PyDTNN: A user-friendly and extensible framework for distributed deep learning

The Journal of Supercomputing - We introduce a framework for training deep neural networks on clusters of computers with the following appealing properties: (1) It is developed in Python, exposing...     

Spatial Optimization Models for Water Supply Allocation

Climate change is likely to result in increased aridity, lower runoff, and declining water supplies for the cities of the Southwestern United States, including Phoenix. The situation in Phoenix is particularly complicated by the large number of water providers, each with its own supply portfolio, demand conditions, and conservation strategies. This paper details spatial optimization models to support water supply allocation between service provider districts, where some districts experience deficits and others experience surpluses in certain years. The approach seeks to reconcile and integrate projections derived from a complex simulation model taking into account current and future climate conditions. The formulated and applied models are designed to help better understand the expected increasingly complex interactions of providers under conditions of climate change. Preliminary results show cooperative agreements would reduce spot shortages that would occur even without climate change. In addition, they would substantially reduce deficits if climate change were to moderately reduce river flows in Phoenix’s major source regions, but have little effect under the most pessimistic scenarios because there are few surpluses available for re-allocation.     

Real-time and low-power processing of 3D direct/inverse discrete cosine transform for low-complexity video codec

The 3D discrete cosine transform and its inverse (3D DCT/IDCT) extend the spatial compression properties of conventional 2D DCT to the spatio-temporal coding of 2D videos. The 3D DCT/IDCT transform is particularly suited for embedded systems needing the low-complexity implementation of both video encoder and decoder, such as mobile terminals with video-communication capabilities. This paper addresses the problem of real-time and low-power 3D DCT/IDCT processing by presenting a context-aware fast transform algorithm and a family of VLSI architectures characterized by different levels of parallelism. Implemented in submicron CMOS technology, the proposed hardware macrocells support the real-time processing of main video formats (up to high definition ones with an input rate of tens of Mpixels/s) with different trade-offs between circuit complexity, power consumption and computational throughput. Voltage scaling and adaptive clock-gating strategies are applied to reduce the power consumption versus the state of the art.     

Evolving neuro-fuzzy network for real-time high impedance fault detection and classification

Neural Computing and Applications - This paper concerns the application of a neuro-fuzzy learning method based on data streams for high impedance fault (HIF) detection in medium-voltage power...     

Anomaly detection with a moving camera using multiscale video analysis

Multidimensional Systems and Signal Processing - This paper addresses the problem of abandoned object detection in a cluttered environment using a camera moving along a straight track. The...     

On an New Algorithm for Function Approximation with Full Accuracy in the Presence of Discontinuities Based on the Immersed Interface Method

This paper is devoted to the construction and analysis of an adapted and nonlinear multiresolution algorithm designed for interpolation or approximation of discontinuous univariate functions. The adaption attained allows to avoid numerical artifacts that appear when using linear algorithms and, at the same time, to obtain a high order of accuracy close to the singularities. It is known that linear algorithms are stable and convergent for smooth functions, but diffusion and Gibbs effect appear if the functions are piecewise continuous. Our aim is to develop an algorithm for function approximation with full accuracy that is capable to adapt to corners (kinks) and jump discontinuities, that uses a centered stencil and that does not use extrapolation. In order to reach this goal, we will need some information about the jumps in the function that we want to approximate and its derivatives. If this information is available, the algorithm is the most compact possible in the sense that the stencil is fixed and we do not need a stencil selection procedure as other algorithms do, such as ENO subcell resolution (ENO-SR). If the information about the jumps is not available, we will show a technique to approximate it. The algorithm is based on linear interpolation plus correction terms that provide the desired accuracy close to corners or jump discontinuities.     

On the structure and surface properties of NiO/MgO–La2O3 catalyst: Influence of the support composition and preparation method

This work addresses the effect of catalyst preparation method and the carrier compositions (MgO–La₂O₃) over the NiO-support interaction, which affect the reducibility, textural properties and the different oxygen species chemisorbed at different temperatures over MgO–La₂O₃ supported NiO catalysts. The materials were prepared by wet sequential impregnation and wet co-impregnation with different Mg molar fractions [Mg/(La+Mg)]. The samples were characterized by X-ray diffraction (XRD), temperature-programmed reduction (TPR), infrared (IR) spectroscopy, scanning electron microscopy (SEM), changes of surface potential and BET surface area measurements. The total oxidation of methane was use as model reaction. It has been found that the catalyst formulations (i.e. NiO/MgO–La₂O₃) and the preparation methods not only affect the interaction among the catalyst components, but also the texture and material morphology as a result of different degrees of particle aggregation. The wet sequential impregnation-prepared catalysts showed a stronger MgO–La₂O₃ interaction than wet co-impregnation-prepared samples. A marked tendency of NiO to react with MgO rather than La₂O₃ following a mechanism of lattice substitution is observed. Mg-free catalyst showed LaNiO₃ and NiO as major crystalline Ni-containing phases. The ternary Ni–Mg–La–O system, on the other hand, facilitates the formation of poorly reducible Ni phase, whereas the La-free catalyst (i.e. NiO/MgO) displayed the lowest content of Ni-reducible phase, owing to the formation of Ni_1−xMg_xO solid solution. Measurements of surface potential changes together with catalytic studies suggest that La-containing catalysts present oxygen vacancies, which markedly affect the chemical nature of the surface oxygen species and hence their catalytic behaviour.     

Epicuticle Lipids Mediate Mate Recognition in <Emphasis Type="Italic">Triatoma infestans</Emphasis>

Epicuticular lipids are contact cues in intraspecific chemical communication in insects, both for aggregation and sexual behavior. Triatomine bugs are vectors of the parasite Trypanosoma cruzi, the cause of Chagas disease. In Triatoma infestans, the major epicuticular lipids are hydrocarbons, fatty alcohols, and free and esterified fatty acids. Previously, we found that epicuticular lipid extracts, or selected fatty acid components, trigger aggregation and arrestment behavior in this bug. Using headspace solid phase microextraction, we found no sexual dimorphism in epicuticular hydrocarbons, but found female-specific fatty alcohols (eicosanol and docosanol). The role of epicuticular lipids in T. infestans copulation behavior was tested by observing male responses to live or various treatments of freeze-killed females. We report that hexane-soluble contact cues on females trigger copulation by males. Freeze-killed intact females were attractive to males, but no response was observed when males were exposed to hexane-washed females. Responses were partially recovered when epicuticular extract was applied to the dorsal surface of dead, hexane-washed females. One female equivalent of docosanol, evoked similar responses.     

Revisiting with updated hardware an old spectroscopic technique: circularly polarized luminescence

Among the various chiroptical spectroscopic techniques available today, circularly polarized luminescence (CPL) plays a minor role and is still used by a limited number of specialists. The cost of the few commercial instruments available and the complexity of homemade apparatuses have strongly limited the widespread application of this technique. New technological approaches, such as the use of light-emitting diode (LED) sources, may significantly simplify the required instrumentation and encourage new potential users. Calibration procedure and the quantitative determination of CPL and total luminescence intensity as well as the corresponding g ratio are described.