This paper focuses on a method to train a regression model from incomplete input values. It is assumed in this paper that there are no missing values in a training dataset while missing values exist during a prediction phase using the trained model. Under this assumption, we propose Intentional-Value-Substitution (IVS) training to obtain a machine learning model that makes the prediction error as minimum as possible. In IVS training, a model is trained to approximate the target function using a modified training dataset in which some feature values are substituted with a certain value even though their values are not missing. It is shown through a series of computational experiments that the substitution values calculated from a mathematical analysis help the models correctly predict outputs for inputs with missing values.

This study presents a repetitive control method based on a multi-stage particle swarm optimization (PSO) algorithm with variable intervals to enhance the tracking performance of Takagi–Sugeno (T–S) fuzzy systems. First, a T–S fuzzy model is used to describe a nonlinear system. A modified repetitive control structure with two repetitive loops guarantees the tracking accuracy of periodic signals. Taking advantage of the two-dimensional (2D) property with continuous control and discrete learning, a continuous-discrete 2D model is presented to describe the nonlinear repetitive control system. Next, a multi-stage PSO algorithm with variable intervals searches for the best parameter combination in the linear matrix inequality-based stability condition to regulate the control and learning actions, which avoids a suboptimal solution and guarantees high control accuracy. Finally, an application to control the speed of synchronous motor with a permanent magnet demonstrates the validity of the method, and comparisons with related methods show its superiority.

Cloud storage technology has attracted a considerable number of users owing to its exponential growth. Storing data in the cloud can save the resources of local storage configuration and reduce the cost of local hardware investment. However, the data stored in the cloud is out of the physical control (out of control) of the user. Based on the service characteristics of the cloud environment and the security requirements of user privacy data in the cloud environment, this paper proposes an improved identity proxy re-encryption algorithm based on the advanced encryption standard algorithm. The performance of the algorithm is optimized by reducing the number of bilinear mapping operations, whose calculation takes the longest time in the proxy re-encryption scheme. Only two bilinear mapping operations are required in this scheme. In addition, the encrypted data are tested to different degrees. The experimental results show that this scheme satisfies encryption and decryption performance requirements of the user.

Genetic Programming (GP) is an Evolutionary Computation (EC) algorithm. Controlling genetic diversity in GP is a fundamental requirement to obtain various types of local minima effectively; however, this control is difficult compared to other EC algorithms because of difficulties in measuring the similarity between solutions. In general, common subtrees and the edit distance between solutions is used to evaluate the similarity between solutions. However, there are no clear guidelines regarding the best features to evaluate it. We hypothesized that the combination of multiple features helps to express the specific genetic similarity of each solution. In this study, we propose a self-organized subpopulation model based on similarity in terms of multiple features. To reconstruct subpopulations, we introduce a novel weighted network based on each normalized feature and utilize network clustering techniques. Although we can regard similarity as a correlation network between solutions, the use of multiple features incurs high computational costs, however, calculating the similarity is very suitable for parallelization on GPUs. Therefore, in the proposed method, we use CUDA to reconstruct subpopulations. Using three benchmark problems widely adopted in studies in the literature, we demonstrate that performance improvement can be achieved by reconstructing subpopulations based on a correlation network of solutions, and that the proposed method significantly outperforms typical methods.

Captions, which are used as a means of information support for deaf and hard-of-hearing students, are usually presented only in text form. Therefore, when mathematical equations and figures are frequently used in a classroom, the captions will show several demonstrative words such as “this equation” or “that figure.” As there is a delay between the teacher’s utterance and the display of the captions sometimes, it is difficult for users to grasp the target of these demonstrative words accurately. In this study, we prepared hybrid captions with mathematical equations and figures and verified their effectiveness via a comparison with conventional text-only captions. The results suggested that the hybrid captions were at least as effective as the conventional captions at helping the students understand the lesson contents. A subjective evaluation with a questionnaire survey also showed that the experimental participants found the hybrid captions to be acceptable without any discomfort. Furthermore, there was no difference in the number of eye movements of the participants during the experiment, suggesting that the physical load was similar for both types of captions.

Focusing on the issue of missing measurement data caused by complex and changeable working conditions during the operation of high-speed trains, in this paper, a framework for the reconstruction of missing measurement data based on a generative adversarial network is proposed. Suitable parameters were set for each frame. Discrete measurement data are taken as the input of the frame for preprocessing the data dimensionality. The convolutional neural network then learns the correlation between different characteristic values of each device in an unsupervised pattern and constrains and improves the reconstruction accuracy by taking advantage of the context similarity of authenticity. It was determined experimentally that when there are different extents of missing measurement data, the model described in the present paper can still maintain a high reconstruction accuracy. In addition, the reconstruction data also conform well to the distribution law of the measurement data.

This paper introduces an improved algorithm for texture-less object detection and pose estimation in industrial scenes. In the template training stage, a multi-scale template training method is proposed to improve the sensitivity of LineMOD to template depth. When this method performs template matching, the test image is first divided into several regions, and then training templates with similar depth are selected according to the depth of each test image region. In this way, without traversing all the templates, the depth of the template used by the algorithm during template matching is kept close to the depth of the target object, which improves the speed of the algorithm while ensuring that the accuracy of recognition will not decrease. In addition, this paper also proposes a method called coarse positioning of objects. The method avoids a lot of useless matching operations, and further improves the speed of the algorithm. The experimental results show that the improved LineMOD algorithm in this paper can effectively solve the algorithm’s template depth sensitivity problem.

A method is proposed for evaluating fuzzy rules independently of each other in fuzzy rules learning. The proposed method is named α-FUZZI-ES (α-weight-based fuzzy-rule independent evaluations) in this paper. In α-FUZZI-ES, the evaluation value of a fuzzy system is divided out among the fuzzy rules by using the compatibility degrees of the learning data. By the effective use of α-FUZZI-ES, a method for fast fuzzy rules learning is proposed. This is named α-FUZZI-ES learning (α-FUZZI-ES-based fuzzy rules learning) in this paper. α-FUZZI-ES learning is especially effective when evaluation functions are not differentiable and derivative-based optimization methods cannot be applied to fuzzy rules learning. α-FUZZI-ES learning makes it possible to optimize fuzzy rules independently of each other. This property reduces the dimensionality of the search space in finding the optimum fuzzy rules. Thereby, α-FUZZI-ES learning can attain fast convergence in fuzzy rules optimization. Moreover, α-FUZZI-ES learning can be efficiently performed with hardware in parallel to optimize fuzzy rules independently of each other. Numerical results show that α-FUZZI-ES learning is superior to the exemplary conventional scheme in terms of accuracy and convergence speed when the evaluation function is non-differentiable.

Several conventional clustering methods use the squared L₂-norm as the dissimilarity. The squared L₂-norm is calculated from only the object coordinates and obtains a linear cluster boundary. To extract meaningful cluster partitions from a set of massive objects, it is necessary to obtain cluster partitions that consisting of complex cluster boundaries. In this study, a JS-divergence-based k-medoids (JSKMdd) is proposed. In the proposed method, JS-divergence, which is calculated from the object distribution, is considered as the dissimilarity. The object distribution is estimated from kernel density estimation to calculate the dissimilarity based on both the object coordinates and their neighbors. Numerical experiments were conducted using five artificial datasets to verify the effectiveness of the proposed method. In the numerical experiments, the proposed method was compared with the k-means clustering, k-medoids clustering, and spectral clustering. The results show that the proposed method yields better results in terms of clustering performance than other conventional methods.

In the context of developing technologies for realizing a user-centric smart society, robot technology is gaining importance for responding to safety issues such as for those living alone and elderly persons. Therefore, in recent years, various robots have been developed to perform social exchanges with people in daily life. We also aim to develop a support system that can be easily used in everyday life through the application of smart device technology that is familiar to people. Therefore, in this paper, we discuss the process of developing robot partners according to various user needs, from the viewpoint of hardware and software development, as human coexistence robot partners. In addition, we show an example of the scalability and application of robot technology using smart devices. First, we describe our smart device-based robot partner system. Next, we describe the development of a robot partner comprising various modules. Finally, we present several examples of robot systems for social implementation and address the applicability of our proposed system.

Although sunlight is a promising renewable energy source, the light is incoherent and difficult to use directly. Therefore, a solar-pumped laser, which directly converts sunlight into coherent laser light of, is a promising technology. A solar-pumped laser collects sunlight into the laser medium to realize laser oscillation. In order to realize an efficient solar-pumped laser system, it is necessary to design a pumping cavity that absorbs maximal sunlight into the laser medium with minimal thermal shock. In this research, the pumping cavity shape was studied using a numerical ray tracing simulation. As a result, it was found that a cone shaped pumping cavity can be expected to improve the absorption rate by approximately 30% over a cylindrically shaped pumping cavity. Furthermore, the absorption power density distribution can be flattened by a vase shaped pumping cavity, while maintaining the same absorption efficiency. The vase shaped pumping cavity has almost half the dispersion of the absorbed power density in the laser medium when compared with the cone shaped pumping cavity.

In recent years, there have been several breakthroughs in the theoretical research of servo control algorithms. However most of these control algorithms remain in the simulation stage. They are difficult to be applied directly to practical platforms or complex industrial sites because of the lack of an experimental system suitable for the verification of their effectiveness. To address this problem, we designed a multi-function servo control algorithm verification experiment system (MVES) within the MATLAB/Simulink theoretical simulation model directly to communicate with the TwinCAT 3 PLC master program to perform different servo control experiments. The MVES supports various Simulink models. However, its and the operation is simple and convenient, which greatly reduces the workload of the algorithm test and has important practical value. Two sets of comparative experiments were used to verify the versatility and superiority of MVES.

Educational benefits of collaborative learning have been demonstrated in several studies and various systems have been developed to date. Numerous efforts have been made to enhance these benefits by supporting collaborative learning with information and communications technology. These efforts have primarily involved support for constructing collaborative learning groups, for collaborative learning in e-learning environments, and for collaborative learning analysis. This study aims to develop a computer-supported collaborative learning system that supports instructors in real time to facilitate collaborative learning in a face-to-face environment with multiple learners at the same time to provide enhanced support. Both the learner and instructor have one tablet terminal and conduct collaborative learning in a single classroom. Herein, the learner can use the tablet to save an educational log and freely browse the educational log of another learner. By referencing the educational logs, learners can learn through face-to-face communication. Additionally, the instructor can determine (1) who is viewing whose educational log and to what extent and (2) which learner is struggling to achieve targets. Herein, an overview of the proposed system is provided and the results obtained using the proposed system are reported to evaluate its effectiveness.