The optimization of energy consumption, with consequent costs reduction, is one the main challenges in present and future smart grids. Of course, this has to occur keeping living comfort for end-user unchanged. In work, an approach based on mixed-integer linear programming paradigm, which able provide optimal solution terms tasks power consumption management renewable resources, developed. proposed algorithm yields task scheduling under dynamic electrical constraints, while simultaneously...

Fault diagnosis of electric motors is a fundamental task for production line testing, and it usually performed by experienced human operators. In the recent years, several methods have been proposed in literature detecting faults automatically. Deep neural networks successfully employed this task, but, up to authors' knowledge, they never used an unsupervised scenario. This paper proposes method diagnosing using novelty detection approach based on deep autoencoders. method, vibration signals...

A recursive algorithm for updating the coefficients of a neural network structure complex signals is presented. Various activation functions are considered and practical definition proposed. The method, associated to mean-square-error criterion, yields form conventional backpropagation algorithm.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Recent papers reporting CMOS RF building blocks have aroused great expectations for receivers using deep-submicron technologies. This paper examines the trend in scaling, order to establish required current levels and achievable performance different feature sizes, if robust, easily manufacturable designs are be implemented cellular applications. The boundary conditions (system-level constraints) such designs, terms of number trimmed untrimmed external components roles they play relaxing...

Research on Smart Grids has recently focused the energy monitoring issue, with objective to maximize user consumption awareness in building contexts one hand, and provide a detailed description of customer habits utilities other. One hottest topic this field is represented by Non-Intrusive Load Monitoring (NILM): it refers those techniques aimed at decomposing aggregated data acquired single point measurement into diverse profiles appliances operating electrical system under study. The focus...

This paper focuses on online learning procedures for locally recurrent neural nets with emphasis multilayer perceptron (MLP) infinite impulse response (IIR) synapses and its variations which include generalized output activation feedback networks (MLN). We propose a new gradient-based procedure called recursive backpropagation (RBP) whose version, causal (CRBP), has some advantages over other methods. CRBP includes as particular cases (BP), temporal BP, Back-Tsoi algorithm (1991) among...

Multilayer perceptrons (MLPs) with weight values restricted to powers of two or sums are introduced. In a digital implementation, these neural networks do not need multipliers but only shift registers when computing in forward mode, thus saving chip area and computation time. A learning procedure, based on backpropagation, is presented for such networks. This procedure requires full real arithmetic therefore must be performed offline. Some test cases presented, concerning MLPs hidden layers...

More than 60 years of biochemical and biophysical studies have accustomed us to think proteins as highly purified entities that act in isolation, more or less freely diffusing until they find their cognate partner bind to. While vitro experiments reproduce these conditions largely remain the only way investigate intrinsic properties molecules, this approach ignores an important factor: natural milieu , are surrounded by several other molecules different chemical nature, crowded environment...

A Speaker Localization algorithm based on Neural Networks for multi-room domestic scenarios is proposed in this paper. The approach fully data-driven and employs a Network fed by GCC-PHAT (Generalized Cross Correlation Phase Transform) Patterns, calculated means of the microphone signals, to determine speaker position room under analysis. In particular, we deal with case study, which acoustic scene each influenced sounds emitted other rooms. tested against home recorded DIRHA dataset,...

When the novel coronavirus disease SARS-CoV2 (COVID-19) was officially declared a pandemic by WHO in March 2020, scientific community had already braced up effort of making sense fast-growing wealth data gathered national authorities all over world. However, despite diversity theoretical approaches and comprehensiveness many widely established models, official figures that recount course outbreak still sketch largely elusive intimidating picture. Here we show unambiguously dynamics COVID-19...

In this paper, a new adaptive spline activation function neural network (ASNN) is presented. Due to the ASNN's high representation capabilities, networks with small number of interconnections can be trained solve both pattern recognition and data processing real-time problems. The main idea use Catmull-Rom cubic as neuron's function, which ensures simple structure suitable for software hardware implementation. Experimental results demonstrate improvements in terms generalization capability...

We introduce a topology-based nonlinear network model of protein dynamics with the aim investigating interplay spatial disorder and nonlinearity. show that spontaneous localization energy occurs generically is site-dependent process. Localized modes origin form spontaneously in stiffest parts structure display activation energies. Our results provide straightforward way for understanding recently discovered link between local stiffness enzymatic activity. They strongly suggest phenomena may...

Diffusion-limited reactions are usually described within the Smoluchowski theory, which neglects interparticle interactions. We propose a simple way to incorporate excluded-volume effects building on simulations of hard sphere in presence sink. For large values sink-to-particle size ratio ${R}_{s}$, measured encounter rate is good agreement with generalization equation at high densities. Reducing substantially depressed and becomes even nonmonotonic for ${R}_{s}\ensuremath{\ll}1$....

In this paper we introduce a fully flexible coarse-grained model of immunoglobulin G (IgG) antibodies parametrized directly on cryo-EM data and simulate the binding dynamics many IgGs to antigens adsorbed surface at increasing densities. Moreover, work out theoretical that allows explain all features observed in simulations. Our combined computational framework is excellent agreement with surface-plasmon resonance us establish number important results. (i) Internal flexibility key maximize...

Novelty detection is the task of recognising events differ from a model normality. This paper proposes an acoustic novelty detector based on neural networks trained with adversarial training strategy. The proposed approach composed feature extraction stage that calculates Log-Mel spectral features input signal. Then, autoencoder network, corpus "normal" signals, employed to detect whether segment contains abnormal event or not. A detected if Euclidean distance between and output exceeds...

Metallic nanoparticles have been used as catalysts for various reactions, and the huge literature on subject is hard to overlook. In many applications, must be affixed a colloidal carrier easy handling during catalysis. These "passive carriers" (e.g., dendrimers) serve controlled synthesis of prevent coagulation Recently, hybrids from polymers developed that allow us change catalytic activity by external triggers. particular, single embedded in thermosensitive network made...

It is well known that the behaviour of a neural network built with classical summing neurons, as in multilayer perceptron, widely depends on activation functions involved neurons. Many authors have proposed use some free parameters which should allow one to reduce size network, trading connection complexity function complexity. Since many implementations are based digital hardware, performing selected through lookup-table (LUT), it could be interesting study networks whose neurons adaptable...

It is known that RNS VLSI processors can parallelize fixed-point addition and multiplication operations by the use of Chinese remainder theorem (CRT). The required modular operations, however, must specialized hardware whose design implementation create several problems. In this paper a modified residue arithmetic, called pseudo-RNS introduced in order to alleviate some problems when digital signal processing (DSP) structures are implemented. Pseudo-RNS requires only binary exhibits speed...