SAM-DNN: Bilevel convergent sequential averaging methods with denoising neural network

We propose a robust way of achieving a well test interpretation by combining the sequential predictive probability method with an artificial neural network approach. The sequential predictive probability method considers all possible reservoir models and determines which candidate model or models best predict the well response. This method is dependent on obtaining good initial estimates for the parameters governing the candidate reservoir models, which is achieved by applying the artificial neural network approach. We use the neural network to identify the characteristic components of the pressure derivative curve corresponding to the flow regimes known to be in each candidate model. Reservoir parameters are then computed using the data in the identified range of the corresponding behavior. As a final step, the candidate models and their initial estimates are evaluated using the sequential probability method. The method discriminates between the candidate models and simultaneously performs nonlinear regression to compute the best estimates of reservoir parameters. The trained neural network was able to identify the characteristic components of the derivative curve in most cases. The algorithm written to interpret the neural network signals into flow regimes required special procedures to take care of the misclassification from the neural network. The initial estimates of reservoir parameters from the neural network were found to be reasonably close to the eventual estimates from the sequential predictive probability method.

This study investigates the advancement of underrepresented groups in leadership roles within the fields of Science, Technology, Engineering, and Mathematics (STEM), areas traditionally dominated by their male counterparts. Through a convergent sequential mixed method, data was collected using 105 questionnaires and six semi-structured interviews involving students, lecturers, government officials, and university admissions administrators in the STEM disciplines. A concurrent triangulation strategy was applied for simultaneous quantitative and qualitative data analysis, utilizing MS Excel for the numerical data and thematic analysis for the narrative information. The research also examined statistical documents to assess the enrollment disparity among underrepresented groups and the presence of specific development programs in universities aimed at addressing this imbalance. Findings suggest that targeted and culturally informed collaborative efforts can effectively encourage the ascent of underrepresented demographics into STEM leadership roles in South African universities. The implications of the study are broad, given that the research spanned two distinct universities, implying that the strategies identified could inform actions across different educational institutions to foster inclusivity and equality in STEM leadership.

This dissertation explores the effectiveness of a mobile-assisted vocabulary learning (MAVL) application, Shanbay Danci (SBDC), in enhancing second language (L2) learners’ vocabulary learning outcomes, self-regulated learning (SRL) skills, perceived social presence during the learning experience, and learning persistence through a series of three interconnected studies. Grounded in frameworks such as the iPAC (Personalization, Authenticity, Collaboration) framework, GAFCC (Goal, Access, Feedback, Challenge, Collaboration) model, and social presence theory, the studies examine how mobile technology, social media, and gamification features can influence learners’ engagement and persistence in informal language learning contexts.Study 1 employed an explanatory sequential mixed methods design to investigate whether SBDC supports vocabulary form-meaning linkage (VFML) and to examine learners’ perceptions of mobile vocabulary learning. While the experimental group showed gains in VFML, learners also expressed both enthusiasm and concerns regarding mobile learning features such as personalization, collaboration, and attention challenges. Study 2 adopted a quasi-experimental design to test the impact of a badge-sharing gamification feature embedded in SBDC on learners’ SRL and persistence. Findings revealed that badge-sharing significantly enhanced SRL skills, though it did not lead to significant increases in vocabulary learning persistence. Learners had mixed emotional responses to the social media-based activity, suggesting both motivational and stress-inducing effects. Study 3 utilized a convergent parallel mixed methods design to explore the effect of a collaborative “learning-pal” feature on learners’ vocabulary learning grit and perceived social presence. Results indicated that pairing learners to track mutual progress significantly improved grit-related learning behaviors. However, contrary to expectations, the learning-pal activity led to lower perceived social presence, highlighting the critical role of instructor facilitation in fostering meaningful social interaction in technology-enhanced environments.Together, these studies suggest that MAVL tools, particularly when enriched with gamification and collaborative features, can support vocabulary acquisition and learner persistence and grit. However, the effectiveness of these tools depends not only on technological design but also on the use of intentional pedagogical strategies by instructors.

Violent crime such as robbery, assault, murder, sexual and domestic violence, undermine the economic and social rights of the people and distorts society’s attempts to achieve a decent and dignified life for all. To reverse the situation, in Ghana, the police have been charged with the primary responsibility of maintaining domestic security. The emerging development in the country, however, has been the blame game between the police and the public in respect of the perceived failure in security governance for the citizenry. This study examined this blame game using Sawla-Tuna-Kalba District of Savannah Region as a case study. To be able to achieve the intended objective, 102 respondents were selected through both probability and non-probability sampling techniques in a convergent sequential mixed method design. Expert administered questionnaire and interviews were employed as the main methods of data collection. The study found that non-prosecution of criminals, bribery and corruption, police-criminal relationship, inadequate logistics and police personnel and stakeholders’ interference in security management, cumulatively were responsible for the low level of security administration in the study area. The study concluded that both the police and the public are responsible for the defective security architecture in the study community. Since crime impacts multiple facets of society in ways that the police and government may not be aware of until those interests are taken into consideration during project development and implementation, the study recommends a multi-sector collaboration involving the government, the police and the public to improve security governance in the area.
 Keywords: Crime, Violence, Security, Police, Citizenry, Peace Architecture.

This study proposed signal processing techniques and neural network with weighted fuzzy membership functions(NEWFM) to detect epileptic seizure from EEG signals. This study used wavelet transform(WT), sequential increment method, and phase space reconstruction(PSR) as signal processing techniques. In the first step of signal processing techniques, wavelet coefficients were extracted from EEG signals using the WT. In the second step, sequential increment method was used to extract peaks from the wavelet coefficients. In the third step, 3D diagram was produced from the extracted peaks using the PSR. The Euclidean distances and statistical methods were used to extract 16 features used as inputs for NEWFM. The proposed methodology shows that accuracy, specificity, and sensitivity are 97.5%, 100%, 95% with 16 features, respectively.

Neural network models, which were initially used to model the processes of recognition of graphic images, are now widely used in the field of recognition of multi-parameter objects and in adjusting the parameters of complex systems. Such an application is effective in solving problems of information security management. With the help of neural networks, the tasks of classifying threats, choosing the parameters of protection measures, and adjusting the operating modes of information systems are successfully solved. In the scientific literature, there is a wide variety of approaches and methods for creating artificial neural networks. At the initial stage of designing information security systems, when it is necessary to determine the main parameters of the software being created, it poses a problem for the designer. Miscalculations made at this stage of the design can lead to the unsuccessful completion of the entire project. The article provides a brief comparative analysis of artificial neural networks that differ in the methods of setting (training) the network and in the form of its structure. The applicability, advantages, and disadvantages of such methods for tuning neural networks are characterized. These are the method of backpropagation of an error, genetic algorithm, iterative Widrow-Hoff algorithm with variable step, modified least squares method, and sequential learning method. The differences in the structure for networks designed to solve the problems of adjusting information processes and recognizing multi-parameter objects are shown. To regulate the processes, the structure of the neural network is studied on the example of the fuzzy neural network ANFIS. For the classification problems, the structure of a multilayer perceptron is given, in which the structure of the inner layers reflects the ontological network of the subject area under consideration. The presented results can be used to justify the type of neural network used for a specific task in the field of information security.

The FE computational homogenization method is a predictive multi‐scale method without the need for constitutive assumptions and/or potential function postulates at the macro engineering scale. Instead, the effective micro‐structural responses are extracted directly from a representative volume element (RVE) underlying each macro point. However, the FE method is still computationally too expensive for most practical uses, since the micro‐macro FE coupling is done at each loading step/iteration for the entire domain. To this end, the machine learning method has been utilized in the literature for the offline training of a surrogate model to predict the RVE homogenized response for general loading conditions. In this contribution, the neural network (NN) is incorporated into the macro finite element framework in a non‐intrusive manner. This is termed as the FE‐NN framework, in analogy to the FE method. In general, online simulations in the FE‐NN method is very efficient, with predictions matching closely to those obtained from reference direct numerical simulations (DNS). A bottleneck with the FE‐NN framework, however, is the high computational cost associated with the data generation for offline NN model setup. In this paper, focusing on the FE‐NN multi‐scale framework for non‐linear elastic deformation of heterogeneous materials, a sequential training strategy with knowledge transfer is proposed, to enable an efficient offline microscopic NN model setup. For a given target RVE, we first consider a simplified source RVE, where data can be generated rapidly, for the NN pre‐training of surrogate model. The pre‐trained network parameters are next downloaded to initialize the target NN surrogate model, followed by a fine‐tuning training process, using only a small dataset generated by the computationally expensive high‐fidelity RVE. The efficiency of the proposed sequential learning method over the conventional NN training, as well as, its excellent predictive capability for multi‐scale analyses, are demonstrated for a multi‐phase composite material. The proposed FE‐NN‐KT approach can be implemented easily without complicated pre‐processing procedures, since the data generation only involves the standard extraction of effective responses from RVEs, together with a standard NN architecture.

A general treatment of a particular class of learning techniques for neural networks, called sequential constructive methods, is proposed. They subsequently add units to the hidden layer until all the input-output relations contained in a given training set are satisfied.Every addition involves the update of a small portion of the whole weight matrix and depends on a subset of samples whose size decreases with time. In most cases this leads to a large reduction of the computational cost.General convergence theorems are presented that ensure the achievement of a good multilayer perceptron within a finite execution time. The output weights need not to be trained but are obtained by the application of simple algebraic equations.KeywordsSupervised learningconstructive methodssequential learningconvergence theoremsmultilayer perceptron

This thesis focuses on the development and applications of efficient sequential and batch learning artificial neural network methods for classification problems. Emphasis is on applications in the bio-informatics area where the problems have a very high input dimension and a small number of samples. Here, by "efficient methods", we imply those methods that produce more accurate classification, low training time and a compact network structure.

Radial basis function neural network (RBFNN) is widely used in nonlinear function approximation. One of the key issues in RBFNN modeling is to improve the approximation ability with samples as few as possible, so as to limit the network’s complexity. To solve this problem, a gradient-based sequential RBFNN modeling method is proposed. This method can utilize the gradient information of the present model to expand the sample set and refine the model sequentially, so as to improve the approximation accuracy effectively. Two mathematical examples and one practical problem are tested to verify the efficiency of this method.

Reliability is a critical indicator for evaluating the safety and performance of mechanical structures. Surrogate models are widely employed in reliability analysis to reduce computational costs; however, their accuracy strongly depends on the quality of training samples. This paper proposes an uncertainty and distance-driven sequential sampling deep neural network method (UDDS-DNN) for structural reliability analysis. First, initial training samples were generated from the Monte Carlo population (MCP) using the Minimax sampling strategy. This ensured global coverage of the random variable space. Then, the predictive uncertainty of the DNN was quantified using the Jackknife-based method. The estimated uncertainty, inter-sample distance, and probabilistic characteristics of the input random variables were considered. The adaptive learning function was constructed to guide sequential sample enrichment. The proposed method was validated through two numerical examples and three engineering case studies. Results demonstrated that UDDS-DNN achieves high accuracy and computational efficiency in multivariate, nonlinear, and complex structural reliability problems. The method provided an effective solution for reliability analysis of complex structures using DNN-based sequential sampling.

In hyperspectral image processing, classification is one of the most popular research topics. In recent years, research progress made in deep-learning-based hierarchical feature extraction and classification has shown a great power in many applications. In this paper, we propose a novel local spatial sequential (LSS) method, which is used in a recurrent neural network (RNN). Using this model, we can extract local and semantic information for hyperspectral image classification. First, we extract low-level features from hyperspectral images, including texture and differential morphological profiles. Second, we combine the low-level features together and propose a method to construct the LSS features. Afterwards, we build an RNN and use the LSS features as the input to train the network for optimizing the system parameters. Finally, the high-level semantic features generated by the RNN is fed into a softmax layer for the final classification. In addition, a nonlocal spatial sequential method is presented for the recurrent neural network model (NLSS-RNN) to further enhance the classification performance. NLSS-RNN finds nonlocal similar structures to a given pixel and extracts corresponding LSS features, which not only preserve the local spatial information, but also integrate the information of nonlocal similar samples. The experimental results on three publicly accessible datasets show that our proposed method can obtain competitive performance compared with several state-of-the-art classifiers.

Personalized search aims to improve the search quality by re-ranking the candidate document list based on user's historical behavior. Existing approaches focus on modeling the order information of user's search history by sequential methods such as Recurrent Neural Network (RNN). However, these methods usually ignore the fine-grained time information associated with user actions. In fact, the time intervals between queries can help to capture the evolution of query intent and document interest of users. Besides, the time intervals between past actions and current query can reflect the re-finding tendency more accurately than discrete steps in RNN. In this paper, we propose PSTIE, a fine-grained Time Information Enhanced model to construct more accurate user interest representations for Personalized Search. To capture the short-term interest of users, we design time-aware LSTM architectures for modeling the subtle interest evolution of users in continuous time. We further leverage time in calculating the re-finding possibility of users to capture the long-term user interest. We propose two methods to utilize the time-enhanced user interest into personalized ranking. Experiments on two datasets show that PSTIE can effectively improve the ranking quality over state-of-the-art models.

A neural network algorithm for selecting paths through a three-stage switching network was developed. Based on the results of computer simulations, a four-dimensional region for the valid combinations was discovered. An analysis is performed to determine the characteristics of this region. The behavior of the neural network algorithm for different switching network configurations and varying traffic patterns was investigated. The effect of initial state of the neural network is analyzed and heuristic improvements to the algorithm are provided. A comparative analysis of the neural network path selection algorithm against a sequential search method is also given. >

In this paper, a method to develop an interactive application in order to detect objects from videos is proposed. The application is able to classify the video according to a particular genre. Also, upon user input, it is also able to detect the particular object being shown at that instant on the screen. A sequential frame extraction method of videos and also deep learning approach of Convolutional Neural Networks along with Fully Connected Neural Networks is used for this task. The method gives good accuracy of average 77 percent.