Search results for: rough neural networks

Authors: YingWei Tan, XueFeng Ding

Abstract:

Attention-based voice activity detection models have gained significant attention in recent years due to their fast training speed and ability to capture a wide contextual range. The inclusion of multi-head style and position embedding in the attention architecture are crucial. Having multiple attention heads allows for differential focus on different parts of the sequence, while position embedding provides guidance for modeling dependencies between elements at various positions in the input sequence. In this work, we propose an approach by considering each head as a node, enabling the application of graph neural networks (GNN) to identify correlations among the different nodes. In addition, we adopt an implementation named rotary position embedding (RoPE), which encodes absolute positional information into the input sequence by a rotation matrix, and naturally incorporates explicit relative position information into a self-attention module. We evaluate the effectiveness of our method on a synthetic dataset, and the results demonstrate its superiority over the baseline CRNN in scenarios with low signal-to-noise ratio and noise, while also exhibiting robustness across different noise types. In summary, our proposed framework effectively combines the strengths of CNN and RNN (LSTM), and further enhances detection performance through the integration of graph neural networks and rotary position embedding.

Keywords: voice activity detection, CRNN, graph neural networks, rotary position embedding

Procedia PDF Downloads 72

Authors: Jérémie Ochin

Abstract:

Soccer analytics are built on two data sources: the frame-by-frame position of each player on the terrain and the sequences of events, such as ball drive, pass, cross, shot, throw-in... With more than 2000 ball-events per soccer game, their precise and exhaustive annotation, based on a monocular video stream such as a TV broadcast, remains a tedious and costly manual task. State-of-the-art methods for spatio-temporal action detection from a monocular video stream, often based on 3D convolutional neural networks, are close to reach levels of performances in mean Average Precision (mAP) compatibles with the automation of such task. Nevertheless, to meet their expectation of exhaustiveness in the context of data analytics, such methods must be applied in a regime of high recall – low precision, using low confidence score thresholds. This setting unavoidably leads to the detection of false positives that are the product of the well documented overconfidence behaviour of neural networks and, in this case, their limited access to contextual information and understanding of the game: their predictions are highly unstructured. Based on the assumption that professional soccer players’ behaviour, pose, positions and velocity are highly interrelated and locally driven by the player performing a ball-action, it is hypothesized that the addition of information regarding surrounding player’s appearance, positions and velocity in the prediction methods can improve their metrics. Several methods are compared to build a proper representation of the game surrounding a player, from handcrafted features of the local graph, based on domain knowledge, to the use of Graph Neural Networks trained in an end-to-end fashion with existing state-of-the-art 3D convolutional neural networks. It is shown that the inclusion of information regarding surrounding players helps reaching higher metrics.

Keywords: fine-grained action recognition, human action recognition, convolutional neural networks, graph neural networks, spatio-temporal action recognition

Procedia PDF Downloads 28

Authors: Tia Claridge

Abstract:

The aim of this study was to explore whether there are gender differences in how early childhood educators view and facilitate rough and tumble play in England. A qualitative approach was used to carry out semi-structured interviews with female and male early years educators. The key rationale for this study was to examine the significant lack of males working in early years education and the consequent impact this has on pedagogical practice. The findings illustrated that there are some gender differences in educators’ perspectives of rough and tumble play. These include how educators use their own childhood experience to inform their professional practice as well as identifying a need for tailored training to upskill and develop confidence in early years staff with regard to this type of play. The most surprising finding was the influence that urban and rural settings had on educators’ perceptions on weapon play, regardless of gender. Awareness of educator positionality was significant throughout the study for male participants, whereas females rarely commented their own gender. These findings indicate that further research is needed to understand the complex narratives underpinning gender and rough and tumble play.

Keywords: rough and tumble play, educators, gender, early years, pedagogy

Procedia PDF Downloads 94

Authors: Rosdyana Mangir Irawan Kusuma, Wei-Chun Kao, Ho-Thi Trang, Yu-Yen Ou, Kai-Lung Hua

Abstract:

Stock market prediction is still a challenging problem because there are many factors that affect the stock market price such as company news and performance, industry performance, investor sentiment, social media sentiment, and economic factors. This work explores the predictability in the stock market using deep convolutional network and candlestick charts. The outcome is utilized to design a decision support framework that can be used by traders to provide suggested indications of future stock price direction. We perform this work using various types of neural networks like convolutional neural network, residual network and visual geometry group network. From stock market historical data, we converted it to candlestick charts. Finally, these candlestick charts will be feed as input for training a convolutional neural network model. This convolutional neural network model will help us to analyze the patterns inside the candlestick chart and predict the future movements of the stock market. The effectiveness of our method is evaluated in stock market prediction with promising results; 92.2% and 92.1 % accuracy for Taiwan and Indonesian stock market dataset respectively.

Keywords: candlestick chart, deep learning, neural network, stock market prediction

Procedia PDF Downloads 449

Authors: Nihal Murali, Kunal Gupta, Surekha Bhanot

Abstract:

Training of artificial neural networks (ANNs) using reinforcement learning (RL) techniques is being widely discussed in the robot learning literature. The high model complexity of ANNs along with the model-free nature of RL algorithms provides a desirable combination for many robotics applications. There is a huge need for algorithms that generalize using raw sensory inputs, such as vision, without any hand-engineered features or domain heuristics. In this paper, the standard control problem of line following robot was used as a test-bed, and an ANN controller for the robot was trained on images from a live video feed using Q-learning. A virtual agent was first trained in simulation environment and then deployed onto a robot’s hardware. The robot successfully learns to traverse a wide range of curves and displays excellent generalization ability. Qualitative analysis of the evolution of policies, performance and weights of the network provide insights into the nature and convergence of the learning algorithm.

Keywords: artificial neural networks, q-learning, reinforcement learning, robot learning

Procedia PDF Downloads 372

Authors: Vishnu Pratap Singh Kirar

Abstract:

The main aim of the communication system is to achieve maximum performance. In cognitive radio, any user or transceiver have the ability to sense best suitable channel, while the channel is not in use. It means an unlicensed user can share the spectrum of licensed user without any interference. Though the spectrum sensing consumes a large amount of energy and it can reduce by applying various artificial intelligent methods for determining proper spectrum holes. It also increases the efficiency of Cognitive Radio Network (CRN). In this survey paper, we discuss the use of different learning models and implementation of Artificial Neural Network (ANN) to increase the learning and decision-making capacity of CRN without affecting bandwidth, cost and signal rate.

Keywords: artificial neural network, cognitive radio, cognitive radio networks, back propagation, spectrum sensing

Procedia PDF Downloads 609

Authors: Hsu-Yung Cheng, Kuo-Chang Hsu, Chi-Chang Chan, Mei-Hui Tseng, Chih-Chang Yu, Ya-Sheng Liu

Abstract:

Photovoltaics modules are usually not installed horizontally to avoid water or dust accumulation. However, the measured irradiance data on tilted surfaces are rarely available since installing pyranometers with various tilt angles induces high costs. Therefore, estimating solar irradiance on tilted surfaces is an important research topic. In this work, artificial neural networks (ANN) are utilized to construct the transfer model to estimate solar irradiance on tilted surfaces. Instead of predicting tilted irradiance directly, the proposed method estimates the differences between the horizontal irradiance and the irradiance on a tilted surface. The outputs of the ANNs in the proposed design are differential values. The experimental results have shown that the proposed ANNs with differential outputs can substantially improve the estimation accuracy compared to ANNs that estimate the titled irradiance directly.

Keywords: photovoltaics, artificial neural networks, tilted irradiance, solar energy

Procedia PDF Downloads 397

Authors: Gaurav Sancheti

Abstract:

This study presents an impacting approach of Artificial Neural Networks (ANNs) in determining the quantity of High Tensile Steel (HTS) strands required in post-tensioned (PT) slabs. Various PT slab configurations were generated by varying the span and depth of the slab. For each of these slab configurations, quantity of required HTS strands were recorded. ANNs with backpropagation algorithm and varying architectures were developed and their performance was evaluated in terms of Mean Square Error (MSE). The recorded data for the quantity of HTS strands was used as a feeder database for training the developed ANNs. The networks were validated using various validation techniques. The results show that the proposed ANNs have a great potential with good prediction and generalization capability.

Keywords: artificial neural networks, back propagation, conceptual design, high tensile steel strands, post tensioned slabs, validation techniques

Procedia PDF Downloads 221

Authors: Ashish James, Anusha James

Abstract:

The impressive predictive performance of deep learning techniques on a wide range of tasks has led to its widespread use. Estimating the confidence of these predictions is paramount for improving the safety and reliability of such systems. However, the uncertainty estimates provided by neural networks (NNs) tend to be overconfident and unreasonable. Ensemble of NNs typically produce good predictions but uncertainty estimates tend to be inconsistent. Inspired by these, this paper presents a framework that can quantitatively estimate the uncertainties by leveraging the advances in transfer learning through slight modification to the existing training pipelines. This promising algorithm is developed with an intention of deployment in real world problems which already boast a good predictive performance by reusing those pretrained models. The idea is to capture the behavior of the trained NNs for the base task by augmenting it with the uncertainty estimates from a supplementary network. A series of experiments with known and unknown distributions show that the proposed approach produces well calibrated uncertainty estimates with high quality predictions.

Keywords: uncertainty estimation, neural networks, transfer learning, regression

Procedia PDF Downloads 136

Authors: R. Boudjelthia

Abstract:

The aim of this study is to determine the influence of mix composition of concrete as the content of water and cement, water–binder ratio, and the replacement of fly ash on the durability of self compacting concrete (SCC) by using artificial neural networks (ANNs). To achieve this, an ANNs model is developed to predict the durability of self compacting concrete which is expressed in terms of chloride ions permeability in accordance with ASTM C1202-97 or AASHTO T277. Database gathered from the literature for the training and testing the model. A sensitivity analysis was also conducted using the trained and tested ANN model to investigate the effect of fly ash on the durability of SCC. The results indicate that the developed model is reliable and accurate. the durability of SCC expressed in terms of total charge passed over a 6-h period can be significantly improved by using at least 25% fly ash as replacement of cement. This study show that artificial neural network have strong potentialas a feasible tool for predicting accurately the durability of SCC containing fly ash.

Keywords: artificial neural networks, durability, chloride ions permeability, self compacting concrete

Procedia PDF Downloads 379

Authors: Philip Symonds, Jon Taylor, Zaid Chalabi, Michael Davies

Abstract:

With the world climate projected to warm and major cities in developing countries becoming increasingly populated and polluted, governments are tasked with the problem of overheating and air quality in residential buildings. This paper presents the development of an adaptable model of these risks. Simulations are performed using the EnergyPlus building physics software. An accurate metamodel is formed by randomly sampling building input parameters and training on the outputs of EnergyPlus simulations. Metamodels are used to vastly reduce the amount of computation time required when performing optimisation and sensitivity analyses. Neural Networks (NNs) are compared to a Radial Basis Function (RBF) algorithm when forming a metamodel. These techniques were implemented using the PyBrain and scikit-learn python libraries, respectively. NNs are shown to perform around 15% better than RBFs when estimating overheating and air pollution metrics modelled by EnergyPlus.

Keywords: neural networks, radial basis functions, metamodelling, python machine learning libraries

Procedia PDF Downloads 447

Authors: F. Allag, S. Bouharati, M. Belmahdi, R. Zegadi

Abstract:

The rapid expansion of deserts in recent decades as a result of human actions combined with climatic changes has highlighted the necessity to understand biological processes in arid environments. Whereas physical processes and the biology of flora and fauna have been relatively well studied in marginally used arid areas, knowledge of desert soil micro-organisms remains fragmentary. The objective of this study is to conduct a diversity analysis of bacterial communities in unvegetated arid soils. Several biological phenomena in hot deserts related to microbial populations and the potential use of micro-organisms for restoring hot desert environments. Dry land ecosystems have a highly heterogeneous distribution of resources, with greater nutrient concentrations and microbial densities occurring in vegetated than in bare soils. In this work, we found it useful to use techniques of artificial intelligence in their treatment especially artificial neural networks (ANN). The use of the ANN model, demonstrate his capability for addressing the complex problems of uncertainty data.

Keywords: desert soil, climatic changes, bacteria, vegetation, artificial neural networks

Procedia PDF Downloads 395

Authors: Guanghua Zhang, Fubao Wang, Weijun Duan

Abstract:

Convolution neural network (CNN) has attracted more and more attention on recent years. Especially in the field of computer vision and image classification. However, unsupervised learning with CNN has received less attention than supervised learning. In this work, we use a new powerful tool which is deep convolutional generative adversarial networks (DCGANs) to generate images from Sloan Digital Sky Survey. Training by various star and galaxy images, it shows that both the generator and the discriminator are good for unsupervised learning. In this paper, we also took several experiments to choose the best value for hyper-parameters and which could help to stabilize the training process and promise a good quality of the output.

Keywords: convolution neural network, discriminator, generator, unsupervised learning

Procedia PDF Downloads 268

Authors: M. Heydari Vini

Abstract:

There is a complicated relation between effective input parameters of cold rolling and output rolling force and exit thickness of strips.in many mathematical models, the effect of some rolling parameters have been ignored and the outputs have not a desirable accuracy. In the other hand, there is a special relation among input thickness of strips,the width of the strips,rolling speeds,mandrill tensions and the required exit thickness of strips with rolling force and the real exit thickness of the rolled strip. First of all, in this paper the effective parameters of cold rolling process modeled using an artificial neural network according to the optimum network achieved by using a written program in MATLAB,it has been shown that the prediction of rolling stand parameters with different properties and new dimensions attained from prior rolled strips by an artificial neural network is applicable.

Keywords: cold rolling, artificial neural networks, rolling force, real rolled thickness of strips

Procedia PDF Downloads 505

Authors: Said Benkaciali, Mourad Haddadi, Abdallah Khellaf, Kacem Gairaa, Mawloud Guermoui

Abstract:

In this study, many experiments were carried out to assess the influence of the input parameters on the performance of multilayer perceptron which is one the configuration of the artificial neural networks. To estimate the daily global solar radiation on the horizontal surface, we have developed some models by using seven combinations of twelve meteorological and geographical input parameters collected from a radiometric station installed at Ghardaïa city (southern of Algeria). For selecting of best combination which provides a good accuracy, six statistical formulas (or statistical indicators) have been evaluated, such as the root mean square errors, mean absolute errors, correlation coefficient, and determination coefficient. We noted that multilayer perceptron techniques have the best performance, except when the sunshine duration parameter is not included in the input variables. The maximum of determination coefficient and correlation coefficient are equal to 98.20 and 99.11%. On the other hand, some empirical models were developed to compare their performances with those of multilayer perceptron neural networks. Results obtained show that the neural networks techniques give the best performance compared to the empirical models.

Keywords: empirical models, multilayer perceptron neural network, solar radiation, statistical formulas

Procedia PDF Downloads 345

Authors: Abdelhamid Louliej, Younes Jabrane

Abstract:

Satellites are becoming indispensable assets to fifth-generation (5G) new radio architecture, complementing wireless and terrestrial communication links. The combination of satellites and 5G architecture allows consumers to access all next-generation services anytime, anywhere, including scenarios, like traveling to remote areas (without coverage). Nevertheless, this solution faces several challenges, such as a significant propagation delay, Doppler frequency shift, and high Peak-to-Average Power Ratio (PAPR), causing signal distortion due to the non-linear saturation of the High-Power Amplifier (HPA). To compensate for HPA non-linearity in 5G satellite transmission, an efficient pre-distorter scheme using Neural Networks (NN) is proposed. To assess the proposed NN pre-distorter, two types of HPA were investigated: Travelling Wave Tube Amplifier (TWTA) and Solid-State Power Amplifier (SSPA). The results show that the NN pre-distorter design presents EVM improvement by 95.26%. NMSE and ACPR were reduced by -43,66 dB and 24.56 dBm, respectively. Moreover, the system suffers no degradation of the Bit Error Rate (BER) for TWTA and SSPA amplifiers.

Keywords: satellites, 5G, neural networks, HPA, TWTA, SSPA, EVM, NMSE, ACPR

Procedia PDF Downloads 91

Authors: Wei Zhang

Abstract:

With the rapid development of deep learning, neural network and deep learning algorithms play a significant role in various practical applications. Due to the high accuracy and good performance, Convolutional Neural Networks (CNNs) especially have become a research hot spot in the past few years. However, the size of the networks becomes increasingly large scale due to the demands of the practical applications, which poses a significant challenge to construct a high-performance implementation of deep learning neural networks. Meanwhile, many of these application scenarios also have strict requirements on the performance and low-power consumption of hardware devices. Therefore, it is particularly critical to choose a moderate computing platform for hardware acceleration of CNNs. This article aimed to survey the recent advance in Field Programmable Gate Array (FPGA)-based acceleration of CNNs. Various designs and implementations of the accelerator based on FPGA under different devices and network models are overviewed, and the versions of Graphic Processing Units (GPUs), Application Specific Integrated Circuits (ASICs) and Digital Signal Processors (DSPs) are compared to present our own critical analysis and comments. Finally, we give a discussion on different perspectives of these acceleration and optimization methods on FPGA platforms to further explore the opportunities and challenges for future research. More helpfully, we give a prospect for future development of the FPGA-based accelerator.

Keywords: deep learning, field programmable gate array, FPGA, hardware accelerator, convolutional neural networks, CNN

Procedia PDF Downloads 128

Authors: Yasser F. Hassan

Abstract:

Deep learning structure is a branch of machine learning science and greet achievement in research and applications. Cellular neural networks are regarded as array of nonlinear analog processors called cells connected in a way allowing parallel computations. The paper discusses how to use deep learning structure for representing neural cellular automata model. The proposed learning technique in cellular automata model will be examined from structure of deep learning. A deep automata neural cellular system modifies each neuron based on the behavior of the individual and its decision as a result of multi-level deep structure learning. The paper will present the architecture of the model and the results of simulation of approach are given. Results from the implementation enrich deep neural cellular automata system and shed a light on concept formulation of the model and the learning in it.

Keywords: cellular automata, neural cellular automata, deep learning, classification

Procedia PDF Downloads 198

Authors: Almagul Assainova , Dariya Abykenova, Liudmila Goncharenko, Sergey Sybachin, Saule Rakhimova, Abay Aman

Abstract:

The recognition of a handwritten Kazakh text is a relevant objective today for the digitization of materials. The study presents a model of a hybrid neural network for handwriting recognition, which includes a convolutional neural network and a multi-layer perceptron. Each network includes 1024 input neurons and 42 output neurons. The model is implemented in the program, written in the Python programming language using the EMNIST database, NumPy, Keras, and Tensorflow modules. The neural network training of such specific letters of the Kazakh alphabet as ә, ғ, қ, ң, ө, ұ, ү, h, і was conducted. The neural network model and the program created on its basis can be used in electronic document management systems to digitize the Kazakh text.

Keywords: handwriting recognition system, image recognition, Kazakh font, machine learning, neural networks

Procedia PDF Downloads 262

Authors: Yilin Liao, Hewen Li, Paula McConvey

Abstract:

Speed skaters often face a risk of concussion when they fall on the ice floor and impact crash mats during practices and competitive races. Several variables, including those related to the skater, the crash mat, and the impact position (body side/head/feet impact), are believed to influence the severity of the skater's concussion. While computer simulation modeling can be employed to analyze these accidents, the simulation process is time-consuming and does not provide rapid information for coaches and teams to assess the skater's injury risk in competitive events. This research paper promotes the exploration of the feasibility of using AI techniques for evaluating skater’s potential concussion severity, and to develop a fast concussion prediction tool using artificial neural networks to reduce the risk of treatment delays for injured skaters. The primary data is collected through virtual tests and physical experiments designed to simulate skater-mat impact. It is then analyzed to identify patterns and correlations; finally, it is used to train and fine-tune the artificial neural networks for accurate prediction. The development of the prediction tool by employing machine learning strategies contributes to the application of AI methods in sports science and has theoretical involvements for using AI techniques in predicting and preventing sports-related injuries.

Keywords: artificial neural networks, concussion, machine learning, impact, speed skater

Procedia PDF Downloads 109

Authors: H. Al-Qassem, L. Cheng, Y. Pan

Abstract:

In this paper, we establish sharp bounds for oscillatory singular integrals with an arbitrary real polynomial phase P. Our kernels are allowed to be rough both on the unit sphere and in the radial direction. We show that the bounds grow no faster than log (deg(P)), which is optimal and was first obtained by Parissis and Papadimitrakis for kernels without any radial roughness. Our results substantially improve many previously known results. Among key ingredients of our methods are an L¹→L² sharp estimate and using extrapolation.

Keywords: oscillatory singular integral, rough kernel, singular integral, orlicz spaces, block spaces, extrapolation, L^{p} boundedness

Procedia PDF Downloads 457

Authors: Seunghee Park, Junkyeong Kim, Eun-Seok Shin, Sang-Hun Han

Abstract:

In this study, two kinds of nondestructive evaluation (NDE) techniques (rebound hardness and ultrasonic pulse velocity methods) are investigated for the effective maintenance of underwater concrete structures. A new methodology to estimate the underwater concrete strengths more effectively, named “artificial neural network (ANN) – based concrete strength estimation with the combination of rebound hardness and ultrasonic pulse velocity methods” is proposed and verified throughout a series of experimental works.

Keywords: underwater concrete, rebound hardness, Schmidt hammer, ultrasonic pulse velocity, ultrasonic sensor, artificial neural networks, ANN

Procedia PDF Downloads 532

Authors: Benjamin Bobbia, Matthias Picard

Abstract:

In recent years active learning methodologies based on the representativity of the data seems more promising to limit overfitting. The presented query methodology for regression using the Wasserstein distance measuring the representativity of our labelled dataset compared to the global distribution. In this work a crucial use of GroupSort Neural Networks is made therewith to draw a double advantage. The Wasserstein distance can be exactly expressed in terms of such neural networks. Moreover, one can provide explicit bounds for their size and depth together with rates of convergence. However, heterogeneity of the dataset is also considered by weighting the Wasserstein distance with the error of approximation at the previous step of active learning. Such an approach leads to a reduction of overfitting and high prediction performance after few steps of query. After having detailed the methodology and algorithm, an empirical study is presented in order to investigate the range of our hyperparameters. The performances of this method are compared, in terms of numbers of query needed, with other classical and recent query methods on several UCI datasets.

Keywords: active learning, Lipschitz regularization, neural networks, optimal transport, regression

Procedia PDF Downloads 80

Authors: Jiangxia Han, Liang Xue, Keda Chen

Abstract:

Despite the significant progress over the last decades in reservoir simulation using numerical discretization, meshing is complex. Moreover, the high degree of freedom of the space-time flow field makes the solution process very time-consuming. Therefore, we present Physics-Informed Convolutional Neural Networks(PICNN) as a hybrid scientific theory and data method for reservoir modeling. Besides labeled data, the model is driven by the scientific theories of the underlying problem, such as governing equations, boundary conditions, and initial conditions. PICNN integrates governing equations and boundary conditions into the network architecture in the form of a customized convolution kernel. The loss function is composed of data matching, initial conditions, and other measurable prior knowledge. By customizing the convolution kernel and minimizing the loss function, the neural network parameters not only fit the data but also honor the governing equation. The PICNN provides a methodology to model and history-match flow and transport problems in porous media. Numerical results demonstrate that the proposed PICNN can provide an accurate physical solution from a limited dataset. We show how this method can be applied in the context of a forward simulation for continuous problems. Furthermore, several complex scenarios are tested, including the existence of data noise, different work schedules, and different good patterns.

Keywords: convolutional neural networks, deep learning, flow and transport in porous media, physics-informed neural networks, reservoir simulation

Procedia PDF Downloads 145

Authors: Lea Lenhardt, Ivana Zeković, Tatjana Dramićanin, Miroslav D. Dramićanin

Abstract:

Fluorescence spectroscopy coupled with parallel factor analysis (PARAFAC) and artificial neural networks (ANN) were used for characterization and classification of honey. Excitation emission spectra were obtained for 95 honey samples of different botanical origin (acacia, sunflower, linden, meadow, and fake honey) by recording emission from 270 to 640 nm with excitation in the range of 240-500 nm. Fluorescence spectra were described with a six-component PARAFAC model, and PARAFAC scores were further processed with two types of ANN’s (feed-forward network and self-organizing maps) to obtain algorithms for classification of honey on the basis of their botanical origin. Both ANN’s detected fake honey samples with 100% sensitivity and specificity.

Keywords: honey, fluorescence, PARAFAC, artificial neural networks

Procedia PDF Downloads 954

Authors: Stuti Modi, Aditi Kambli

Abstract:

This paper addresses the need for intelligent water management and distribution system in smart cities to ensure optimal consumption and distribution of water for drinking and sanitation purposes. Water being a limited resource in cities require an effective system for collection, storage and distribution. In this paper, applications of two mostly widely used particular types of data-driven models, namely artificial neural networks (ANN) and fuzzy logic-based models, to modelling in the water resources management field are considered. The objective of this paper is to review the principles of various types and architectures of neural network and fuzzy adaptive systems and their applications to integrated water resources management. Final goal of the review is to expose and formulate progressive direction of their applicability and further research of the AI-related and data-driven techniques application and to demonstrate applicability of the neural networks, fuzzy systems and other machine learning techniques in the practical issues of the regional water management. Apart from this the paper will deal with water storage, using ANN to find optimum reservoir level and predicting peak daily demands.

Keywords: artificial neural networks, fuzzy systems, peak daily demand prediction, water management and distribution

Procedia PDF Downloads 186

Authors: Apoorva Vinod, Archana Mathur, Snehanshu Saha

Abstract:

Though there exists a plethora of Activation Functions (AFs) used in single and multiple hidden layer Neural Networks (NN), their behavior always raised curiosity, whether used in combination or singly. The popular AFs –Sigmoid, ReLU, and Tanh–have performed prominently well for shallow and deep architectures. Most of the time, AFs are used singly in multi-layered NN, and, to the best of our knowledge, their performance is never studied and analyzed deeply when used in combination. In this manuscript, we experiment with multi-layered NN architecture (both on shallow and deep architectures; Convolutional NN and VGG16) and investigate how well the network responds to using two different AFs (Sigmoid-Tanh, Tanh-ReLU, ReLU-Sigmoid) used alternately against a traditional, single (Sigmoid-Sigmoid, Tanh-Tanh, ReLUReLU) combination. Our results show that using two different AFs, the network achieves better accuracy, substantially lower loss, and faster convergence on 4 computer vision (CV) and 15 Non-CV (NCV) datasets. When using different AFs, not only was the accuracy greater by 6-7%, but we also accomplished convergence twice as fast. We present a case study to investigate the probability of networks suffering vanishing and exploding gradients when using two different AFs. Additionally, we theoretically showed that a composition of two or more AFs satisfies Universal Approximation Theorem (UAT).

Keywords: activation function, universal approximation function, neural networks, convergence

Procedia PDF Downloads 158

Authors: Mst Shapna Akter, Hossain Shahriar

Abstract:

Neural network approaches are machine learning methods used in many domains, such as healthcare and cyber security. Neural networks are mostly known for dealing with image datasets. While training with the images, several fundamental mathematical operations are carried out in the Neural Network. The operation includes a number of algebraic and mathematical functions, including derivative, convolution, and matrix inversion and transposition. Such operations require higher processing power than is typically needed for computer usage. Central Processing Unit (CPU) is not appropriate for a large image size of the dataset as it is built with serial processing. While Graphics Processing Unit (GPU) has parallel processing capabilities and, therefore, has higher speed. This paper uses advanced Neural Network techniques such as VGG16, Resnet50, Densenet, Inceptionv3, Xception, Mobilenet, XGBOOST-VGG16, and our proposed models to compare CPU and GPU resources. A system for classifying autism disease using face images of an autistic and non-autistic child was used to compare performance during testing. We used evaluation matrices such as Accuracy, F1 score, Precision, Recall, and Execution time. It has been observed that GPU runs faster than the CPU in all tests performed. Moreover, the performance of the Neural Network models in terms of accuracy increases on GPU compared to CPU.

Keywords: autism disease, neural network, CPU, GPU, transfer learning

Procedia PDF Downloads 118

Authors: Shahzad Yousaf, Imran Shafi

Abstract:

This paper presents a new method for calculation of pressure loss coefficients by use of the artificial neural network (ANN) in tee junctions. Geometry and flow parameters are feed into ANN as the inputs for purpose of training the network. Efficacy of the network is demonstrated by comparison of the experimental and ANN based calculated data of pressure loss coefficients for combining flows in a tee junction. Reynolds numbers ranging from 200 to 14000 and discharge ratios varying from minimum to maximum flow for calculation of pressure loss coefficients have been used. Pressure loss coefficients calculated using ANN are compared to the models from literature used in junction flows. The results achieved after the application of ANN agrees reasonably to the experimental values.

Keywords: artificial neural networks, combining flow, pressure loss coefficients, solar collector tee junctions

Procedia PDF Downloads 391

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 142