Artificial Intelligence & Neural Networking

Speech recognition techniques are one of the most important modern technologies. Many different systems have been developed in terms of methods used in the extraction of features and methods of classification. Voice recognition includes two areas: speech recognition and speaker recognition, where the research is confined to the field of speech recognition.

Induction motors are widely used in transportation, mining, petrochemical, manufacturing and in almost every other field dealing with electrical power. These motors are simple, efficient, highly robust and rugged thus offering a very high degree of reliability. But like any other machine, they are vulnerable to faults, which if left unmonitored, might lead to catastrophic failure of the machine in the long run. On-line condition monitoring of the induction motors has been widely used in the detection of faults. This paper delves into the various faults and study of conventional and innovative techniques for induction motor faults with an identification of future research areas.

This paper examines the cobalt-doped ceria/reduced graphene oxide (Co-CeO2/rGO) nanocomposite as a supercapacitor and modeling of its cyclic voltammetry behavior using Artificial Neural Network (ANN) and Random Forest Algorithm (RFA). Good agreement was found between experimental results and the predicted values generated by using ANN and RFA. Simulation results confirmed the accuracy of the models, compared to measurements from supercapacitor module power-cycling. A comparison of the best performance between ANN and RFA models shows that the ANN models performed better (value of coefficient of determination >0.95) than the RFA models for all datasets used in this study. The results of the ANN and RFA models could be useful in designing the unique nanocomposites for supercapacitors and other strategies related with energy and the environment.

Forest fire causes serious damage to the Flora and fauna of a country. This is one of major environmental concern. Early prediction of fires saves large number of Flora and fauna and prevents the ecosystem. By predicting the area burnt we can also classify whether the fire into small or big. The key motivation for this prediction is to help fire management team in proper resource allocation and to help the fire-fighters in a best possible way. The meteorological conditions of the forest are the key factors of the forest fire. These climatic data can be obtained using the local sensors which are incorporated in the nearest meteorological stations. This research proposes various Machine learning approaches such as Naïve Bias, Decision Trees, SVR, Random forest, Stochastic Gradient Descent and Bagging for predicting the amount of land burnt in the forest. Here the predictive model is build using the outbreaks of fire caused in the northeast region of Portugal.

COVID-19 has changed the way we live, communicate and work, as well as altering our
feelings. The higher education sector, alongside other sectors, has been severely affected by the pandemic and its serious repercussions. Academic and teaching staff have had to work from home and convert to online teaching, a change which has been met with both negative and positive feelings. The need for new competencies and upskilling, among other challenges, has been encountered. Therefore, the objectives of this study are aligned with exploring the impact of three constructs—self efficacy,
positive feelings and negative feelings—on the performance of academic and teaching staff at public and private universities in Bahrain during the COVID-19 lock down. Additionally, the impact of self-efficacy on these feelings was explored. A cross-sectional quantitative survey instrument was developed, validated and distributed using 83 valid responses. A two-way approach was followed to evaluate the model using the partial least squares (PLS-SEM) and multi-layer perception-artificial
neural network (MLP-ANN) techniques. Tests support the validity, reliability and consistency of the measurement scale, as well as the validity of the postulated model. The results revealed a statistically significant relationship between the three constructs and performance. Interestingly, attention is drawn to the impact of self-efficacy on increasing positive feelings and task performance. The impact
of self-efficacy on reducing negative feelings is also evident. Analyses of PLS-SEM augmented by MLP-ANN enhanced our understanding of the relationships and gave more support to the use of dual approach analyses in future research. This research adds to COVID-19 global research and the findings increase the knowledge within the literature. The implications of the study’s outcomes should be given attention from higher education authorities and management to raise staff morale and offer training to help sustain performance and mental well being. Lasty, limitations and future
directions are discussed.

Implementations of mini hydro schemes with conventional hydraulic, electrical equipment's and controllers have proven very expensive and uneconomical. Many developing countries that are in need of rural electrification have encountered economical problem when setting up these mini hydro schemes. To address this problem, alternative options and new designs of these equipment's have been explored by many researchers around the world. The application of these new designs would reduce the overall cost of mini hydro development and would help in making it a cost effective technology. These new designs will also help developing countries to provide electricity to rural areas or remote regions where interconnection of transmission line from the electrical grid is uneconomical. The new designs can also be an enabling factor in boosting up electricity generation using a renewable energy source. This paper provides survey of all these alternative options and new designs in the controller, hydraulic turbine and generators that have been implemented in different countries of the world.

Accurate and fast islanding detection of distributed generation is highly important for its successful operation
in distribution networks. Up to now, various islanding detection technique based on communication,
passive, active and hybrid methods have been proposed. However, each technique suffers from
certain demerits that cause inaccuracies in islanding detection. Computational intelligence based techniques,
due to their robustness and flexibility in dealing with complex nonlinear systems, is an option
that might solve this problem. This paper aims to provide a comprehensive review of computational
intelligence based techniques applied for islanding detection of distributed generation. Moreover, the
paper compares the accuracies of computational intelligence based techniques over existing techniques
to provide a handful of information for industries and utility researchers to determine the best method
for their respective system.

Recently, process control in wastewater treatment plants (WWTPs) is, mostly accomplished through examining the quality of the water effluent and adjusting the processes through the operator's experience. This practice is inefficient, costly and slow in control response. A better control of WTPs can be achieved by developing a robust mathematical tool for performance prediction. Due to their high accuracy and quite promising application in the field of engineering, Artificial Neural Networks (ANNs) are attracting attention in the domain of WWTP predictive performance modeling. This work focuses on applying ANN with a feed-forward, back propagation learning paradigm to predict the effluent water quality of the Habesha brewery WTP. Data of influent and effluent water quality covering approximately an 11-month period (May 2016 to March 2017) were used to develop, calibrate and validate the models. The study proves that ANN can predict the effluent water quality parameters with a correlation coefficient (R) between the observed and predicted output values reaching up to 0.969. Model architecture of 3-21-3 for pH and TN, and 1-76-1 for COD were selected as optimum topologies for predicting the Habesha Brewery WTP performance. The linear correlation between predicted and target outputs for the optimal model architectures described above were 0.9201 and 0.9692, respectively.

The evaluation of deep eutectic solvents (DESs) as a new generation of solvents for various practical application requires an insight of the main physical, chemical, and thermodynamic properties. In this study, the experimental measurements of the electrical conductivity of two classes of DESs based on ammonium and phosphonium salts at different compositions and temperatures were reported. The results revealed that electrical conductivity of DESs has temperature-dependency. In addition, molar conductivities of ammonium and phosphonium salts in DESs were obtained using DESs experimental values of electrical conductivities. The feasibility of using an artificial neural network (ANN) model to predict the electrical conductivity of ammonium and phosphonium based DESs at different temperatures and compositions was also examined. A feed-forward back propagation neural network with 8 hidden neurons was successfully developed and trained with the measured electrical conductivity data. The results indicated that among the different networks tested, the network with 8 hidden neurons had the best prediction performance and gave the smallest value of Normalized Mean Square Error (NMSE) (0.0010) and acceptable values of Index of Agreement (IA) (0.9999) and Regression Coefficient (R 2 ) (0.9988). The comparison of the predicted electrical conductivity of DESs by the proposed model with those obtained by experiments confirmed the reliability of the ANN model with an average absolute relative deviation (AARD%) of 4.40%.

During the last decades, many cognitive architectures (CAs) have been realized adopting different assumptions about the organization and the representation of their knowledge level. Some of them (e.g. SOAR [35]) adopt a classical symbolic approach, some (e.g. LEABRA[48]) are based on a purely con-nectionist model, while others (e.g. CLARION [59]) adopt a hybrid approach combining connectionist and symbolic representational levels. Additionally, some attempts (e.g. biSOAR) trying to extend the representational capacities of CAs by integrating diagrammatical representations and reasoning are also available [34]. In this paper we propose a reflection on the role that Conceptual Spaces, a framework developed by Peter Gärdenfors [24] more than fifteen years ago, can play in the current development of the Knowledge Level in Cognitive Systems and Architectures. In particular, we claim that Conceptual Spaces offer a lingua franca that allows to unify and generalize many aspects of the symbolic, sub-symbolic and diagrammatic approaches (by overcoming some of their typical problems) and to integrate them on a common ground. In doing so we extend and detail some of the arguments explored by Gärdenfors [23] for defending the need of a conceptual, intermediate, representation level between the symbolic and the sub-symbolic one. In particular we focus on the advantages offered by Conceptual Spaces (w.r.t. symbolic and sub-symbolic approaches) in dealing with the problem of compositionality of representations based on typicality traits. Additionally, we argue that Conceptual Spaces could offer a unifying framework for interpreting many kinds of diagrammatic and analogical representations. As a consequence, their adoption could also favor the integration of diagrammatical representation and reasoning in CAs.

The best feature of artificial intelligent Feed Forward Neural Network (FFNN) classification models is learning of input data through their weights. Data preprocessing and pre-training are the contributing factors in developing efficient techniques for low training time and high accuracy of classification. In this study, we investigate and review the powerful preprocessing functions of the FFNN models. Currently initialization of the weights is at random which is the main source of problems. Multilayer auto-encoder networks as the latest technique like other related techniques is unable to solve the problems. Weight Linear Analysis (WLA) is a combination of data pre-processing and pretraining to generate real weights through the use of normalized input values. The FFNN model by using the WLA increases classification accuracy and improve training time in a single epoch without any training cycle, the gradient of the mean square error function, updating the weights. The results of comparison and evaluation show that the WLA is a powerful technique in the FFNN classification area yet.

The evaluation of deep eutectic solvents (DESs) as a new generation of solvents for various practical application requires an insight of the main physical, chemical, and thermodynamic properties. In this study, the experimental measurements of the electrical conductivity of two classes of DESs based on ammonium and phosphonium salts at different compositions and temperatures were reported. The results revealed that electrical conductivity of DESs has temperature-dependency. In addition, molar conductivities of ammonium and phosphonium salts in DESs were obtained using DESs experimental values of electrical conductivities. The feasibility of using an artificial neural network (ANN) model to predict the electrical conductivity of ammonium and phosphonium based DESs at different temperatures and compositions was also examined. A feed-forward back propagation neural network with 8 hidden neurons was successfully developed and trained with the measured electrical conductivity data. The results indicated that among the different networks tested, the network with 8 hidden neurons had the best prediction performance and gave the smallest value of Normalized Mean Square Error (NMSE) (0.0010) and acceptable values of Index of Agreement (IA) (0.9999) and Regression Coefficient (R 2 ) (0.9988). The comparison of the predicted electrical conductivity of DESs by the proposed model with those obtained by experiments confirmed the reliability of the ANN model with an average absolute relative deviation (AARD%) of 4.40%.

Recently, process control is, mostly, accomplished through examining the quality of the product water and adjusting the processes through an operator’s experience. This practice is inefficient, costly and slow in control response. A better control of WTPs can be achieved by developing a robust mathematical tool for predicting the performance. Owing to their high accuracy and quite promising application in the field of engineering, Artificial Neural Networks (ANNs) are getting attention in the predictive performance modeling of WTPs. This paper focus on applying ANN with a feed-forward, back propagation learning paradigm to predict the effluent water quality of Habesha Brewery’s WTP. About 11 months of data (from May 2016 to March 2017) of influent and effluent water quality were used to build, train and evaluate the models. The study signifies that ANN can predict the effluent water quality parameters with a correlation coefficient (R) between the observed and predicted output varia...

New design in propeller turbine Pump as turbine (PAT) Three-phase and six-phase induction generator Six-phase synchronous generator Intelligent controllers a b s t r a c t Implementations of mini hydro schemes with conventional hydraulic, electrical equipment's and controllers have proven very expensive and uneconomical. Many developing countries that are in need of rural electrification have encountered economical problem when setting up these mini hydro schemes. To address this problem, alternative options and new designs of these equipment's have been explored by many researchers around the world. The application of these new designs would reduce the overall cost of mini hydro development and would help in making it a cost effective technology. These new designs will also help developing countries to provide electricity to rural areas or remote regions where interconnection of transmission line from the electrical grid is uneconomical. The new designs can also be an enabling factor in boosting up electricity generation using a renewable energy source. This paper provides survey of all these alternative options and new designs in the controller, hydraulic turbine and generators that have been implemented in different countries of the world.

The evaluation of deep eutectic solvents (DESs) as a new generation of solvents for various practical application requires an insight of the main physical, chemical, and thermodynamic properties. In this study, the experimental measurements of the electrical conductivity of two classes of DESs based on ammonium and phosphonium salts at different compositions and temperatures were reported. The results revealed that electrical conductivity of DESs has temperature-dependency. In addition, molar conductivities of ammonium and phosphonium salts in DESs were obtained using DESs experimental values of electrical conductivities. The feasibility of using an artificial neural network (ANN) model to predict the electrical conductivity of ammonium and phosphonium based DESs at different temperatures and compositions was also examined. A feed-forward back propagation neural network with 8 hidden neurons was successfully developed and trained with the measured electrical conductivity data. The results indicated that among the different networks tested, the network with 8 hidden neurons had the best prediction performance and gave the smallest value of Normalized Mean Square Error (NMSE) (0.0010) and acceptable values of Index of Agreement (IA) (0.9999) and Regression Coefficient (R2) (0.9988). The comparison of the predicted electrical conductivity of DESs by the proposed model with those obtained by experiments confirmed the reliability of the ANN model with an average absolute relative deviation (AARD%) of 4.40%. © 2013 Elsevier B.V.

http://ac.els-cdn.com/S0378381213003518/1-s2.0-S0378381213003518-main.pdf?_tid=7fb4f1be-0ca1-11e4-9941-00000aacb35e&acdnat=1405485030_8b454bad8f3dd84f4ff2ad7da2ccafca 
http://www.sciencedirect.com/science/article/pii/S0378381213003518

Guided-based systems are an alternative to natural language systems regarding the query construction method. We present a new solution that bridges the gap between these two types of systems by providing a hybrid interface which combines a graphical query interface for query construction with the correspondent natural language expression automatically generated by the system.

The main problem for Supervised Multi-layer Neural Network (SMNN) model such
as Back propagation network lies in finding the suitable weights during training in order to
improve training time as well as achieve high accuracy. The important issue in the training
process of the existing SMNN model is initialization of the weights which is random and
creates paradox, and leads to low accuracy with high training time. In this paper, a new
Supervised Feed Forward Multi-layer Neural Network (SFFMNN) model for classification
problem is proposed. It consists of a new preprocessing technique which combines data
preprocessing and pre-training that offer a number of advantages; training cycle, gradient
of mean square error function, and updating weights are not needed in this model. In new
SMFFNN model, thresholds of training set and test set are computed by using input values
and potential weights. In training set each instance has one special threshold and class
label. In test set the threshold of each instance will be compared with the range of
thresholds of training set and the class label of each instance will be predicted. To evaluate
the performance of the proposed SMFFNN model, a series of experiment on XOR problem
and two datasets, which are SPECT Heart and SPECTF Heart was implemented with 10-
fold cross-validation. As quoted by literature, these two datasets are difficult for
classification and most of the conventional methods do not process well on these datasets.
Our results, however, show that the proposed model has been given high accuracy in one
epoch without training cycle.

Application of Artificial Neural Network Committee Machine (ANNCM) for the inversion of magnetic
anomalies caused by a long-2D horizontal circular cylinder is presented. Although, the subsurface targets
are of arbitrary shape, they are assumed to be regular geometrical shape for convenience of mathematical
analysis. ANNCM inversion extract the parameters of the causative subsurface targets include depth to the
centre of the cylinder (Z), the inclination of magnetic vector(Ɵ)and the constant term (A)comprising the
radius(R)and the intensity of the magnetic field(I). The method of inversion is demonstrated over a
theoretical model with and without random noise in order to study the effect of noise on the technique and
then extended to real field data. It is noted that the method under discussion ensures fairly accurate results
even in the presence of noise. ANNCM analysis of vertical magnetic anomaly near Karimnagar, Telangana,
India, has shown satisfactory results in comparison with other inversion techniques that are in vogue.The
statistics of the predicted parameters relative to the measured data, show lower sum error (<9.58%) and
higher correlation coefficient (R>91%) indicating that good matching and correlation is achieved between
the measured and predicted parameters.

In recent years, there is a significant increase in the number of devices with low power consumption. The energy requirements of these devices are provided by chemical batteries. The batteries must be charged at regular times, and cause some problems such as environmental pollution. RF energy harvesters are an alternative energy source for the batteries. In this study, the responses of 900 MHz RF energy harvester, which was previously tested, are estimated using an Artificial Neural Network (ANN) method in different states. For this aim, the output power values are determined by using the input power and the frequency of the signal and the load resistances connected to the energy harvester.

Learning is the important property of Back Propagation Network (BPN) and finding the suitable weights and thresholds during training in order to improve training time as well as achieve high accuracy. Currently, data pre-processing such as dimension reduction input values and pre-training are the contributing factors in developing efficient techniques for reducing training time with high accuracy and initialization of the weights is the important issue which is random and creates paradox, and leads to low accuracy with high training time. One good data preprocessing technique for accelerating BPN classification is dimension reduction technique but it has problem of missing data. In this paper, we study current pre-training techniques and new preprocessing technique called Potential Weight Linear Analysis (PWLA) which combines normalization, dimension reduction input values and pre-training. In PWLA, the first data preprocessing is performed for generating normalized input values and then applying them by pre-training technique in order to obtain the potential weights. After these phases, dimension of input values matrix will be reduced by using real potential weights. For experiment results XOR problem and three datasets, which are SPECT Heart, SPECTF Heart and Liver disorders (BUPA) will be evaluated. Our results, however, will show that the new technique of PWLA will change BPN to new Supervised Multi Layer Feed Forward Neural Network (SMFFNN) model with high accuracy in one epoch without training cycle. Also PWLA will be able to have power of non linear supervised and unsupervised dimension reduction property for applying by other supervised multi layer feed forward neural network model in future work.

Learning is the important property of Back Propagation Network (BPN) and finding the suitable weights and thresholds during training in order to improve training time as well as achieve high accuracy. Currently, data pre-processing such as dimension reduction input values and pre-training are the contributing factors in developing efficient techniques for reducing training time with high accuracy and initialization of the weights is the important issue which is random and creates paradox, and leads to low accuracy with high training time. One good data preprocessing technique for accelerating BPN classification is dimension reduction technique but it has problem of missing data. In this paper, we study current pre-training techniques and new preprocessing technique called Potential Weight Linear Analysis (PWLA) which combines normalization, dimension reduction input values and pre-training. In PWLA, the first data preprocessing is performed for generating normalized input values and then applying them by pre-training technique in order to obtain the potential weights. After these phases, dimension of input values matrix will be reduced by using real potential weights. For experiment results XOR problem and three datasets, which are SPECT Heart, SPECTF Heart and Liver disorders (BUPA) will be evaluated. Our results, however, will show that the new technique of PWLA will change BPN to new Supervised Multi Layer Feed Forward Neural Network (SMFFNN) model with high accuracy in one epoch without training cycle. Also PWLA will be able to have power of non linear supervised and unsupervised dimension reduction property for applying by other supervised multi layer feed forward neural network model in future work.

In recent years, there is a significant increase in the number of devices with low power consumption. The energy requirements of these devices are provided by chemical batteries. The batteries must be charged at regular times, and cause some problems such as environmental pollution. RF energy harvesters are an alternative energy source for the batteries. In this study, the responses of 900 MHz RF energy harvester, which was previously tested, are estimated using an Artificial Neural Network (ANN) method in different states. For this aim, the output power values are determined by using the input power and the frequency of the signal and the load resistances connected to the energy harvester.

This paper is in continuation to what we did previously in the field of Soft Computing. In the previous workthe author proposed several solutions to the XOR problem. In this paper we will see some solutions to theEx-NOR problem using the concept of MLFFN. Ex-NOR problem is a non-linearly separable problem thesolutions of which are discussed in this paper. The Artificial Neural Network proposed is mathematicallyproved to be successful in providing solution to the Ex-NOR problem. The solutions proposed areexplained by means of Architectural Graph and Signal flow graph.