Search

A neural-network-assisted numerical method is proposed for the solution of Laplace and Poisson problems. Finite differences are applied to approximate the spatial Laplacian operator on nonuniform grids. For this, a neural network is trained to compute the corresponding coefficients for general quadrilateral meshes. Depending on the position of a given grid point x 0 and its neighbors, we face with a nonlinear optimization problem to obtain the finite difference coefficients in x 0 . This computing step is executed with an artificial neural network. In this way, for any geometric setup of the neighboring grid points, we immediately obtain the corresponding coefficients. The construction of an appropriate training data set is also discussed, which is based on the solution of overdetermined linear systems. The method was experimentally validated on a number of numerical tests. As expected, it delivers a fast and reliable algorithm for solving Poisson problems. [ABSTRACT FROM AUTHOR]

Copyright of Modern Educational Technology is the property of Editorial Board of Modern Educational Technology, Tsinghua University and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Drawing from a larger CIPP program evaluation (Context, Input, Processes, Products) comprised of 11 analyses of the ETS online teacher training course titled Teaching Academic English with the TOEFL iBT® Test (TAE), this paper triangulates three qualitatively different data sets and seeks to understand how participants responded to the TAE course. The research team considered self-reported data (closed- and open-answer survey questions), engagement data, and learning data. The self-reported data were strongly positive and illustrated the value participants reported having about the course content, especially ETS-created materials. Some gaps were indicated: calls were made for more materials--especially additional resources participants could use when teaching--and for increased interactions during the course workshop. In partial juxtaposition, the engagement data indicated that participants were engaging the course less than anticipated. The learning data simultaneously indicated that some assessments were quite easy for participants while elsewhere participants did not have the desired uptake of some topics. Viewing the three data sets together, the research team arrived at a more nuanced conclusion: despite participant satisfaction, development processes should continue, to ensure the highest quality teacher training program possible, in terms of its content, resources, and interactions. [ABSTRACT FROM AUTHOR]

The article focuses on disturbing context in which so much is said about the so-called "artificial intelligence" and problematization of the question and his attempt to define the fallacies that, already in the fifties of the last century. It mentions sixty years impacts its epistemological relevance and its potential ethical warning and there is a serious risk of falling imperceptibly into a techno-totalitarian status quo and disturbing context so-called "artificial intelligence."

We propose neural-network-based algorithms for the numerical solution of boundary-value problems for the Laplace equation. Such a numerical solution is inherently mesh-free, and in the approximation process, stochastic algorithms are employed. The chief challenge in the solution framework is to generate appropriate learning data in the absence of the solution. Our main idea was to use fundamental solutions for this purpose and make a link with the so-called method of fundamental solutions. In this way, beyond the classical boundary-value problems, Dirichlet-to-Neumann operators can also be approximated. This problem was investigated in detail. Moreover, for this complex problem, low-rank approximations were constructed. Such efficient solution algorithms can serve as a basis for computational electrical impedance tomography. [ABSTRACT FROM AUTHOR]

Abstract This study analyzed libraries' management of the data generated by library digital learning objects (DLOs) such as forms, surveys, quizzes, and tutorials. A substantial proportion of respondents reported having a policy relevant to learning data, typically a campus-level policy, but most did not. Other problems included a lack of access to library learning data, concerns about student privacy, inadequate granularity or standardization, and a lack of knowledge about colleagues' practices. We propose more dialogue on learning data within libraries, between libraries and administrators, and across the library profession. [ABSTRACT FROM AUTHOR]

A neural-network-assisted numerical method is proposed for the solution of Laplace and Poisson problems. Finite differences are applied to approximate the spatial Laplacian operator on nonuniform grids. For this, a neural network is trained to compute the corresponding coefficients for general quadrilateral meshes. Depending on the position of a given grid point x0 and its neighbors, we face with a nonlinear optimization problem to obtain the finite difference coefficients in x0. This computing step is executed with an artificial neural network. In this way, for any geometric setup of the neighboring grid points, we immediately obtain the corresponding coefficients. The construction of an appropriate training data set is also discussed, which is based on the solution of overdetermined linear systems. The method was experimentally validated on a number of numerical tests. As expected, it delivers a fast and reliable algorithm for solving Poisson problems.

The aim of the Knowledge Acquisition Monitoring (KAM) approach created is to teach faster and/or more in a particular time frame. The KAM Question Framework that was designed according to the KAM approach ensures the active involvement of the students in the study process and provides them with relevant and updated content in real-time. The obtained data enables understanding the nature of knowledge acquisition - to improve the content, and enhance the efficiency of the learning process. 3-D coordinates were used to design the Knowledge Acquisition Surface (KAS). This surface describes the relevance of the e-course content to the needs of a student or group of students. The course suitability evaluation is performed considering the area in which the obtained knowledge acquisition data point is located. The approach provides the teacher with insight into the specific needs of students, and facilitates e-course improvement. [ABSTRACT FROM AUTHOR]

During the COVID-19 pandemic, most education has been conducted through online classes. While feedback and interaction between students and instructors are significant in programming education or engineering practice, online education today cannot satisfy these aspects of learning. Therefore, this study proposes a learning support system for programming education and presents the results of designing and implementing this system. The proposed system consists of an online development environment module, a learning monitoring module, and a learning support module. It also provides a web-based programming environment, real-time chat and code mirroring, error guide messages and related lectures, e-learning quizzes, and learning activity analysis features. The system standardizes the development environment between the instructor and students, helps students take the initiative in solving errors, and enables code-oriented interactions between the instructor and students. It also collects data from all learning situations in the database. Conducting a big data analysis with the collected data will enable individual guidance for students by finding errors that frequently occur in programming and recommending learning materials to solve them. [ABSTRACT FROM AUTHOR]

Acquiring actual data for developing data-based prediction models for the optimization of control setpoints (hereinafter referred to as optimal control (OC)) in heating, ventilation, and air conditioning (HVAC) systems is challenging, especially when changing control setpoints while considering all possible boundary conditions. Thus, considering the scalability of OC, it is a promising solution to generate sufficient learning data according to the control settings using a highly accurate simulation model, build a source model, and expand the learning model based on the actual data of the target system. This is the first study to improve the scalability of OC aiming to develop a deep neural network (DNN) model to predict the power consumption of a heat source system. First, this study will construct a calibrated heat source system simulation model and then use the generated learning data to train a DNN model. The actual operation conditions of the target building and heating system located in Vietnam were analyzed using operation data. A physics-based simulation model of the target heat source system was developed and calibrated using actual operation data to improve accuracy. By using the calibrated simulation, sufficient learning data was generated for all control settings allowing for the development of a learning model based on DNN that predicted the power consumption of the heat source system. In addition, a highly accurate prediction model through hyperparameter optimization was secured. [ABSTRACT FROM AUTHOR]

We propose neural-network-based algorithms for the numerical solution of boundary-value problems for the Laplace equation. Such a numerical solution is inherently mesh-free, and in the approximation process, stochastic algorithms are employed. The chief challenge in the solution framework is to generate appropriate learning data in the absence of the solution. Our main idea was to use fundamental solutions for this purpose and make a link with the so-called method of fundamental solutions. In this way, beyond the classical boundary-value problems, Dirichlet-to-Neumann operators can also be approximated. This problem was investigated in detail. Moreover, for this complex problem, low-rank approximations were constructed. Such efficient solution algorithms can serve as a basis for computational electrical impedance tomography.

Online education is growing because of its benefits and advantages that students enjoy. Educational technologies (e.g., learning analytics, student modelling, and intelligent tutoring systems) bring great potential to online education. Many online courses, particularly in self-paced online learning (SPOL), face some inherent barriers such as learning awareness and academic intervention. These barriers can affect the academic performance of online learners. Recently, learning analytics has been shown to have great potential in removing these barriers. However, it is challenging to achieve the full potential of learning analytics with the traditional online course learning design model. Thus, focusing on SPOL, this study proposes that learning analytics should be included in the course learning design loop to ensure data collection and pedagogical connection. We propose a novel learning design-analytics model in which course learning design and learning analytics can support each other to increase learning success. Based on the proposed model, a set of online course design strategies are recommended for online educators who wish to use learning analytics to mitigate the learning barriers in SPOL. These strategies and technologies are inspired by Jim Greer's work on student modelling. By following these recommended design strategies, a computer science course is used as an example to show our initial practices of including learning analytics in the course learning design loop. Finally, future work on how to develop and evaluate learning analytics enabled learning systems is outlined. [ABSTRACT FROM AUTHOR]

Nowadays, schools use many information systems to automate their activities for different stakeholders’ groups – learning management systems, student diary, library systems, digital repositories, financial management and accounting systems, document processing systems, etc. The huge amount of data generated by the users of these systems, led to increased interest in the collection and analysis of data to encourage students to achieve higher results, teachers to provide personalized support and school managers to make data-driven decisions at all levels of school, and stimulates research into the application of Learning Analytics (LA) in schools. The paper presents a LA model and a software prototype of the LA tool designed for the needs of Bulgarian school education from the perspective of different stakeholder groups (students, teachers, class teachers, parents, school managers, inspectors from evaluation agencies), aiming to improve school methods of approaching and analyzing learning data. The tool allows stakeholders to track data for students’ learning or training for different purposes, e.g. monitoring, analysis, forecast, intervention, recommendations, etc., but finally to improve the quality of learning and teaching processes. Research and experiments with the model and the LA tool under consideration are conducted based on the information infrastructure of a typical Bulgarian school. [ABSTRACT FROM AUTHOR]

During the COVID-19 pandemic, most education has been conducted through online classes. While feedback and interaction between students and instructors are significant in programming education or engineering practice, online education today cannot satisfy these aspects of learning. Therefore, this study proposes a learning support system for programming education and presents the results of designing and implementing this system. The proposed system consists of an online development environment module, a learning monitoring module, and a learning support module. It also provides a web-based programming environment, real-time chat and code mirroring, error guide messages and related lectures, e-learning quizzes, and learning activity analysis features. The system standardizes the development environment between the instructor and students, helps students take the initiative in solving errors, and enables code-oriented interactions between the instructor and students. It also collects data from all learning situations in the database. Conducting a big data analysis with the collected data will enable individual guidance for students by finding errors that frequently occur in programming and recommending learning materials to solve them.

Many educational institutions use a large number of information systems to automate their activities for different stakeholders' groups – learning management systems, student diary, library system, digital repositories, management system, etc. This leads to a significant increase in the volume and variety of data that can be captured, stored, and harnessed to improve student learning and school effectiveness. Educational institutions are realising that, with the help of technology, they are collecting data that could be very valuable when properly analysed, aligned with learning outcomes, and integrated into a tighter feedback loop with stakeholders. In addition, the analysis of data can help their managers to take data-driven decision making at all levels of educational institutions. The paper presents a comprehensive approach to Learning Analytics in the field of Secondary Education from the perspective of all different stakeholders, which aims to improve its methods of approaching and analysing learning data. On the basis of a literature review in the field and an investigation of requirements for quality evaluation of learning in school education, the corresponding stakeholder groups are identified - students, teachers, class teachers, managers, parents, inspectors and 6 models (1 model per each stakeholder group) for data collection and personalized and meaningful analysis are proposed for the needs of Learning Analytics. Each model consists of measurable indicators allowing the relevant stakeholder to track data for students' learning or training for different purposes, e.g. monitoring, analysis, forecast, intervention, recommendations, etc., but finally to improve the quality of learning and teaching processes. The proposed models are evaluated by the representatives of 4 stakeholder groups – students, teachers, class teachers, parents. [ABSTRACT FROM AUTHOR]

Learning methods are challenged when there is not enough labelled data. It gets worse when the existing learning data have different distributions in different domains. To deal with such situations, deep unsupervised domain adaptation techniques have newly been widely used. This study surveys such domain adaptation methods that have been used for classification tasks in computer vision. The survey includes the very recent papers on this topic that have not been included in the previous surveys and introduces a taxonomy by grouping methods published on unsupervised domain adaptation into five groups of discrepancy‐, adversarial‐, reconstruction‐, representation‐, and attention‐based methods. [ABSTRACT FROM AUTHOR]

In general, medical pattern is a collection of characters formed using the characters such as 'a', 'c', 'g' and 't'. The length of the pattern varies from one disease to another disease, and the pattern also seems to be different for different patients. Identifying an unusual pattern from the information design is troublesome a decade ago. In this paper, with a specific end goal to enhance the precision of the anomalous distinguishing pattern, an Artificial Immune System (AIS) is framed. In the present study, AIS is used to obtain the abnormal pattern by learning the characteristics of the entire data set. Due to the powerful and adaptive nature of AIS, detecting and identifying abnormal pattern is more accurate. This proposed idea is implemented in MATLAB software and experimented on DNA/RNA dataset and the performance is verified. [ABSTRACT FROM AUTHOR]

The user behavior data generated in the TELECI learning environment with additional short, easy-to-use multiple-choice questions before and after each content subunit are used for visualization and correlation analysis. Three user behavior data clusters were identified in data landscape. The student behavior change among the TELECI-clusters was used for TELECI learning support algorithm design. The student performance data before and after learning the e-content were used for knowledge acquisition model design. This model is based on the assumption that knowledge acquisition of real e-content can be quantified by superposition of the impact of learning “perfect” content, too easy content, and too complicated content. The learner knowledge acquisition surface is calculated on this assumption. The data of real course learner knowledge acquisition are located on this surface as “telecides”. Telecides are the visualization of the appropriateness of an e-content unit for the needs of the specific learner or learners target group. [ABSTRACT FROM AUTHOR]

The purpose of this study is to generate training data for failure detection by Neural Network, which estimates extent of damage from vibration data of meshing plastic gear pairs. Generally, measurement of meshing vibration with accelerometers installed on a gear system is not difficult. Therefore, one can automatically acquire vibration data under operation and obtain a large amount of vibration data. However, it is not easy to obtain a large amount of evaluation data of the damage extent. Evaluation of the damage extent is possible if the operation has stopped. It means that the procedure of this data acquisition requires additional work, and there is a need for acquiring data of the damage extent, automatically. This research addresses this need by developing an automatic data acquisition system of the meshing vibration and the extent of damage. In the developed system, an accelerometer installed on the housing of bearings is used for acquisition of vibration data, and a high-speed camera observes cracks occurring at the root of the target tooth. The measurement condition was discussed, and the effectiveness of the system was confirmed through experiments. Additionally, a damage index of the crack at the root was proposed, and the computed index was used as labels for training data. Finally, the measured meshing vibration data was visualized for generating training data of Convolutional Neural Network.

In this study, an adaptive continuous wavelet transform (CWT)‐based overcurrent protection for smart grids is proposed to enhance the overcurrent protection performance encountering high‐impedance fault (HIF) and current transformer (CT) saturation, which are extremely complex phenomena and their impacts often cause mis‐coordination or mal‐operation. The proposed algorithm samples three phase current waveforms and imports them to CWT to extract high frequency coefficients. Afterwards, the sum of absolute values of the coefficients, Scoef, is calculated for each sample during the last cycle. Meanwhile, several simulations related to HIFs with different impedances and CT saturations with different severities are executed and the fault currents and the sum of absolute values of the coefficients are achieved and saved in the relay memory as (X, Y) coordinates of points (IL‐fault, SL‐coef), named 'learning data'. Thereafter, each Scoef is imported to the X −Y plane and, consequently, the occurrence of HIF or CT saturation is detected and the real‐fault current is estimated by both non‐linear interpolation and extreme learning machine approaches. Subsequently, new time dial setting and Ipickup of the relays are computed and reloaded. Security, dependability, and sensitivity of the proposed adaptive protection method are confirmed by numerous simulation studies. [ABSTRACT FROM AUTHOR]

Learning analytics has been demonstrated as a great tool to evaluate quality of instructional design. This study examined how instructors can utilize learning analytics to improve online learning support design. 128 students from a southern university in United States participated in this study. All learning data related to learners and online learning environment was collected and analyzed. Results showed that learning analytics results were very helpful for instructors to improve online instructional support timing, adjust learning module design, and enhance learning materials accessibility. [ABSTRACT FROM AUTHOR]

As higher education increasingly relies on e-learning, the need for tools that will allow teachers themselves to develop effective e-learning objects as simply and quickly as possible has also been increasingly recognized. This article discusses the design and development of a novel tool, Enook (Evolutionary note book), for creating activity-based learning objects (LOs). This tool has three general foci: 1) creation of various activity-based objects for experimental learning or learning by doing while simultaneously offering as many learning resources as a real e-book that services as both a textbook and a notebook, dual functions that paperback books can never provide; 2) management of all aspects of live instructor-led classroom training in a network environment, such the Web and or a mobile network; 3) facilitated use of data on students' progress to monitor their understanding of the material and to provide other meaningful information through learning analytics. [ABSTRACT FROM AUTHOR]