Error Learning Research Articles

Artificial intelligence for load forecasting: A stacking learning approach based on ensemble diversity regularization

State-of-art artificial intelligence (AI) has made great breakthroughs in various industries. Ensemble learning mixed with various predictors provides a considerable solution for electric load forecasting in power system. In our paper, the generalization error of ensemble learning is statistically decomposed to exhibit the significance of base-learner diversity. A diversity regularized Stacking learning approach is proposed to solve the electric load forecasting issue. In our model, the input features are comprehensively selected by various tree-based embedded methods to understand the feature contribution. The robust candidate base-learners are extracted from sub-model pool depending on diversity regularization besides the individual learning capability. Mutual information theory and hierarchical clustering quantitatively assess the dissimilarity degree among base-leaners by exploiting error distribution. The Stacking ensemble framework is utilized to avoid the over-fitting occurrence by employing leave-one-out data splitting procedure for raw dataset block. At last, various cases from different time horizons or geographical scopes are deployed to verify the validity of the model. The case shows that the diversity regularized Stacking learning has better prediction performance compared with the traditional ensemble model or single model. Load forecasting results become more accurate and stable when elaborately selecting base-learners portfolio.

An Adaptive Reinforcement Learning-Based Scheduling Approach with Combination Rules for Mixed-Line Job Shop Production

Affected by economic globalization and market diversification, more manufacturing enterprises realize that large-scale production cannot adapt to the current market environment. The new trend of multivariety customized mixed-line production brings a higher level of disturbances and uncertainties to production planning. Traditional methods cannot be directly applied to the classic flexible job shop scheduling problem (FJSP). Therefore, this paper presents an adaptive scheduling method for mixed-line job shop scheduling. First, the scheduling problem caused by combined processing constraints is studied and transformed by introducing the definition of virtual operation. According to the situation of the coexistence of trial-production and batch production, the disturbance processing mechanism is established. And a scheduling decision model is established based on contextual bands (CBs) in reinforcement learning to overcome the shortcoming of poor performance of traditional single dispatching rule strategy. Through continuous trial and error learning, each scheduler can select the most suitable scheduling rules according to the environment state. Finally, we benchmark the performance of the scheduling algorithm with scheduling methods based on a variety of single scheduling rules. The results show that the proposed algorithm not only improves the performance in the mixed production scheduling problem but also effectively copes with emergency trial-production orders.

Differential Graphical Games for Constrained Autonomous Vehicles Based on Viability Theory.

This article proposes an optimal-distributed control protocol for multivehicle systems with an unknown switching communication graph. The optimal-distributed control problem is formulated to differential graphical games, and the Pareto optimum to multiplayer games is sought based on the viability theory and reinforcement learning techniques. The viability theory characterizes the controllability of a wide range of constrained nonlinear systems; and the viability kernel and the capture basin are the pillars of the viability theory. The capture basin is the set of all initial states, in which there exist control strategies that enable the states to reach the target in finite time while remaining inside a set before reaching the target. In this regard, the feasible learning region is characterized by the reinforcement learner. In addition, the approximation of the capture basin provides the learner with prior knowledge. Unlike the existing works that employ the viability theory to solve control problems with only one agent and differential games with only two players, the viability theory, in this article, is utilized to solve multiagent control problems and multiplayer differential games. The distributed control law is composed of two parts: 1) the approximation of the capture basin and 2) reinforcement learning, which are computed offline and online, respectively. The convergence properties of the parameters' estimation errors in reinforcement learning are proved, and the convergence of the control policy to the Pareto optimum of the differential graphical game is discussed. The guaranteed approximation results of the capture basin are provided and the simulation results of the differential graphical game are provided for multivehicle systems with the proposed distributed control policy.

Feedforward Error Learning Deep Neural Networks for Multivariate Deterministic Power Forecasting

This article proposes a deep neural network (DNN) framework for multivariate deterministic power forecasting in the context of the high penetration of variable and uncertain renewable energy sources. The deep learning model is organized based on the 1-D convolutional neural network to lessen the computational burden, typical of recurrent neural network based models, and combines WaveNet and EfficientNet to improve the forecasting accuracy. Motivated by the inefficiency that all the models conduct the same tasks in the popular ensemble approach, we also designed a feedforward error learning DNN, which computes the error of the basic model separately. We further incorporated embedded and filter methods for feature selection to enhance the model visibility and the utility of the framework. Comprehensive studies on the public load and PV datasets demonstrate that the proposed framework outperforms the conventional methods in applicability, computational efficiency, and forecasting accuracy.

고등학교 운동선수의 실수인식과 스포츠자신감 및 인지된 경기력의 인과관계

The purpose of this study was to investigate the relationship among perception of error, sports confidence and perceived performance of high school Athletes. For this purpose, samples were collected from 300 high school athletes using the convenient sampling method. the collected data, a total of 292 data were used for the final analysis, excluding data from 8 people with insufficient response consistency. For the processing of descriptive statistics, descriptive statistics, confirmatory factor analysis, correlation analysis, and structural equation model were performed using SPSS 23.0 and Amos 23.0 programs. The study results are as follows. First, the recognition of perception of error(learning, challenge, and consideration) was found to had a positive(+) effect on sports confidence. Second, the burden of perception of error was found to had a negative(-) effect on sports confidence. Third, the recognition of perception of error(learning, challenge, consideration) was found to had a positive(+) effect on the perceived performance. Fourth, the burden of perception of error was found to have a negative(-) effect on the perceived performance. Fifth, sports confidence was found to had a positive(+) effect on perceived performance.

체육수업에서 교사와 학생 및 동료 관계가 학생들의 실수 인식과 도전감에 미치는 영향

The purpose of this study is to investigate the effects of teacher-student and peer relationships on students" perception of errors and sense of challenge in PE lessons. For this purpose, the investigator found a population of middle and high school students in Gyeonggi Province and selected 337 students through convenience sampling. A multiple regression analysis was conducted with the collected data in the SPSS Win Ver 23.0 statistical program. The findings were as follows: first, teacher-student, and peer relationships had statistically significant effects on students" perception of errors(error consideration, error learning, and error challenge) in PE lessons. Teacher and student relationship had no significant effects on error burden, one of the subfactors of the perception of errors; secondly, students" perception of errors had statistically significant impacts on their sense of challenge in PE lessons, but error burden under the perception of errors had no significant effects on it; and finally, teacher-student and peer relationships had statistically significant influences on students" sense of challenge in PE lessons.

A longitudinal study of learners’ writing errors in French

Several scholars have carried out investigations on writing errors in second language learning. However, longitudinal studies that focus on the linguistic development in general and lexical competence of the anglophone learners of French, in particular, are still very scarce. An investigation of this sort will give a concise scope of the language development of the anglophone learners of French and some of the factors that are responsible for the errors found in their writing. This study investigates the writing errors of the Obafemi Awolowo University learners of French. The subjects who participated in the study were 14 beginner students. The learners were monitored from their first year of study at the University all through the end of their third year of study. An essay writing exercise was administered on the subjects at the end of their first academic year in 2015/2016. By the end of their second year at the university in 2016/2017 and the end of their third year in 2017/2018, the same essay writing exercise was administered to the same set of students. The study found out that Obafemi Awolowo University learners of French-made frequent errors in their writing. They made repeated errors ranging from wrong spellings, determinants, prepositions, choice of words, overgeneralization, and wrong analogy. The study thus suggested that the learning techniques, teaching methods, and teaching curriculum be regularly reviewed to reflect the language needs of the learners.

A unified fixed-time framework of adaptive fuzzy controller design for unmodeled dynamical systems with intermittent feedback

Although conventional dynamic surface filters can eliminate the issue of complexity explosion in backstepping design, the convergence of filter errors significantly influences the overall control performance. This paper proposes a unified design framework of an event-triggered adaptive fuzzy design scheme with fixed-time performance for unmodeled strict-feedback nonlinear systems, where the tracking performance, filter error, and parameter learning convergence are considered simultaneously. The unified framework guarantees the smoothness of all the closed-loop signals. A novel dynamic surface filter is designed to avoid the repeated differentiation of recursive virtual control while the filter error converges in fixed time. Compared with the conventional backstepping design, the proposed adaptive fuzzy controller design in this paper can guarantee fixed-time tracking performance. An event-triggered condition with intermittent feedback is developed to decrease the computational and communication burden. Two simulation examples are provided to validate the effectiveness of the unified fixed-time framework.

Optimizing the human learnability of abstract network representations

Precisely how humans process relational patterns of information in knowledge, language, music, and society is not well understood. Prior work in the field of statistical learning has demonstrated that humans process such information by building internal models of the underlying network structure. However, these mental maps are often inaccurate due to limitations in human information processing. The existence of such limitations raises clear questions: Given a target network that one wishes for a human to learn, what network should one present to the human? Should one simply present the target network as-is, or should one emphasize certain parts of the network to proactively mitigate expected errors in learning? To investigate these questions, we study the optimization of network learnability in a computational model of human learning. Evaluating an array of synthetic and real-world networks, we find that learnability is enhanced by reinforcing connections within modules or clusters. In contrast, when networks contain significant core-periphery structure, we find that learnability is best optimized by reinforcing peripheral edges between low-degree nodes. Overall, our findings suggest that the accuracy of human network learning can be systematically enhanced by targeted emphasis and de-emphasis of prescribed sectors of information.

Design of a Fuzzy Algorithm-Based Evaluation System for the Effectiveness of International Online Chinese Listening and Speaking Teaching.

With the increasing popularity of online foreign language teaching and learning practice, learners and teachers have developed a high demand for evaluation of teaching effectiveness. When we focus discussion on international Chinese teaching, which has been developed for a relatively short time and not experienced enough, online teaching effectiveness evaluation has become an important obstacle to the development of teaching. This paper introduces a hybrid technique based on fuzzy evaluation method, for determining and suggesting possible types of errors in international Chinese online listening and speaking instruction and giving suggestions for improvement. The system can help learners to identify and determine the types of errors in Chinese listening and speaking learning in a timely manner and make a more objective and comprehensive evaluation of learning performance; at the same time, it helps teachers to trace the effectiveness of teaching design and implementation in a targeted manner and make corresponding scientific decisions. This hybrid technology combines existing language teaching evaluation models, takes advantage of data from online education, and creates corresponding criteria through machine learning fuzzy algorithms and large data sample training, combined with the theory of effective teaching evaluation, which is beneficial for all participants of online Chinese listening and speaking teaching to improve their learning effectiveness.

생식세포분열 문제해결과정에서 나타나는 과학고와 일반고 학생의 개념과 표상의 오류

Objectives The purpose of this study was to analyze error types in the connection between concepts and representations used by high school students in solving the meiosis problem, and to analyze the difference in the frequency of error types between science high school and general high school students. Methods The subjects were 75 science high school and 93 general high school, a total of 168 students. The connection between concepts and representations used by students in solving the meiosis problem was analyzed. First, errors in each stage of meiosis were classified by Masnick & Klahr(2003), and then error types were analyzed in terms of the connection between concepts and representations by Kang, Lee, & Noh(2008). Results As a result of the study, error types in the connection of concepts and representations that appeared in solving the meiosis problem could be divided into 10 types. As a result of analyzing the frequency of error types, it was confirmed that the level of understanding of meiosis was not relatively high because each study subject had an average of 2.3 error types. The most common types of connection errors between concepts and representations were ‘failing to connect’, ‘impossible connection’, and ‘insufficient connection’. The error types of science high school and general high school students showed a significant difference only in ‘impossible connection’, but there was no significant difference in the other types. Conclusions Through the causes and solutions of connection errors between concepts and representations, it will be possible to help students prevent errors in genetic learning as well as in meiosis learning. In addition, it is expected that the wrong use of representations will be prevented, and students' connection errors will be reduced and used in learning strategies to improve the understanding of scientific concepts.

A Stable Generalized Finite Element Method Coupled with Deep Neural Network for Interface Problems with Discontinuities

The stable generalized finite element method (SGFEM) is an improved version of generalized or extended FEM (GFEM/XFEM), which (i) uses simple and unfitted meshes, (ii) reaches optimal convergence orders, and (iii) is stable and robust in the sense that conditioning is of the same order as that of FEM and does not get bad as interfaces approach boundaries of elements. This paper designs the SGFEM for the discontinuous interface problem (DIP) by coupling a deep neural network (DNN). The main idea is to construct a function using the DNN, which captures the discontinuous interface condition, and transform the DIP to an (approximately) equivalent continuous interface problem (CIP) based on the DNN function such that the SGFEM for CIPs can be applied. The SGFEM for the DIP is a conforming method that maintains the features (i)–(iii) of SGFEM and is free from penalty terms. The approximation error of the proposed SGFEM is analyzed mathematically, which is split into an error of SGFEM of the CIP and a learning error of the DNN. The learning dimension of DNN is one dimension less than that of the domain and can be implemented efficiently. It is known that the DNN enjoys advantages in nonlinear approximations and high-dimensional problems. Therefore, the proposed SGFEM coupled with the DNN has great potential in the high-dimensional interface problem with interfaces of complex geometries. Numerical experiments verify the efficiency and optimal convergence of the proposed method.

Learning from Generational Products: Effect on Rapid Utilization of Newly Introduced Resources

Although generational product introduction is crucial for organizations, existing studies on the products tend to focus on generational product launches by innovators. In addition, organizations learn from heterogeneity due to knowledge transfer effects; however, prior studies have found both positive and negative impacts of the heterogeneity for improving outcomes. This study resolves the conflict by focusing on generational overlap volume in generational products. This study investigates how the overlap is associated with the rapid resource utilization of newly introduced resources that contribute to competitive advantages. Using data on newly introduced aircraft between 1960 and 2019, this study found an inverted U-shaped relationship between generational overlap volume and rapid utilization. This study also considered learning errors in the learning processes and found generational overlap scope reflects proximity between old and new generations, and firm introduction experience moderates the inverted U-shaped relationship. When the generational proximity is moderate, a negative effect in the second half of the inverted U-shaped relationship is dampened. Moreover, when firm introduction experiences are higher, the inverted U-shaped relationship becomes flatter. These findings contribute to the literature on organizational learning and generational product launch.

Inferring causal networks of health care resilience and safety performance indicators: A two-stage fuzzy cognitive map approach

In the literature, there are three major concepts used to model resilience and safety performance: (i) integrated resilience engineering (IRE), (ii) work motivational factors, (iii) health, safety, environment, and ergonomics (HSEE). Over the last few decades, health care has developed a compartmentalized vision of performance, safety, and resilience. The issues caused by this are not yet clearly understood since the developments in each dimension are in their infancy. This paper presents a two-stage fuzzy cognitive map (FCM) using a non-linear Hebbian learning (NHL) algorithm and three evolutionary algorithms (EAs) to model the causal relations among these concepts and their impact on health care performance. This paper aims to infer the causal network of these concepts. We use an FCM trained in two stages to show the synergistic relationships among the health and safety paradigms and their favorable effects on the organization's safety performance. The methodology developed in this study targets achieving the minimum learning error by running an NHL algorithm followed by three EAs comprising genetic algorithm (GA), particle swarm optimization (PSO), and imperialist competitive algorithm (ICA). The method is applied to a case study of a large general hospital. While a mere defuzzification of the experts' judgments yields virtually no distinguishability between the individual factors and their impact on performance, our approach extracts a much-improved degree of differentiation. One striking example is the obtained high impact of self-organization and overall workload on resilience and safety performance.

Pupil dilation and skin conductance as measures of prediction error in aversive learning

Notwithstanding the success of CBT, it is relatively unknown how individuals can better profit from corrective learning experiences. Various theories postulate that prediction errors – the difference between what is occurring and what is expected – are the driving force of associative (re)learning. While prediction errors are typically operationalized as violations of cognitive outcome expectancies, direct physiological indices of prediction errors could capture potentially more essential automatic and emotional processes in associative learning. Although physiological responses have previously been suggested to reflect prediction errors, it remains elusive if these measures actually predict changes in subsequent conditioned responding. In three fear-conditioning experiments, we compared pupil dilation and skin conductance responses to unexpected outcomes – unconditioned stimulus (US) presentations or omissions – with expected outcomes, and tested whether outcome responses predicted actual changes in subsequent conditioned responding. We found evidence for increased physiological responses to unexpected outcomes, but the results were inconsistent across experiments. Furthermore, only pupil responses to US presentations consistently predicted an increase in conditioned responding, making it difficult to reconcile our findings with associative learning models. Both pupil dilation and skin conductance can thus index unexpected outcomes, but the relationship of these responses to future learning is not evident and requires further investigation.

Optimization algorithm for feedback and feedforward policies towards robot control robust to sensing failures

Background and problem statementModel-free or learning-based control, in particular, reinforcement learning (RL), is expected to be applied for complex robotic tasks. Traditional RL requires that a policy to be optimized is state-dependent, that means, the policy is a kind of feedback (FB) controllers. Due to the necessity of correct state observation in such a FB controller, it is sensitive to sensing failures. To alleviate this drawback of the FB controllers, feedback error learning integrates one of them with a feedforward (FF) controller. RL can be improved by dealing with the FB/FF policies, but to the best of our knowledge, a methodology for learning them in a unified manner has not been developed.ContributionIn this paper, we propose a new optimization problem for optimizing both the FB/FF policies simultaneously. Inspired by control as inference, the proposed optimization problem considers minimization/maximization of divergences between trajectories, one is predicted by the composed policy and a stochastic dynamics model, and others are inferred as optimal/non-optimal ones. By approximating the stochastic dynamics model using variational method, we naturally derive a regularization between the FB/FF policies. In numerical simulations and a robot experiment, we verified that the proposed method can stably optimize the composed policy even with the different learning law from the traditional RL. In addition, we demonstrated that the FF policy is robust to the sensing failures and can hold the optimal motion.

A gradual temporal shift of dopamine responses mirrors the progression of temporal difference error in machine learning.

A large body of evidence has indicated that the phasic responses of midbrain dopamine neurons show a remarkable similarity to a type of teaching signal (temporal difference (TD) error) used in machine learning. However, previous studies failed to observe a key prediction of this algorithm: that when an agent associates a cue and a reward that are separated in time, the timing of dopamine signals should gradually move backward in time from the time of the reward to the time of the cue over multiple trials. Here we demonstrate that such a gradual shift occurs both at the level of dopaminergic cellular activity and dopamine release in the ventral striatum in mice. Our results establish a long-sought link between dopaminergic activity and the TD learning algorithm, providing fundamental insights into how the brain associates cues and rewards that are separated in time.

How Can Servant Leadership Promote Employees' Voice Behavior? A Moderated Chain Mediation Model.

PurposeBased on Social Exchange Theory, this paper constructs and evaluates the mediation model of servant leadership on employees’ voice behavior through psychological security and error learning ability and discusses the moderation role of work autonomy in the model.DesignThis research used 424 employees and their direct superiors as the research objects and then conducted a paired survey at three points. Mplus7.4 software analyzed the empirical data.FindingsThe results are shown servant leadership is positively correlated with employees’ voice behavior, but psychological security and error learning ability play a partial mediation role in the positive impact of servant leadership on employees’ voice behavior. while psychological security and error learning ability play a chain mediation role in the positive impact of servant leadership on employees’ voice behavior. Besides, work autonomy moderates the chain mediation path by enhancing the positive impact of servant leadership on psychological security. the higher the employees’ work autonomy is, the stronger the positive impact of servant leadership on psychological security will be, and the stronger the chain mediation effect of psychological security and error learning ability between servant leadership and employees’ voice behavior.Practical ImplicationsManagers should abandon the traditional “command and control” management mode, focus on the servant leadership style and improve employees’ psychological security through a comfortable and independent working environment. Besides, managers should set up a suitable error learning atmosphere mechanism, appropriately allow employees to work, give them a specific work autonomy, increase employees’ work flexibility, and encourage employees to provide a cheerful voice for the development and progress of the organization.Origin/valueFrom the perspective of Social Exchange, this study more comprehensively discusses the internal action path of the impact of servant leadership on employees’ voice behavior, enriches the antecedent variables of employees’ voice behavior, expands the existing research on the mediation mechanism of servant leadership, deepens the understanding of the efficiency mechanism of servant leadership, and has enlightenment significance for managers to stimulate employees’ voice better.

VoLearn

Conventional motion tutorials rely mainly on a predefined motion and vision-based feedback that normally limits the application scenario and requires professional devices. In this paper, we propose VoLearn, a cross-modal system that provides operability for user-defined motion learning. The system supports the ability to import a desired motion from RGB video and animates the motion in a 3D virtual environment. We built an interface to operate on the input motion, such as controlling the speed, and the amplitude of limbs for the respective directions. With exporting of virtual rotation data, a user can employ a daily device (i.e., smartphone) as a wearable device to train and practice the desired motion according to comprehensive auditory feedback, which is able to provide both temporal and amplitude assessment. The user study demonstrated that the system helps reduce the amplitude and time errors of motion learning. The developed motion-learning system maintains the characteristics of high user accessibility, flexibility, and ubiquity in its application.

A Fuzzy System Based Iterative Learning Control for Nonlinear Discrete-Time Systems with Iteration-Varying Uncertainties

In this paper, we consider an iterative learning control problem for a class of unknown discrete-time nonlinear systems with iteration-varying initial error, iteration-varying system parameters, iteration-varying external disturbance, iteration-varying desired output, and iteration-varying control direction. These iteration-varying uncertainties are not required to take any particular structure such as the high-order internal model and only need to satisfy certain boundedness conditions. We propose an iterative learning control law with an adaptive iteration-varying fuzzy system to overcome all the uncertainties and achieve the learning control objective. Furthermore, we present a sufficient condition for designing adaptive gains and prove the convergence of the learning error to a small value as the trial number is large enough. Finally, we use two simulation examples to demonstrate all the theoretical results.