In this paper, the estimation of parameters for a three-parameter Weibull distribution based on progressively Type-II right censored sample is studied. Different procedures complete are generalized to case with data. These methods include maximum likelihood estimators (MLEs), corrected MLEs, weighted product spacing and least squares estimators. We also proposed use method one-step bias-correction obtain reliable initial estimates iterative procedures. compared via Monte Carlo simulation...

From the point of information theory, baseband binary pulse amplitude modulated (PAM) signals transmission, via tuning nonlinear receiver's parameters, is studied over an additive white Gaussian noise (AWGN) channel. It demonstrated that channel capacity communication systems, for a given signal added noise, can be maximized by optimal designed receivers. This new form stochastic resonance (SR) referred to as parameter-induced (PSR) in broad sense. PSR effect does not require input are...

Two methods of realizing aperiodic stochastic resonance (ASR) by adding noise and tuning system parameters in a bistable system, after scale transformation, can be compared real parameter space. In this space, the point ASR via denotes extremum line segment, whereas method presents extrema plane. We demonstrate that, terms performance, takes precedence approach for an adjustable system. Besides, viewed as specific case parameters. Further research shows that optimal found may subthreshold or...

Abstract From the perspective of information transmission and visualization in deep neural networks, bottleneck theory offers a method to measure learning representations within networks. This thereby provides valuable framework for understanding benefits noise In this paper, we derive noise-boosted activation function based on suprathreshold stochastic resonance model analyze its characteristics. Our findings demonstrate that mutual between input output exhibits resonant behavior,...

We analyze the parametric estimation that can be performed on a signal buried in noise based parsimonious representation provided by parallel array of threshold devices. The Fisher information contained output about input parameter is used as measure performance task. For suprathreshold signal, we establish enhancement obtained addition independent noises to thresholds array. Similar improvement also shown possible for error maximum likelihood estimator. These results extend applicability...

In this paper, we investigate the first exploitation of vibrational resonance (VR) effect to detect weak signals in presence strong background noise. By injecting a series sinusoidal interference same amplitude but with different frequencies into generalized correlation detector, show that detection probability can be maximized at an appropriate amplitude. Based on dual-Dirac density model, compare VR method stochastic approach via adding dichotomous The compared results indicate achieve...

In this paper, we investigate the noise benefits to maximum likelihood type estimators (M-estimator) for robust estimation of a location parameter. Two distinct are shown be accessible under these conditions. With symmetric heavy-tailed distributions, asymptotic efficiency can enhanced by injecting extra into M-estimators. an asymmetric contaminated model having convex cumulative distribution function, demonstrate that addition reduce bias median estimator. These findings extend analysis...

This paper investigates the benefits of intentionally adding noise to a Bayesian estimator, which comprises linear combination number individual estimators that are perturbed by mutually independent sources and multiplied set adjustable weighting coefficients. We prove Bayes risk for mean square error (MSE) criterion is minimized when same optimum coefficients assigned identical in combiner. property leads simplified analysis benefit MSE combined estimator even tends infinity. It shown that,...

Remaining useful life (RUL) prediction methods for rotating machines have been successfully developed in recent decades. More attention should be paid to predictions with inconsistent data distributions under different conditions. To solve this problem, article proposes a new RUL method that includes two phases. In the first phase, degradation features are extracted from both training and testing sets using probabilistic principal component analysis (PPCA). second additive white Gaussian...

Various situations where a signal is enhanced by noise through stochastic resonance are now known. This paper contributes to determining general conditions under which improvement can be priori decided as feasible or not. We focus on the detection of known in additive white noise. Under assumptions weak and sufficiently large sample size, it proved, with an inequality based Fisher information, that adding never possible, generically, these conditions. However, less restrictive conditions,...

Using a gradient-based algorithm, we investigate signal estimation and filtering in large-scale summing network of single-bit quantizers. Besides adjusting weights, the proposed learning algorithm also adaptively updates level added noise components that are intentionally injected into Experimental results show minimization mean-squared error requires nonzero optimal noise. The process achieves this way form stochastic resonance or noise-aided processing. This adaptive optimization method...

Aiming to ensure the feasibility of backpropagation training feedforward threshold neural networks, each hidden unit layer is designed be composed a sufficiently large number hard-limiting activation functions that are excited by mutually independent external noise components and weighted inputs simultaneously. The application nondifferentiable enables proper definition gradients, injected treated as network parameter can adaptively updated stochastic gradient descent learning rule. This...

We analyze signal detection with nonlinear test statistics in the presence of colored noise. In limits small and weak noise correlation, optimal statistic its performance are derived under general conditions, especially concerning type also analyze, for a threshold nonlinearity–a key component neural model, conditions noise-enhanced performance, establishing that is superior to white detection. For parallel array elements, approximating neurons, we demonstrate even broader allowing...