Artificial Intelligence in Education Technologies: New Development and Innovative Practices

STEM education has taken on high importance in Hong Kong K-12 education landscape. Despite policy advocacy and curriculum endeavour, the quality of STEM education varies significantly between schools. Research literature indicates that high-quality STEM education requires teachers’ rigorous delivery of topics and appropriate pedagogies, and one approach to improve such practices is teacher professional development (TPD). However, because current research on TPD has not given explicit consideration to the complex nature of STEM education, there remains a lack of a clear blueprint of how TPD should be conducted to build teachers’ capacity for STEM education effectively. This paper presents a case study that explores the necessary attributes and identifies the missing links of STEM education TPD by understanding how various TPD models supported a Hong Kong K-12 school embracing STEM education. Qualitative data collection methods, including semi-structured interviews and classroom observations, were employed to draw a picture of TPD implementations in the selected school. The findings suggest that, at the macro-level, effective STEM TPD should not stop at employing mixed use of TPD models; the models have to be integrated organically with respect to a school-based STEM curriculum implementation approach. A collaborative culture between teachers must be cultivated for effective inter-disciplinary integration. Collaborative action research should also be promoted to develop collective wisdom of STEM pedagogies. At the micro-level, TPACK and cross-disciplinary integration skills need to be focusing areas of STEM TPD. With these guiding principles, some possible strategies for effective STEM education TPD are suggested.

As an important part of public safety throughout the world, primary and secondary school safety has been faced with increasingly diverse and complex risks. Literature mentions different methods to identify and assess safety risks in primary and secondary schools. However, none of these methods help in defining a complete scope. Differently, this study aims to clarify the relationship between risk factors and their impact on school safety by integrating qualitative and quantitative research methods. Firstly, adopting grounded theory as the methodology, data collection entailed from two sources: Semi-structured interviews with experienced primary and secondary school staff (n = 20) across diverse schools, 231 cases of school safety accidents in primary and secondary schools in China from 2011 to 2021. The results show that 76 initial concepts, 19 initial categories, and 8 main categories are obtained through the coding step of grounded theory. Then, the 19 initial categories are defined as 19 risk factors in primary and secondary schools. We employed the combination of Decision-making Trial and Evaluation Laboratory (DEMATEL) and Interpretive Structural Modeling (ISM) techniques for understanding the hierarchal and contextual relationship structure among the 19 risk factors. This present novel model helps the policy and decision-makers to find out a mutual relationship and interlinking with school safety.

The paper addresses the problem of academic course prerequisites and learning outcomes generation in learning analytics systems. For prerequisites generation, collaborative filtering, i.e., ALS algorithm for Matrix Factorization, is used. For learning outcomes generation, the study discusses an approach based on Computational Linguistics data extraction methods and content-based filtering to recommend potential outcomes. The recommendation mechanisms are designed to be implemented in the Educational Program Maker service for working with education process elements. The study's primary goal is to simplify, formalize and speed up the course development process. Implementation of the approach will make it possible to build unambiguous interdisciplinary connections, identify the closest intersections of the curriculum courses, and build individual learning pathways.

Understanding how the UAE K12 education system performs with data-driven evidence is key to inform better policy making to support UAE vision to upskill human capital growth for its economic transformation. In this study, we investigate the potential of using machine learning techniques to understand key learning factors contributing to UAE student math performance on the TIMSS 2019 assessment. Due to the fact that learning factors co-exist and interact with one another, we explore the SHapley Additive exPlanations (SHAP) approach to explain the complexity of the model. The results highlight the importance and contributions of each learning factor and uncover the relationships between the learning factors. Understanding key learning factors and identifying evidence-based intervention opportunities will help policymakers with informed education intervention designs to improve student mathematics learning, in order to improve UAE student TIMSS math performance over the long run.

Massive open online courses and other online study opportunities are providing easier access to education for more and more people around the world. To cope with the large number of exams to be assessed in these courses, AI-driven automatic short answer grading can recommend teaching staff to assign points when evaluating free text answers, leading to faster and fairer grading. But what would be the best way to work with the AI? In this paper, we investigate and evaluate different methods for explainability in automatic short answer grading. Our survey of over 70 professors, lecturers and teachers with grading experience showed that displaying the predicted points together with matches between student answer and model answer is rated better than the other tested explainable AI (XAI) methods in the aspects trust, informative content, speed, consistency and fairness, fun, comprehensibility, applicability, use in exam preparation, and in general.

In teaching, assessment tasks are often used as an important way to evaluate students’ learning abilities. In traditional education, to design an assessment task, e.g., an assignment, teachers are often required to manually design by themselves. It is usually a challenging task to design high-quality assignments, especially for less-experienced teachers. In addition, students often have different learning abilities and it is often difficult and unreasonable to evaluate students’ learning abilities using only the same assignments. Moreover, to design an assignment with decent quality, teachers have to consider the knowledge items to be covered and the difficulty of the assignment. Therefore, it is worthwhile to do the research of automatically providing students with high-quality assessment tasks, taking into account the coverage of knowledge items and the appropriate difficulty of the designed assessment tasks, i.e., proposing an approach of personalized assessment task recommendation systems. The current literature on personalized assessment task recommendations shows that the recommendation process does not take into account the difficulty of the designed questions and students’ knowledge. To overcome the limitations of the current related work, this paper proposes a framework design of intelligent assessment tasks recommendation by considering several aspects, such as students’ learning ability, mastery of student knowledge, the difficulty of assessment tasks, and students’ forgetting characteristics. The proposed framework design consists of two components: data processing and personalized assessment task generation. The data processing component is designed to proceed with data, e.g., generating the initial question bank, analyses on the question bank generated, auto-generation of the assessment task, and the result collection of auto-correcting on the designed assessment task. Besides, the forgetting characteristics of students are also considered in this study for the intelligent assessment tasks recommendation framework design.

Academic scholarships motivate graduates to work hard. However, the tedious application process frustrates students and brings adverse effects. The paper presents a rule-based cloud computing system to assess academic scholarship applications. The criteria can be encoded with if-then expressions. The rule-based system enables the administrator to add new standards or update existing ones flexibly. Students can upload their achievements through a web browser on different devices. Graduates know the mark of each item input into the system. The system calculates final grades automatically. Members of evaluation committees can review and return applications to applicants. All criteria are available for students. The transparency encourages students to participate in applying for scholarships. The experimental results show that the designed system was helpful in evaluating scholarship applications. Students spent less time applying for academic scholarships with the rule-based cloud system than with traditional approaches.

More and more educational institutions are making lecture videos available online. Since 100+ empirical studies document that captioning a video improves comprehension of, attention to, and memory for the video [1], it makes sense to provide those lecture videos with captions. However, studies also show that the words themselves contribute only 7% and how we say those words with our tone, intonation, and verbal pace contributes 38% to making messages clear in human communication [2]. Consequently, in this paper, we address the question of whether an AI-based visualization of voice characteristics in captions helps students further improve the watching and learning experience in lecture videos. For the AI-based visualization of the speaker’s voice characteristics in the captions we use the WaveFont technology [3‐5], which processes the voice signal and intuitively displays loudness, speed and pauses in the subtitle font. In our survey of 48 students, it could be shown that in all surveyed categories—visualization of voice characteristics, understanding the content, following the content, linguistic understanding, and identifying important words—always a significant majority of the participants prefers the WaveFont captions to watch lecture videos.

Micro-expression is the facial expression that is extremely quick and lasts less than half a second. As a spontaneous expression, it is usually produced when people try to suppress their emotions and can reveal the true emotions of human beings. It plays an important role in lie detection. In recent years, with the progress of neural network technology, the research on micro-expression recognition has made significant progress. However, because our understanding of the psychological process of micro-expression recognition is far from complete, the existing method of recognizing micro-expression still cannot meet the standard of practical application. In the present research, we investigated the effects of facial feedback and social identity of the expresser on the recognition of micro-expressions by one behavioral experiment. The results showed that facial feedback can moderate the intergroup bias in micro-expression recognition, which suggests that humans will imitate other people's facial expressions to different degrees in the recognition of micro-expression. At shorter duration, facial feedback has a stronger effect on the recognition of micro-expression of outgroup members. And at longer duration, facial feedback has a stronger effect on the recognition of micro-expressions of ingroup members. This further suggests that we need to consider the identity of the model and the identity of the coder to obtain more accurate and effective data coding when establishing the micro-expression database.

Faced with the challenge of online teaching-learning, university teachers continued with the responsibility of developing their learning sessions, innovating teaching material and methodology during this process, changing the way of generating learning in health sciences students, through the application of videos, summary readings and practices carried out with family members who acted as patients, in order to achieve the planned competition. The importance of letting students know their achievements in relation to what is evaluated, helps them to understand their way of learning, assess their learning result and self-regulate. This is how feedback motivates the student to rethink their learning strategies. The purpose of this study was to determine the effect of feedback on the online learning outcome of health sciences university students, in a non-experimental research, descriptive-correlational level, with a sample of 294 students. The results obtained showed that feedback in university students of Health Sciences in virtual environments is effective when applied in a timely manner and can be planned, based on the evidence of the learning outcome. To achieve this, they must be previously trained, from the first semesters of study, in feedback literacy, making it part of the self-regulation of their learning.

Software Engineering is a brand-new engineering course in the application-oriented skills training program, and an essential professional skill for high-demand employment. Because of the fast growth of science and technology, many students have high expectations for this topic. Students’ excitement and expectations for this topic, however, are dwindling as a result of uninspiring texts and an antiquated teaching method. To improve the problem, this study innovates in the areas of teaching idea, teaching substance, teaching style, and assessment mode. Based on the education concept of “student-centered” and “project-driven,” three basic teaching modalities of “role-playing,” “project case teaching,” and “workshop” were chosen to optimize the teaching content and assessment procedures. The blended teaching method shifts the function of learning from passive to active, and students are liberated from the monotonous theoretical learning environment. Students have made significant progress in course scores, competitions, projects, and other areas as a result of practical instruction, and the effect is palpable.

Learning outcomes have attracted more and more attention in higher education. Many teaching and learning methods have been invented to improve learning outcomes. Teaching and learning pedagogies will attract intensive focus because the educational paradigm is moving from teacher-centered to student-centered learning. However, there is little research on improving learning outcomes based on existing teaching and learning contexts. This paper presented a creative strategy to enhance learning outcomes with two instructors teaching by turns based on learning theories. This novel method utilized the characteristics of different instructors to facilitate student learning and was adopted in the C programming language course to verify its effectiveness. Participants were divided into two groups according to their majors. We randomly selected one group as the treatment group, and the left group was the control group. One instructor taught the control group, and another lectured the treatment group. Students took the pre-test after twelve weeks. The instructor from the control group led the treatment group for a week before the end of the course. Students who failed the pre-test took the post-test. Our experimental results showed that the pair instructing strategy could increase learning outcomes.

Anxiety is a common experience among all of us in every daily situation. It is considered one of the affective factors in foreign language learning. Past research has pointed out that foreign language anxiety negatively correlates with academic performance in all four traditional language skills (speaking, listening, writing, and reading). Of all the skills, reading has been once deemed less susceptible to anxiety; however, it has been increasingly gaining more attention from researchers since findings of the pioneers (Saito and the associates) as a phenomenon related to foreign language anxiety. Notwithstanding increasing studies on foreign language reading anxiety, not much has been conducted in Vietnamese contexts, especially the research site. Thus, a survey design was conducted on 207 advanced level students at an educational Vietnamese institution with the aim to explore participants’ reading anxiety level, possible differences between FLRA and demographic variables, and correlation between FLRA and reading test scores. Results revealed that 96.6% of respondents were at a moderate level of reading anxiety. There were no statistically significant differences between FLRA and demographic variables such as gender, birthplace, and English text reading time per week. Similarly, no statistically significant correlation between FLRA and reading test scores could be found. Pedagogical implications were also discussed.

Objective: In order to explore the multiple chain mediating effects of positive affect and self-esteem between nostalgia and life satisfaction in college students. Methods: 477 college students were investigated with Southampton Nostalgia Scale, Positive Affect Scale, Rosenberg Self-esteem Scale, Life Satisfaction Scale. Results: ① Nostalgia was significantly and positively correlated with positive emotion (r = 0.12, p < 0.01) and life satisfaction (r = 0.13, p < 0.01), also, significant positive correlations exists among other key variables. ② Structural equation models showed that, nostalgia could exert effects on life satisfaction (β = 0.14, P < 0.05), and also indirectly through the independent mediating effect of positive affect, the chain mediating effect of positive affect and self-esteem (β = 0.10, P < 0.05). Conclusion: Nostalgia could affect life satisfaction, not only through direct path, but also through the indirect path of multiple mediating effects of positive affect and self-esteem.

Industrial College has become an effective organizational form of cultivating applied talents. Exploring and constructing the talent training mode based on the industrial college is of great significance of deepening the cooperation between schools and enterprises, promoting the integration of industry, university and research, and improving the quality of talent training. Therefore, based on the analysis of the successful experience of applied talent training in home and abroad, this paper defines the construction idea of the talent training mode of the cold chain industry college. According to the practice and experience summary of the exploration of the new talent training mode by the cold chain industry college of Guangdong University of Science and Technology, we construct “3 + 1” talent training operation mechanism (which Let students study theoretical knowledge in school for the first three years, and enter the industrial class jointly established with enterprises in the fourth year to start practical learning), and analyze its theoretical basis, organization and leadership and curriculum system. It defines the key links such as learning centered teaching method, double qualified tutor team construction, practice base construction, assessment and evaluation reform, considers the construction of relevant systems, and summarizes the practical effect.

Skilled professional personnel are the human resources that enterprises rely on for survival, and their professional competence level affects the long-term development of enterprises. The paper takes 86 vocational skilled talents as the sample, collects the data on the growth of professional ability such as education level, learning, and training, service experience, mentoring, organization and management, extracts the parameter indexes that can objectively reflect their professional ability quality, integrates the characteristics of various methods such as behavioral event interview method, questionnaire survey method, expert group discussion method and focus interview method for comprehensive analysis. Through combing and analyzing, the professional competency quality model of vocationally skilled personnel is constructed and its effectiveness for talent screening and assessment is verified. The model and its creation method can provide a useful reference for enterprises or related personnel.