This paper details the development of an educational chatbot designed to assist non-IT undergraduate students in a computer programming course. The chatbot, accessible through a LINE Bot interface, focuses conversations on course materials and activities, providing students with relevant learning support. Utilizing OpenAI’s API, the chatbot generates human-like conversational content. The paper includes preliminary evaluation results, emphasizing the chatbot’s ability to respond accurately to course-related student queries. This study highlights the potential of chatbots in enhancing educational experiences in specialized subjects.

The integration of educational robots and IoT-based tangible objects (R&T system) holds great potential for enhancing the learning outcome and providing scaffoldings for language learners. This study was focused on examining the learners' behaviors during language learning activities, to analyze the effectiveness of the scaffoldings provided by the R&T system. The aim was to assess the task completion and responsiveness to learners' learning states. A method called Lag sequential analysis was employed to identify significant sequential behavioral patterns. The results showed that, in addition to confirming the effectiveness of the scaffoldings, the sequential pattern analysis was useful in capturing the dynamics of interaction among learners, robots, teachers, and peers in real-life learning scenarios.

This study used deep learning techniques with Moodle log data to predict student performance in introductory computer programming courses. Particularly, this study would like to use prediction results to identify potential low-performing students who may need assistance from teachers. The results suggested that deep learning models are promising to predict student performance and identify low-performing students in the researched context. What the prediction results provided by the models can inform teachers in learning settings was also further discussed in this paper.

This study conducted a cross-classified multilevel analysis of undergraduate students' peer-/self-assessment scores in a collaborative learning environment. The analysis intended to investigate group, rater and ratee effects on the peer-/self-assessment scores. The findings showed that group, rater and ratee effects were all statistically significant on peer-/self-assessment scores. The raters' and ratees' gender as well as the diversity in group members' majors influence peer-/self-assessment scores significantly. In addition, there was a statistically significant difference in students' self and peer ratings. The instructional implications of using cross-classified random effects model to analyze peer-/self-assessment ratings were also discussed in the paper.

This paper reports an analysis of elementary school students' learning process as they participate in English learning activities using educational robots and IoT-based toys. Grounded on Vygotsky's notion of zone of proximal development, the analysis is conducted through the lenses of obstacles learners encounter in real learning situations as well as the causes of and the solutions to the encountered obstacles. The techniques of video and qualitative analysis are adopted to conduct this naturalistic inquiry, extracting insight from real-world data for enhancing language learning with emerging technology.