Search results for: hidden layers

Authors: Slimane Ouhmad, Abdellah Halimi

Abstract:

In this work, we have applied neural networks method MLP type to a database from an array of six sensors for the detection of three toxic gases. As the choice of the number of hidden layers and the weight values has a great influence on the convergence of the learning algorithm, we proposed, in this article, a mathematical formulation to determine the optimal number of hidden layers and good weight values based on the method of back propagation of errors. The results of this modeling have improved discrimination of these gases on the one hand, and optimize the computation time on the other hand, the comparison to other results achieved in this case.

Keywords: MLP Neural Network, back-propagation, number of neurons in the hidden layer, identification, computing time

Procedia PDF Downloads 320

Authors: Zhifeng Kong

Abstract:

Over-parameterized neural networks have attracted a great deal of attention in recent deep learning theory research, as they challenge the classic perspective of over-fitting when the model has excessive parameters and have gained empirical success in various settings. While a number of theoretical works have been presented to demystify properties of such models, the convergence properties of such models are still far from being thoroughly understood. In this work, we study the convergence properties of training two-hidden-layer partially over-parameterized fully connected networks with the Rectified Linear Unit activation via gradient descent. To our knowledge, this is the first theoretical work to understand convergence properties of deep over-parameterized networks without the equally-wide-hidden-layer assumption and other unrealistic assumptions. We provide a probabilistic lower bound of the widths of hidden layers and proved linear convergence rate of gradient descent. We also conducted experiments on synthetic and real-world datasets to validate our theory.

Keywords: over-parameterization, rectified linear units ReLU, convergence, gradient descent, neural networks

Procedia PDF Downloads 120

Authors: Vibhushan Subba

Abstract:

B-movies are almost always ‘glanced over’, ‘swept beneath’, ‘hidden from’ and ‘locked away’ to live a secret life; a life that exists but enjoys only a mummified existence behind layers of protective covering. They are more often than not discarded as ‘trash’, ‘sleaze’, ‘porn’ and put down for their ‘bad taste’ or at least that has been the case in India. With the art film entering the realm of high art, the popular and the mainstream has been increasingly equated with the A grade Bollywood film. This leaves the B-movie to survive as a degraded cultural artifact on the fringes of the mainstream. Kanti Shah’s films are part of a secret, traversing the libidinal circuits of the B and C grade through history. His films still circulate like a corporeal reminder of the forbidden and that which is taboo, like a hidden fracture that threatens to split open bourgeois respectability. Seeking to find answers to an aesthetic that has been rejected and hidden, this paper looks at three films of Kanti Shah to see how the notion of taboo, censorship and the unseen coincide, how they operate in the domain of his cinema and try and understand a form that draws our attention to the subterranean forces at work.

Keywords: B-movies, trash, taboo, censorship

Procedia PDF Downloads 433

Authors: Joselito Medina-Marin, Maria G. Serna-Diaz, Juan C. Seck-Tuoh-Mora, Norberto Hernandez-Romero, Irving Barragán-Vite

Abstract:

Wind speed forecasting is an important issue for planning wind power generation facilities. The accuracy in the wind speed prediction allows a good performance of wind turbines for electricity generation. A model based on artificial neural networks is presented in this work. A dataset with atmospheric information about air temperature, atmospheric pressure, wind direction, and wind speed in Pachuca, Hidalgo, México, was used to train the artificial neural network. The data was downloaded from the web page of the National Meteorological Service of the Mexican government. The records were gathered for three months, with time intervals of ten minutes. This dataset was used to develop an iterative algorithm to create 1,110 ANNs, with different configurations, starting from one to three hidden layers and every hidden layer with a number of neurons from 1 to 10. Each ANN was trained with the Levenberg-Marquardt backpropagation algorithm, which is used to learn the relationship between input and output values. The model with the best performance contains three hidden layers and 9, 6, and 5 neurons, respectively; and the coefficient of determination obtained was r²=0.9414, and the Root Mean Squared Error is 1.0559. In summary, the ANN approach is suitable to predict the wind speed in Pachuca City because the r² value denotes a good fitting of gathered records, and the obtained ANN model can be used in the planning of wind power generation grids.

Keywords: wind power generation, artificial neural networks, wind speed, coefficient of determination

Procedia PDF Downloads 90

Authors: Tet H. Yeap

Abstract:

An analog restricted Hopfield Network is presented in this paper. It consists of two layers of nodes, visible and hidden nodes, connected by directional weighted paths forming a bipartite graph with no intralayer connection. An energy or Lyapunov function was derived to show that the proposed network will converge to stable states. By introducing hidden nodes, the proposed network can be trained to store patterns and has increased memory capacity. Training to be an associative memory, simulation results show that the associative memory performs better than a classical Hopfield network by being able to perform better memory recall when the input is noisy.

Keywords: restricted Hopfield network, Lyapunov function, simultaneous perturbation stochastic approximation

Procedia PDF Downloads 108

Authors: Eunji Choi, Borang Park, Youngjae Choi, Jinwoo Moon

Abstract:

In this study, a predictive model for estimating the metabolism (MET) of human body was developed for the optimal control of indoor thermal environment. Human body images for indoor activities and human body joint coordinated values were collected as data sets, which are used in predictive model. A deep learning algorithm was used in an initial model, and its number of hidden layers and hidden neurons were optimized. Lastly, the model prediction performance was analyzed after the model being trained through collected data. In conclusion, the possibility of MET prediction was confirmed, and the direction of the future study was proposed as developing various data and the predictive model.

Keywords: deep learning, indoor quality, metabolism, predictive model

Procedia PDF Downloads 237

Authors: Elabed Saud

Abstract:

This paper introduces the new ability of Artificial Neural Networks (ANNs) in estimating speed and controlling the separately excited DC motor. The neural control scheme consists of two parts. One is the neural estimator which is used to estimate the motor speed. The other is the neural controller which is used to generate a control signal for a converter. These two neutrals are training by Levenberg-Marquardt back-propagation algorithm. ANNs are the standard three layers feed-forward neural network with sigmoid activation functions in the input and hidden layers and purelin in the output layer. Simulation results are presented to demonstrate the effectiveness of this neural and advantage of the control system DC motor with ANNs in comparison with the conventional scheme without ANNs.

Keywords: Artificial Neural Network (ANNs), excited DC motor, convenional controller, speed Controller

Procedia PDF Downloads 698

Authors: Joselito Medina-Marin, Maria G. Serna-Diaz, Alejandro Tellez-Jurado, Juan C. Seck-Tuoh-Mora, Eva S. Hernandez-Gress, Norberto Hernandez-Romero, Iaina P. Medina-Serna

Abstract:

Trichoderma harzianum is a fungus that has been utilized as a low-cost fungicide for biological control of pests, and it is important to determine the optimal conditions to produce the highest amount of conidiospores of Trichoderma harzianum. In this work, the conidiospore production of Trichoderma harzianum is modeled and optimized by using Artificial Neural Networks (AANs). In order to gather data of this process, 30 experiments were carried out taking into account the number of hours of culture (10 distributed values from 48 to 136 hours) and the culture humidity (70, 75 and 80 percent), obtained as a response the number of conidiospores per gram of dry mass. The experimental results were used to develop an iterative algorithm to create 1,110 ANNs, with different configurations, starting from one to three hidden layers, and every hidden layer with a number of neurons from 1 to 10. Each ANN was trained with the Levenberg-Marquardt backpropagation algorithm, which is used to learn the relationship between input and output values. The ANN with the best performance was chosen in order to simulate the process and be able to maximize the conidiospores production. The obtained ANN with the highest performance has 2 inputs and 1 output, three hidden layers with 3, 10 and 10 neurons in each layer, respectively. The ANN performance shows an R2 value of 0.9900, and the Root Mean Squared Error is 1.2020. This ANN predicted that 644175467 conidiospores per gram of dry mass are the maximum amount obtained in 117 hours of culture and 77% of culture humidity. In summary, the ANN approach is suitable to represent the conidiospores production of Trichoderma harzianum because the R2 value denotes a good fitting of experimental results, and the obtained ANN model was used to find the parameters to produce the biggest amount of conidiospores per gram of dry mass.

Keywords: Trichoderma harzianum, modeling, optimization, artificial neural network

Procedia PDF Downloads 130

Authors: N. V. Kuznetsov, O. A. Kuznetsova, G. A. Leonov, M. V. Yuldashev, R. V. Yuldashev

Abstract:

Nonlinear analysis of the phase locked loop (PLL)-based circuits is a challenging task. Thus, the simulation is widely used for their study. In this work, we consider a mathematical model of the optical Costas loop and demonstrate the limitations of simulation approach related to the existence of so-called hidden oscillations in the phase space of the model.

Keywords: optical Costas loop, mathematical model, simulation, hidden oscillation

Procedia PDF Downloads 414

Authors: Poonam A. Trambadia, Ashish V. Trambadia

Abstract:

The architecture of sultanate period of Ahmedabad had evolved just before the establishment of Mughal rule in North India. After shifting the capital of the kingdom from Patan to Ahmedabad, when the buildings and structures were being built, an interesting cultural blend happened in architecture. Many sultanate buildings in Ahmedabad historic city have resemblance with Patan including the names. Outer fortification walls and Gates were built during the rule of the third ruler in the late 15th century. All the gates had sandstone slabs supported by three arched entrance in sandstone with wooden shutter. A restoration project of Khanpur Gate was initiated in 2016. The paper identifies some evidences and some hidden layers of structures as important elements of cultural interchange while some were just forgotten in the process. The recycling of pre-existing elements of structures are examined and compared. There were layers uncovered that were hidden behind later repairs using traditional brick arch, which was taken out in the process. As the gate had partially collapsed, the restoration included piece by piece dismantling and restoring in the same sequence wherever required. The recycled materials found in the process were recorded and provided the basis for this study. The gate after this discovery sets a new example of fortification Gate built in Sultanate era. The comparison excludes Maratha and British Period Gates to avoid further confusion and focuses on 15th – 16th century sultanate architecture of Ahmedabad.

Keywords: Ahmedabad World Heritage, fortification, Indo-Islamic style, Sultanate architecture, cultural interchange

Procedia PDF Downloads 99

Authors: R. Islam, Rafiqul A. Tarefder

Abstract:

This study presents the moisture variations of unbound layers from April 2012 to January 2014 in the Interstate 40 (I-40) pavement section in New Mexico. Three moisture probes were installed at different layers inside the pavement which measure the continuous moisture variations of the pavement. Data show that the moisture contents of unbound layers are typically constant throughout the day and month unless there is rainfall. Moisture contents of all unbound layers change with rainfall. Change in ground water table may affect the moisture content of unbound layers which has not investigated in this study. In addition, the Level 3 predictions of moisture contents using the Pavement Mechanistic-Empirical (ME) Design software are compared and found quite reasonable. However, results presented in the current study may not be applicable for pavement in other regions.

Keywords: asphalt pavement, moisture probes, resilient modulus, climate model

Procedia PDF Downloads 470

Authors: Grzegorz Zinkiewicz

Abstract:

The paper explores affective potential implicit in Bergman’s movies. This is done by the use of affect theory and the concept of affect in terms of paradigmatic and syntagmatic relations, from both diachronic and synchronic perspective. Since its inception in the early 2000s, affect theory has been applied to a number of academic fields. In Film Studies, it offers new avenues for discovering deeper, hidden layers of a given film. The aim is to show that the form and content of the films by Ingmar Bergman are determined by their inner affects that function independently of the viewer and, to an extent, are autonomous entities that can be analysed in separation from the auteur and actual characters. The paper discovers layers in Ingmar Bergman films and focuses on aspects that are often marginalised or studied from other viewpoints such as the connection between the content and visual side. As a result, a revaluation of Bergman films is possible that is more consistent with his original interpretations and comments included in his lectures, interviews and autobiography.

Keywords: affect theory, experimental cinema, Ingmar Bergman, viewer response

Procedia PDF Downloads 82

Authors: S. Mousavian, A. Abedianpour, A. Khanmohammadi, S. Hematian, Gh. Eidi Veisi

Abstract:

Ionic liquids are finding a wide range of applications from reaction media to separations and materials processing. In these applications, Vapor–Liquid equilibrium (VLE) is the most important one. VLE for six systems at 353 K and activity coefficients at infinite dilution 〖(γ〗_i^∞) for various solutes (alkanes, alkenes, cycloalkanes, cycloalkenes, aromatics, alcohols, ketones, esters, ethers, and water) in the ionic liquids (1-ethyl-3-methylimidazolium bis (trifluoromethylsulfonyl)imide [EMIM][BTI], 1-hexyl-3-methyl imidazolium bis (trifluoromethylsulfonyl) imide [HMIM][BTI], 1-octyl-3-methylimidazolium bis(trifluoromethylsulfonyl) imide [OMIM][BTI], and 1-butyl-1-methylpyrrolidinium bis (trifluoromethylsulfonyl) imide [BMPYR][BTI]) have been used to train neural networks in the temperature range from (303 to 333) K. Densities of the ionic liquids, Hildebrant constant of substances, and temperature were selected as input of neural networks. The networks with different hidden layers were examined. Networks with seven neurons in one hidden layer have minimum error and good agreement with experimental data.

Keywords: ionic liquid, neural networks, VLE, dilute solution

Procedia PDF Downloads 274

Authors: Justyna Chrzanowska, Jacek Hoffman, Dariusz Garbiec, Łukasz Kurpaska, Piotr Denis, Tomasz Moscicki, Zygmunt Szymanski

Abstract:

Tungsten and boron compounds belong to the group of superhard materials and its hardness could exceed 40 GPa. In this study, the properties of the tungsten boride (WB) layers deposited in magnetron sputtering process are investigated. The sputtering process occurred from specially prepared targets that were composed of boron and tungsten mixed in molar ratio of 2.5 or 4.5 and sintered in spark plasma sintering process. WB layers were deposited on silicon (100) and stainless steel 304 substrates at room temperature (RT) or in 570 °C. Layers deposited in RT and in elevated temperature varied considerably. Layers deposited in RT are amorphous and have low adhesion. In contrast, the layers deposited in 570 °C are crystalline and have good adhesion. All deposited layers have a hardness about 40 GPa. Moreover, the friction coefficient of crystalline layers is 0.22 and wear rate is about 0.67•10-6 mm3N-1m-1. After material characterization the WB layers were annealed in argon atmosphere in 1000 °C for 1 hour. On the basis of X-Ray Diffraction analysis, it has been noted that the crystalline layers are thermally stable and do not change their phase composition, whereas the amorphous layers change their phase composition. Moreover, after annealing, on the surface of WB layers some cracks were observed. It is probably connected with the differences of the thermal expansion between the layer and the substrate. Despite of the presence of cracks, the wear resistance of annealed layers is still higher than the wear resistance of uncoated substrate. The analysis of the structure and properties of tungsten boride layers lead to the discussion about the application area of this material.

Keywords: hard coatings, hard materials, magnetron sputtering, mechanical properties, tungsten boride

Procedia PDF Downloads 257

Authors: Zh. M. Blednova, P. O. Rusinov

Abstract:

Results of research on the formation of the surface layers of a material with shape memory effect (SME) based on TiNi diffusion metallization in molten Pb-Bi under isothermal conditions in an argon atmosphere are presented. It is shown that this method allows obtaining of uniform surface layers in nanostructured state of internal surfaces on the articles of complex shapes with stress concentrators. Structure, chemical and phase composition of the surface layers provide a manifestation of TiNi shape memory. The average grain size of TiNi coatings ranges between 60 ÷ 160 nm.

Keywords: diffusion metallization, nikelid titanium surface layers, shape memory effect, nanostructures

Procedia PDF Downloads 302

Authors: Johnson Joseph Kwabina Arhinful, Owusu-Ansah Emmanuel Degraft Johnson, Okyere Gabrial Asare, Adebanji Atinuke Olusola

Abstract:

This paper proposes a model to describe the blood types distribution of new Coronavirus (COVID-19) cases using the Bayesian Poisson - Hidden Markov Model (BP-HMM). With the help of the Gibbs sampler algorithm, using OpenBugs, the study first identifies the number of hidden states fitting European (EU) and African (AF) data sets of COVID-19 cases by blood type frequency. The study then compares the state-dependent mean of infection within and across the two geographical areas. The study findings show that the number of hidden states and infection rates within and across the two geographical areas differ according to blood type.

Keywords: BP-HMM, COVID-19, blood types, GIBBS sampler

Procedia PDF Downloads 106

Authors: Afnan Khoja, Ittisak Subrungruang, Kritchaya Ritruechai, Linda Jones, David Wall

Abstract:

Background: Some characteristics might be stated as an objective of the curriculum and some might be hidden. The hidden curriculum is the unwritten and unintended lessons and perspectives that students absorb in school. Though, the hidden characteristics are expected that tutors hope students might have in order to become medical educators. We suspected our faculty hoped we would develop skills, know and develop beyond the written outcomes. Our research question aimed to explore the hidden curriculum; as part of our learning; we had to design and report findings. Summary of Work: We undertook semi-structured interviews with a sample of the centre for medical education faculty at Dundee. Participants answered the question , of what are the hidden characteristics that they hope Dundee MMed graduates might have after graduation. Thematic analysis was carried out on the interview scripts. Summary of Results: A thematic analysis was carried out on the interview transcripts. Three main themes were identified from all respondents' comments. These were lifelong learners, being flexible and problem solvers. In addition individual respondents also described sense of humour, collaboration, humility, role model, inquisitiveness, optimism, and ability to express oneself clearly. Discussion: Tutors put great value on three behaviours lifelong learner, flexible, and problem solver, which are part of professional characteristics in leadership. Therefore, leadership characteristics is incorporated as the outcomes of hidden characteristics that tutors would like to see. Conclusion: Tutors in the Master's program of medical education at the University of Dundee hope that medical education students should present the three main hidden characteristics, which are lifelong learner, flexible, and problem solver after graduation. Take-home Messages: These hidden characteristics are considered as informal unless a change has been made to the formal curriculum. Therefore, to reach the tutors’ expectations, further studies might be held to make this personal characteristics transformation more accessible.

Keywords: characteristics, hidden curriculum, transformation, informal

Procedia PDF Downloads 62

Authors: J. K. Chawla, S. K. Jain, M. K. Mishra

Abstract:

Ion-acoustic double layers have been studied in magnetized plasma. The modified Korteweg-de Vries (m-KdV) equation using reductive perturbation method is derived. It is found that for the selected set of parameters, the system supports rarefactive double layers depending upon the value of nonthermal parameters. It is also found that the magnetization affects only the width of the double layer. For a given set of parameter values, increases in the magnetization and the obliqueness angle (θ) between wave vector and magnetic field, affect the width of the double layers, however the amplitude of the double layers have no effect. An increase in the values of nonthermal parameter decreases the amplitude of the rarefactive double layer. The effect of the ion temperature ratio on the amplitude and width of the double layers are also discussed in detail.

Keywords: ion-acoustic double layers, magnetized electronegative plasma, reductive perturbation method, the modified Korteweg-de Vries (KdV) equation

Procedia PDF Downloads 583

Authors: Ayuko Itsuki, Sachiyo Aburatani

Abstract:

We assessed the ecology of the organic and mineral soil layers of laurel-leaved (BB-1) and Cryptomeria japonica (BB-2 and Pw) forests in the Kasugayama Hill Primeval Forest (Nara, Japan). The soil respiration rate was higher in the deeper horizons (F and H) of organic layers than in those of mineral soil layers, suggesting organic layers may be where active microbial metabolism occurs. Respiration rates in the soil of BB-1, BB-2 and Pw forests were closely similar at 5 and 10°C. However, the soil respiration rate increased in proportion to temperatures of 15°C or above. We therefore consider the activity of soil microorganisms to markedly decrease at temperatures below 10°C. At a temperature of 15°C or above, the soil respiration rate in the BB-1 organic layers was higher than in those of the BB-2 and Pw organic layers, due to differences in forest vegetation that appeared to influence several salient soil properties, particularly pH and the carbon (C) and nitrogen (N) content of the F and H horizons.

Keywords: forest soil, mineralization rate, heterotroph, soil respiration rate

Procedia PDF Downloads 311

Authors: S. Curteanu, F. Leon, M. Gavrilescu, S. A. Floria

Abstract:

Optimization algorithms inspired from human behavior were applied in this approach, associated with neural networks models. The algorithms belong to human behaviors of learning and cooperation and human competitive behavior classes. For the first class, the main strategies include: random learning, individual learning, and social learning, and the selected algorithms are: simplified human learning optimization (SHLO), social learning optimization (SLO), and teaching-learning based optimization (TLBO). For the second class, the concept of learning is associated with competitiveness, and the selected algorithms are sports-inspired algorithms (with Football Game Algorithm, FGA and Volleyball Premier League, VPL) and Imperialist Competitive Algorithm (ICA). A real process, the synthesis of polyacrylamide-based multicomponent hydrogels, where some parameters are difficult to obtain experimentally, is considered as a case study. Reaction yield and swelling degree are predicted as a function of reaction conditions (acrylamide concentration, initiator concentration, crosslinking agent concentration, temperature, reaction time, and amount of inclusion polymer, which could be starch, poly(vinyl alcohol) or gelatin). The experimental results contain 175 data. Artificial neural networks are obtained in optimal form with biologically inspired algorithm; the optimization being perform at two level: structural and parametric. Feedforward neural networks with one or two hidden layers and no more than 25 neurons in intermediate layers were obtained with values of correlation coefficient in the validation phase over 0.90. The best results were obtained with TLBO algorithm, correlation coefficient being 0.94 for an MLP(6:9:20:2) – a feedforward neural network with two hidden layers and 9 and 20, respectively, intermediate neurons. Good results obtained prove the efficiency of the optimization algorithms. More than the good results, what is important in this approach is the simulation methodology, including neural networks and optimization biologically inspired algorithms, which provide satisfactory results. In addition, the methodology developed in this approach is general and has flexibility so that it can be easily adapted to other processes in association with different types of models.

Keywords: artificial neural networks, human behaviors of learning and cooperation, human competitive behavior, optimization algorithms

Procedia PDF Downloads 89

Authors: Andrew N. Saylor, James R. Peters

Abstract:

Scoliosis is a complex 3D deformity of the thoracic and lumbar spines, clinically diagnosed by measurement of a Cobb angle of 10 degrees or more on a coronal X-ray. The Cobb angle is the angle made by the lines drawn along the proximal and distal endplates of the respective proximal and distal vertebrae comprising the curve. Traditionally, Cobb angles are measured manually using either a marker, straight edge, and protractor or image measurement software. The task of measuring the Cobb angle can also be represented by a function taking the spine geometry rendered using X-ray imaging as input and returning the approximate angle. Although the form of such a function may be unknown, it can be approximated using artificial neural networks (ANNs). The performance of ANNs is affected by many factors, including the choice of activation function and network architecture; however, the effects of these parameters on the accuracy of scoliotic deformity measurements are poorly understood. Therefore, the objective of this study was to systematically investigate the effect of ANN architecture and activation function on Cobb angle measurement from the coronal X-rays of scoliotic subjects. The data set for this study consisted of 609 coronal chest X-rays of scoliotic subjects divided into 481 training images and 128 test images. These data, which included labeled Cobb angle measurements, were obtained from the SpineWeb online database. In order to normalize the input data, each image was resized using bi-linear interpolation to a size of 500 × 187 pixels, and the pixel intensities were scaled to be between 0 and 1. A fully connected (dense) ANN with a fixed cost function (mean squared error), batch size (10), and learning rate (0.01) was developed using Python Version 3.7.3 and TensorFlow 1.13.1. The activation functions (sigmoid, hyperbolic tangent [tanh], or rectified linear units [ReLU]), number of hidden layers (1, 3, 5, or 10), and number of neurons per layer (10, 100, or 1000) were varied systematically to generate a total of 36 network conditions. Stochastic gradient descent with early stopping was used to train each network. Three trials were run per condition, and the final mean squared errors and mean absolute errors were averaged to quantify the network response for each condition. The network that performed the best used ReLU neurons had three hidden layers, and 100 neurons per layer. The average mean squared error of this network was 222.28 ± 30 degrees2, and the average mean absolute error was 11.96 ± 0.64 degrees. It is also notable that while most of the networks performed similarly, the networks using ReLU neurons, 10 hidden layers, and 1000 neurons per layer, and those using Tanh neurons, one hidden layer, and 10 neurons per layer performed markedly worse with average mean squared errors greater than 400 degrees2 and average mean absolute errors greater than 16 degrees. From the results of this study, it can be seen that the choice of ANN architecture and activation function has a clear impact on Cobb angle inference from coronal X-rays of scoliotic subjects.

Keywords: scoliosis, artificial neural networks, cobb angle, medical imaging

Procedia PDF Downloads 103

Authors: H. Hruschka

Abstract:

This paper presents the first comparison of the performance of the restricted Boltzmann machine and the deep belief net on binary market basket data relative to binary factor analysis and the two best-known topic models, namely Dirichlet allocation and the correlated topic model. This comparison shows that the restricted Boltzmann machine and the deep belief net are superior to both binary factor analysis and topic models. Managerial implications that differ between the investigated models are treated as well. The restricted Boltzmann machine is defined as joint Boltzmann distribution of hidden variables and observed variables (purchases). It comprises one layer of observed variables and one layer of hidden variables. Note that variables of the same layer are not connected. The comparison also includes deep belief nets with three layers. The first layer is a restricted Boltzmann machine based on category purchases. Hidden variables of the first layer are used as input variables by the second-layer restricted Boltzmann machine which then generates second-layer hidden variables. Finally, in the third layer hidden variables are related to purchases. A public data set is analyzed which contains one month of real-world point-of-sale transactions in a typical local grocery outlet. It consists of 9,835 market baskets referring to 169 product categories. This data set is randomly split into two halves. One half is used for estimation, the other serves as holdout data. Each model is evaluated by the log likelihood for the holdout data. Performance of the topic models is disappointing as the holdout log likelihood of the correlated topic model – which is better than Dirichlet allocation - is lower by more than 25,000 compared to the best binary factor analysis model. On the other hand, binary factor analysis on its own is clearly surpassed by both the restricted Boltzmann machine and the deep belief net whose holdout log likelihoods are higher by more than 23,000. Overall, the deep belief net performs best. We also interpret hidden variables discovered by binary factor analysis, the restricted Boltzmann machine and the deep belief net. Hidden variables characterized by the product categories to which they are related differ strongly between these three models. To derive managerial implications we assess the effect of promoting each category on total basket size, i.e., the number of purchased product categories, due to each category's interdependence with all the other categories. The investigated models lead to very different implications as they disagree about which categories are associated with higher basket size increases due to a promotion. Of course, recommendations based on better performing models should be preferred. The impressive performance advantages of the restricted Boltzmann machine and the deep belief net suggest continuing research by appropriate extensions. To include predictors, especially marketing variables such as price, seems to be an obvious next step. It might also be feasible to take a more detailed perspective by considering purchases of brands instead of purchases of product categories.

Keywords: binary factor analysis, deep belief net, market basket analysis, restricted Boltzmann machine, topic models

Procedia PDF Downloads 172

Authors: Brahim-Fares Zaidi, Malika Boudraa, Sid-Ahmed Selouani

Abstract:

Our work aims to improve our Automatic Recognition System for Dysarthria Speech (ARSDS) based on the Hidden Models of Markov (HMM) and the Hidden Markov Model Toolkit (HTK) to help people who are sick. With pronunciation problems, we applied two techniques of speech parameterization based on Mel Frequency Cepstral Coefficients (MFCC's) and Perceptual Linear Prediction (PLP's) and concatenated them with JITTER and SHIMMER coefficients in order to increase the recognition rate of a dysarthria speech. For our tests, we used the NEMOURS database that represents speakers with dysarthria and normal speakers.

Keywords: hidden Markov model toolkit (HTK), hidden models of Markov (HMM), Mel-frequency cepstral coefficients (MFCC), perceptual linear prediction (PLP’s)

Procedia PDF Downloads 136

Authors: Pavan K. Rallabandi, Kailash C. Patidar

Abstract:

In this paper, we present an optimization technique or a learning algorithm using the hybrid architecture by combining the most popular sequence recognition models such as Recurrent Neural Networks (RNNs) and Hidden Markov models (HMMs). In order to improve the sequence or pattern recognition/ classification performance by applying a hybrid/neural symbolic approach, a gradient descent learning algorithm is developed using the Real Time Recurrent Learning of Recurrent Neural Network for processing the knowledge represented in trained Hidden Markov Models. The developed hybrid algorithm is implemented on automata theory as a sample test beds and the performance of the designed algorithm is demonstrated and evaluated on learning the deterministic finite state automata.

Keywords: hybrid systems, hidden markov models, recurrent neural networks, deterministic finite state automata

Procedia PDF Downloads 363

Authors: Sandy Bao, Yankang Bao

Abstract:

The understanding of information that is being processed in the brain, especially in the lateral geniculate nucleus (LGN), has been proven challenging for modern neuroscience and for researchers with a focus on how neurons process signals and images. In this paper, we are proposing a method to image process different colors within different layers of LGN, that is, green information in layers 4 & 6 and red & blue in layers 3 & 5 based on the surface dimension of layers. We take into consideration the images in LGN and visual cortex, and that the edge detected information from the visual cortex needs to be considered in order to return back to the layers of LGN, along with the image in LGN to form the new image, which will provide an improved image that is clearer, sharper, and making it easier to identify objects in the image. Matrix Laboratory (MATLAB) simulation is performed, and results show that the clarity of the output image has significant improvement.

Keywords: lateral geniculate nucleus, matrix laboratory, neuroscience, visual cortex

Procedia PDF Downloads 241

Authors: Ricardo A. Martins, Matheus S. da Silva, Gabriel H. F. Iarossi, Helen C. M. Senefonte, Cinthyan R. S. C. de Barbosa

Abstract:

This paper is the development of a concept of the man/robot interaction. More accurately in developing of an autistic child that have more troubles with interaction, here offers an efficient solution, even though simple; however, less studied for this public. This concept is based on code applied thought out the Lego NXT kit, built for the interpretation of the robot, thereby can create this interaction in a constructive way for children suffering with Autism.

Keywords: lego NXT, interaction, BricX, autismo, ANN (Artificial Neural Network), MLP back propagation, hidden layers

Procedia PDF Downloads 541

Authors: Khenfer Koummich Fatma, Hendel Fatiha, Mesbahi Larbi

Abstract:

This paper presents an approach based on Hidden Markov Models (HMM: Hidden Markov Model) using HTK tools. The goal is to create a human-machine interface with a voice recognition system that allows the operator to teleoperate a mentor robot to execute specific tasks as rotate, raise, close, etc. This system should take into account different levels of environmental noise. This approach has been applied to isolated words representing the robot commands pronounced in two languages: French and Arabic. The obtained recognition rate is the same in both speeches, Arabic and French in the neutral words. However, there is a slight difference in favor of the Arabic speech when Gaussian white noise is added with a Signal to Noise Ratio (SNR) equals 30 dB, in this case; the Arabic speech recognition rate is 69%, and the French speech recognition rate is 80%. This can be explained by the ability of phonetic context of each speech when the noise is added.

Keywords: Arabic speech recognition, Hidden Markov Model (HMM), HTK, noise, TIMIT, voice command

Procedia PDF Downloads 349

Authors: Nana Saralidze, Nino Sharashenidze, Zurab Kevanishvili

Abstract:

Hidden hearing loss may occur in humans exposed to prolonged high-level sound. It is the loss of ability to hear high-level background noise while having normal hearing in quiet. We compared the hearing of people who regularly listen 3 hours and more to personal music players and those who do not. Forty participants aged 18-30 years were divided into two groups: regular users of music players and people who had never used them. And the third group – elders aged 50-55 years, had 15 participants. Pure-tone audiometry (125-16000 Hz), auditory brainstem response (ABR) (70dB SPL), and ability to identify speech in noise (4-talker babble with a 65-dB signal-to-noise ratio at 80 dB) were measured in all participants. All participants had normal pure-tone audiometry (all thresholds < 25 dB HL). A significant difference between groups was observed in that regular users of personal audio systems correctly identified 53% of words, whereas the non-users identified 74% and the elder group – 63%. This contributes evidence supporting the presence of a hidden hearing loss in humans and demonstrates that speech-in-noise audiometry is an effective method and can be considered as the GOLD standard for detecting hidden hearing loss.

Keywords: mp3 player, hidden hearing loss, speech audiometry, pure tone audiometry

Procedia PDF Downloads 46

Authors: R. K. Pateriya, Jyoti Bharti

Abstract:

This paper presents an approach for data hiding methods which provides a secret communication between sender and receiver. The data is hidden in gray-scale images and the boundary of gray-scale image is used to store the mapping information. In this an approach data is in ASCII format and the mapping is in between ASCII value of hidden message and pixel value of cover image, since pixel value of an image as well as ASCII value is in range of 0 to 255 and this mapping information is occupying only 1 bit per character of hidden message as compared to 8 bit per character thus maintaining good quality of stego image.

Keywords: ASCII value, cover image, PSNR, pixel value, stego image, secret message

Procedia PDF Downloads 395

Authors: Seyed Abolhasan Naeini, Eisa Aliagahei

Abstract:

Today, the use of geosynthetic materials in geotechnical activities is increasing significantly. One of the main uses of these materials is to increase the compressive strength of clay reinforced by geotextile layers. In the present study, the effect of clay reinforcement by geotextile layers in increasing the compressive strength of clay has been investigated using modeling in ABAQUS 6.11.3 software. For this purpose, the modified Drager Prager model has been chosen to simulate the stress-strain behavior of soil layers and the linear elastic model for the geotextile layer. Unreinforced samples and reinforced samples are modeled by geotextile layers (1, 2 and 3 geotextile layers) by software. In order to validate the results, an article in the same field was used and the numerical modeling results were calibrated with the laboratory results. Based on the obtained results, the software has a suitable capability for modeling and the results of the numerical model overlap with the laboratory results to a very acceptable extent, by increasing the number of geotextile layers, the error between the results of the laboratory sample and the software model increases. The highest amount of error is related to the sample reinforced with three layers of geotextile and is 7.3%.

Keywords: Abaqus, cap model, clay, geotextile layer, reinforced soil

Procedia PDF Downloads 68