This abstract represents our full paper categorized under the ‘Research to Practice’ category. Program tracing tasks ask about the behavior of a program, such as “the next expression to be executed” and “the value of variables during the execution of that expression”. The program tracing exercise system developed by Tomoto et al. provides an environment for efficiently conducting exercises using program tracing tasks. This system poses questions to learners based on program tracing tasks, adapting to their answers. However, implementing questions for new programs in this system requires considerable effort and time. Therefore, we propose a method to automatically generate tracing tasks by identifying the targets of questions from the program and following predefined question guidelines. Specifically, we first create a model that formalizes the tracing tasks. This model consists of the targets of questions, their execution order, and execution conditions. Then, we develop a method to generate data from the program to implement this model. This method derives an abstract syntax tree from the program, identifies question targets based on the attribute values of nodes, and specifies the order and conditions for executing questions based on the relationships between nodes. Consequently, tracing tasks can be generated by implementing the model with the generated data. We found that tracing tasks could be generated using the proposed model and the data extracted by the proposed method from programs extracted from a programming exercise textbook.

In programming education, refining code is crucial. Trial and error is one effective method for achieving this refinement. However, it can be challenging for learners to maintain motivation during conventional programming exercises due to them not knowing how good their code is and not being able to continuously perform trial and error. To address these issues, it is important to establish criteria for evaluating code quality and have learners objectively recognize them. Therefore, we propose and implement a method that uses a code visualization system that visualizes quality indicators, a ranking system, and a code viewing function based on that ranking. Evaluation of this system when implemented in a university classroom suggests that students write better code, based on these quality indicators, when they perform trial and error by themselves or by viewing other learners' code.

It is important for learners to understand the relationship between mathematical problems and their answers to those problems. One way for learners to recognize the significance of their answers method for a problem is through trial and error. However, in ordinary mathematics learning, the correct answer is generally fed back from the textbook or teacher after the problem is attempted, which does not give learners an opportunity to reflect on their answers. So far, we have developed a system that encourages trial and error by visualizing the learner's answer for the mathematics of vectors and having them interact with it. However, there was a possibility that some learners would only look at the visualized figures to determine correctness and would not reflect on the errors they were making. In this study, we proposed and developed a system that requires learners to interact with figures on the system side, thereby promoting greater understanding of the relationship between mathematical problems and their answers.

It is important to reuse knowledge acquired through problem-solving in programming. To reuse knowledge, it is effective to first understand differences between knowledge items and then to organize that knowledge. Therefore, we propose a method and develop a support system for facilitating knowledge organization in programming. We further examine a model of parts and problem-solving process of parts based on function–behavior– structure aspects. In this paper, we expand a previously developed system to propose an improved system that provides support based on feedback from these models. The aim of this system is to allow learners to consider behavior rather than thinking directly from function to structure.

It is sometimes challenging for learners of elementary physics to correct their mistaken notions regarding physics concepts. The need to address this phenomenon is illustrated by the existence of common misconceptions such as ‘motion implies a force’ (MIF), in which learners mistakenly believe that force always acts in an object’s direction of motion, as well as educational approaches using error-based simulation (EBS). In a conventional EBS, a learner’s mistake is usually output as a physical phenomenon, but there are cases in which the learner’s mistake cannot actually be shown. For this reason, we propose an EBS in which measurement information is displayed to the learner in the form of a measurement tool together with the physical phenomenon and shows error in response to the learner's mistake, something the conventional EBS is unable to do. In this paper, we propose an EBS learning exercise using a measurement tool in which the learners themselves must determine the measurement in order to confirm the correct answers. Using this method, it is possible for the learner to use trial and error after getting the wrong answer and eventually find the correct answer. It is also possible to support learning in a virtual experiment environment using a computer. Experiments were performed using a system equipped with the measurement tool, and it was verified that the learners made significant progress learning physics through trial and error.

In this study, we aim to develop an educational environment and platform based on color science, merging information technology, robotics and artificial intelligence. In the educational environment, the robot requires functional capability for human-like communication. However, the robot may have difficulty in grasping the small cues to represent feelings and emotional changes of human by itself. This paper presents a measurement system with pneumatic pressure sensor for heart rate variability analysis. We proposed a fuzzy spiking neural network to extract heartbeat signal from the measured data. We furthermore conducted experiments to investigate the effects on student's psychological states while solving some conundrums by using the measurement system.

A framework for indexing problems is proposed, based on which explanation generation and problem sequencing for conceptual understanding in science can be automatized. It helps student learn how to make appropriate models to solve problems and prevents them from superficially read the solution of a worked example and apply it wrongly to others. The results of a preliminary experiment is also described in which the explanations generated based on the framework promoted subjects’ conceptual understanding in mechanics.

e-Testing is effective in relieving time and space limitations for examinations. However, a drawback is that user authentication employs only login credentials , making cheating easy. We examine variations in personal writing data in e-testing with the aim of detecting cheating. We assume two behaviors: answering questions by oneself and replicating others’ answers. We used the sub-stroke method for data analysis, which allowed confirmation of differences in the change over time in writing speed and pen pressure, and differences in average pen angle and direction.

We analyze text discussions with the aim of using statement data to assess the status of an in-progress discussion. We assume convergent discussions with three discussion status types: Publication, Active, and Convergent. Then we propose an automatic method for evaluating discussion status by network analysis. Experimental evaluation indicates that the method is useful. We also achieved positive results in comparison with manual analysis.

We propose a method for evaluating participants’ roles in online text discussion using network analysis of chat log data. In our method, we use log data to construct a statement network indicating the relativity of the statements. We evaluate the participants’ role by calculating four types of point centralities. Experiments have been conducted to evaluate our method. The results from our method closely correspond to participants’ peer assessment data. This indicates that the proposed method is effective.

This paper reports on the practical use of Error-based Simulation (EBS) for learning dynamics in a junior high school. EBS is simulation generated from learners’ erroneous answers. Differences between simulations showing correct phenomena and EBS showing unusual phenomena promote learner awareness of errors. Previously, we implemented EBS for learning about normal reaction forces in statics problems, and confirmed that EBS was effective for learners’ self-correction of errors. In that problem, however, visualization of errors is easy because a correct simulation is always motionless. In this research, we implemented EBS for a dynamics problem, and put it into practical use in a junior high school. Because errors can be visualized as unexpected motion, this is an important generalization of EBS. The learning effect of EBS in dynamics problems was evaluated by comparing the results of pre-tests, post-tests, and delayed post-tests between two classes, a experimental class where students learned with the EBS and a control class where students learned as usual. The results suggest that EBS is also useful in dynamics problems.

In this paper, we describe developing and evaluation of a learning environment for Solution-based Problem-Posing in the case of multi-digit subtraction. One of the ways to learn by problem-posing is to make learners pose problems which can be solved by a solution. We call the type of problem posing Solution-based Problem-osing. The Solution-based Problem-Posing helps the learners understand the solution. In the case of multi-digit subtraction, however, it is difficult to provide the solutions directly because the solutions in the multi-digit subtraction are complex for learners who are elementary school students when they learn them. We suggest a method they are required to solve problems to be conscious the solution of the problems and then pose problems based on the solution.

Error-based simulation (EBS) has been developed to generate phenomenon by using students' erroneous idea and also it is considered as a promising method to promote the students be aware of their errors. In this paper, we used the EBS in learning of “normal reaction force” in a junior high school, and then compared the EBS class (where students learned the concept with EBS) and the usual class (where students learned the concept as usual) by using the results of a pre-test, post-test, delayed post-test and an interview after the delayed post-test. The results suggest that EBS contributed to generalization and retention of the learned concept as for “normal reaction force”.