trying to predict current feature detector outputs from other feature detector outputs. Simultaneously, however, the feature detectors try to become as unpredictable as possible, resulting in a co-evolution of predictors and feature detectors. This paper describes the implementation of a visual processing system trained by semi-linear predictability minimization, and presents many experiments that examine its response to artificial and real-world images. In particular, we observe that under a wide variety of conditions, predictability minimization results in the development of well-known visual feature detectors.

Feature extraction through LOCOCODE.

Low-complexity coding and decoding (LOCOCODE) is a novel approach to sensory coding and unsupervised learning. Unlike previous methods, it explicitly takes into account the information-theoretic complexity of the code generator. It computes lococodes that convey information about the input data and can be computed and decoded by low-complexity mappings. We implement LOCOCODE by training autoassociators with flat minimum search, a recent, general method for discovering low-complexity neural nets. It turns out that this approach can unmix an unknown number of independent data sources by extracting a minimal number of low-complexity features necessary for representing the data. Experiments show that unlike codes obtained with standard autoencoders, lococodes are based on feature detectors, never unstructured, usually sparse, and sometimes factorial or local (depending on statistical properties of the data). Although LOCOCODE is not explicitly designed to enforce sparse or factorial codes, it extracts optimal codes for difficult versions of the "bars" benchmark problem, whereas independent component analysis (ICA) and principal component analysis (PCA) do not. It produces familiar, biologically plausible feature detectors when applied to real-world images and codes with fewer bits per pixel than ICA and PCA. Unlike ICA, it does not need to know the number of independent sources. As a preprocessor for a vowel recognition benchmark problem, it sets the stage for excellent classification performance. Our results reveal an interesting, previously ignored connection between two important fields: regularizer research and ICA-related research. They may represent a first step toward unification of regularization and unsupervised learning.

Long short-term memory.

Learning to store information over extended time intervals by recurrent backpropagation takes a very long time, mostly because of insufficient, decaying error backflow. We briefly review Hochreiter's (1991) analysis of this problem, then address it by introducing a novel, efficient, gradient-based method called long short-term memory (LSTM). Truncating the gradient where this does not do harm, LSTM can learn to bridge minimal time lags in excess of 1000 discrete-time steps by enforcing constant error flow through constant error carousels within special units. Multiplicative gate units learn to open and close access to the constant error flow. LSTM is local in space and time; its computational complexity per time step and weight is O(1). Our experiments with artificial data involve local, distributed, real-valued, and noisy pattern representations. In comparisons with real-time recurrent learning, back propagation through time, recurrent cascade correlation, Elman nets, and neural sequence chunking, LSTM leads to many more successful runs, and learns much faster. LSTM also solves complex, artificial long-time-lag tasks that have never been solved by previous recurrent network algorithms.

Probabilistic incremental program evolution

Probabilistic incremental program evolution (PIPE) is a novel technique for automatic program synthesis. We combine probability vector coding of program instructions, population-based incremental learning, and tree-coded programs like those used in some variants of genetic programming (GP). PIPE iteratively generates successive populations of functional programs according to an adaptive probability distribution over all possible programs. Each iteration, it uses the best program to refine the distribution. Thus, it stochastically generates better and better programs. Since distribution refinements depend only on the best program of the current population, PIPE can evaluate program populations efficiently when the goal is to discover a program with minimal runtime. We compare PIPE to GP on a function regression problem and the 6-bit parity problem. We also use PIPE to solve tasks in partially observable mazes, where the best programs have minimal runtime.

Flat minima.

We present a new algorithm for finding low-complexity neural networks with high generalization capability. The algorithm searches for a "flat" minimum of the error function. A flat minimum is a large connected region in weight space where the error remains approximately constant. An MDL-based, Bayesian argument suggests that flat minima correspond to "simple" networks and low expected overfitting. The argument is based on a Gibbs algorithm variant and a novel way of splitting generalization error into underfitting and overfitting error. Unlike many previous approaches, ours does not require gaussian assumptions and does not depend on a "good" weight prior. Instead we have a prior over input-output functions, thus taking into account net architecture and training set. Although our algorithm requires the computation of second-order derivatives, it has backpropagation's order of complexity. Automatically, it effectively prunes units, weights, and input lines. Various experiments with feedforward and recurrent nets are described. In an application to stock market prediction, flat minimum search outperforms conventional backprop, weight decay, and "optimal brain surgeon/optimal brain damage".

Sequential neural text compression.

The purpose of this paper is to show that neural networks may be promising tools for data compression without loss of information. We combine predictive neural nets and statistical coding techniques to compress text files. We apply our methods to certain short newspaper articles and obtain compression ratios exceeding those of the widely used Lempel-Ziv algorithms (which build the basis of the UNIX functions "compress" and "gzip"). The main disadvantage of our methods is that they are about three orders of magnitude slower than standard methods.