Set Of Multilayer Perceptrons Research Articles

Estimation of electrical resistivity using artificial neural networks: a case study from Lublin Basin, SE Poland

Artificial neural networks method (ANNs) is a common estimation tool used for geophysical applications. Considering borehole data, when the need arises to supplement a missing well log interval or whole logging—ANNs provide a reliable solution. Supervised training of the network on a reliable set of borehole data values with further application of this network on unknown wells allows creation of synthetic values of missing geophysical parameters, e.g., resistivity. The main assumptions for boreholes are: representation of similar geological conditions and the use of similar techniques of well data collection. In the analyzed case, a set of Multilayer Perceptrons were trained on five separate chronostratigraphic intervals of borehole, considered as training data. The task was to predict missing deep laterolog (LLD) logging in a borehole representing the same sequence of layers within the Lublin Basin area. Correlation between well logs data exceeded 0.8. Subsequently, magnetotelluric parametric soundings were modeled and inverted on both boreholes. Analysis showed that congenial Occam 1D models had better fitting of TM mode of MT data in each case. Ipso facto, synthetic LLD log could be considered as a basis for geophysical and geological interpretation. ANNs provided solution for supplementing datasets based on this analytical approach.

On species distribution modelling, spatial scales and environmental flow assessment with Multi–Layer Perceptron Ensembles: A case study on the redfin barbel (Barbus haasi; Mertens, 1925)

Inconsistent performance of Species Distribution Models (SDMs), which may depend on several factors such as the initial conditions or the applied modelling technique, is one of the greatest challenges in ecological modelling. To overcome this problem, ensemble modelling combines the forecasts of several individual models. A commonly applied ensemble modelling technique is the Multi–Layer Perceptron (MLP) Ensemble, which was envisaged in the 1990s. However, despite its potential for ecological modelling, it has received little attention in the development of SDMs for freshwater fish. Although this approach originally included all the developed MLPs, Genetic Algorithms (GA) now allow selection of the optimal subset of MLPs and thus substantial improvement of model performance. In this study, MLP Ensembles were used to develop SDMs for the redfin barbel ( Barbus haasi ; Mertens, 1925) at two different spatial scales: the micro–scale and the meso–scale. Finally, the potential of the MLP Ensembles for environmental flow (e–flow) assessment was tested by linking model results to hydraulic simulation. MLP Ensembles with a candidate selection based on GA outperformed the optimal single MLP or the ensemble of the whole set of MLPs. The micro–scale model complemented previous studies, showing high suitability of relatively deep areas with coarse substrate and corroborating the need for cover and the rheophilic nature of the redfin barbel. The meso–scale model highlighted the advantages of using cross–scale variables, since elevation (a macro–scale variable) was selected in the optimal model. Although the meso–scale model also demonstrated that redfin barbel selects deep areas, it partially contradicted the micro–scale model because velocity had a clearer positive effect on habitat suitability and redfin barbel showed a preference for fine substrate in the meso–scale model. Although the meso–scale model suggested an overall higher habitat suitability of the test site, this did not result in a notable higher minimum environmental flow. Our results demonstrate that MLP Ensembles are a promising tool in the development of SDMs for freshwater fish species and proficient in e–flow assessment.

Binaural Detection, Localization, and Segregation in Reverberant Environments Based on Joint Pitch and Azimuth Cues

We propose an approach to binaural detection, localization and segregation of speech based on pitch and azimuth cues. We formulate the problem as a search through a multisource state space across time, where each multisource state encodes the number of active sources, and the azimuth and pitch of each active source. A set of multilayer perceptrons are trained to assign time-frequency units to one of the active sources in each multisource state based jointly on observed pitch and azimuth cues. We develop a novel hidden Markov model framework to estimate the most probable path through the multisource state space. An estimated state path encodes a solution to the detection, localization, pitch estimation and simultaneous organization problems. Segregation is then achieved with an azimuth-based sequential organization stage. We demonstrate that the proposed framework improves segregation relative to several two-microphone comparison systems that are based solely on azimuth cues. Performance gains are consistent across a variety of reverberant conditions.

A spectral clustering approach to underdetermined postnonlinear blind source separation of sparse sources

This letter proposes a clustering-based approach for solving the underdetermined (i.e., fewer mixtures than sources) postnonlinear blind source separation (PNL BSS) problem when the sources are sparse. Although various algorithms exist for the underdetermined BSS problem for sparse sources, as well as for the PNL BSS problem with as many mixtures as sources, the nonlinear problem in an underdetermined scenario has not been satisfactorily solved yet. The method proposed in this letter aims at inverting the different nonlinearities, thus reducing the problem to linear underdetermined BSS. To this end, first a spectral clustering technique is applied that clusters the mixture samples into different sets corresponding to the different sources. Then, the inverse nonlinearities are estimated using a set of multilayer perceptrons (MLPs) that are trained by minimizing a specifically designed cost function. Finally, transforming each mixture by its corresponding inverse nonlinearity results in a linear underdetermined BSS problem, which can be solved using any of the existing methods.

Hybrid knowledge-based system/multilayer perceptron approach for the post-assembly tuning of electronic filters

A generic approach to the development of a hybrid knowledge-based system (KBS)-multilayer perceptron (MLP) system for the post-assembly tuning of filters is described. This KBS uses rules derived using the Iterative Dichotomiser 3 to advise upon which filter component to adjust and in which direction. The problem is divided into a number of search spaces corresponding to three levels of decision, and numerical attribute values are represented by numerical ranges with logical names. How far to turn is determined by a set of MLPs, trained on the amplitude responses of a typical filter. It is established that the task requires a hybrid system and cannot be satisfactorily performed by either a KBS or MLP alone. When applied to the amplitude response tuning of four-pole asymmetric bandpass crystal filters three adjustments are typically required, on average, to tune the stopbands of 79% of those tested. The lengthy development times needed per filter type of 1-2 weeks are compensated for as the task is within the capabilities of a trained operator.

Context-dependent connectionist probability estimation in a hybrid hidden Markov model-neural net speech recognition system

In this paper we present a training method and a network architecture for estimating context-dependent observation probabilities in the framework of a hybrid hidden Markov model (HMM)/multi layer perceptron (MLP) speaker-independent continuous speech recognition system. The context-dependent modeling approach we present here computes the HMM context-dependent observation probabilities using a Bayesian factorization in terms of context-conditioned posterior phone probabilities which are computed with a set of MLPs, one for every relevant context. The proposed network architecture shares the input-to-hidden layer among the set of context dependent MLPs in order to reduce the number of independent parameters. Multiple states for phone models with different context dependence for each state are used to model the different context effects at the beginning and end of phonetic segments. A new training procedure that "smooths" networks with different degrees of context dependence is proposed to obtain a robust estimate of the context-dependent probabilities. We have used this new architecture to model generalized biphone phonetic contexts. Tests with the speaker-independent DARPA Resource Management database have shown average reductions in word error rates of 28% using a word-pair grammar, compared to our earlier context-independent HMM/MLP hybrid.