Automated essay scoring (AES) is to estimate the scores of essays automatically. Two types of AES models are commonly used: handcrafted feature-based and neural-based models. In this paper, we introduce AES systems based on the two types for evaluating the logical consistency of Japanese essays. In addition, to enhance the performance of models, we integrate the neural-based model with the handcrafted features: a hybrid AES system. In the experiment, we show the effectiveness of our hybrid AES system. Besides, most of our AES models obtained higher QWK scores than human evaluators. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Pedagogical agents offer significant promise for engaging students in learning. In this paper, we investigate students' conversational interactions with a pedagogical agent in a game-based learning environment for middle school science education. We utilize word embeddings of student-agent conversations along with features distilled from students' in-game actions to induce predictive models of student engagement. An evaluation of the models' accuracy and early prediction performance indicates that features derived from students' conversations with the pedagogical agent yield the highest accuracy for predicting student engagement. Results also show that combining student problem-solving features and conversation features yields higher performance than a problem solving-only feature set. Overall, the findings suggest that student-agent conversations can greatly enhance student models for game-based learning environments.

Across the past decade, open science has increased in momentum, making research more openly available and reproducible. Artificial Intelligence (AI), especially within education, has produced effective models to better predict student outcomes, generate content, and provide a greater number of observable features for teachers. While completed, generalized AI models take advantage of available open science practices, models used during the actual research process are not made available. In this tutorial, we will provide an overview of open science practices and their benefits and mitigation within AI education research. In the second part of this tutorial, we will use the Open Science Framework to make, collaborate, and share projects - demonstrating how to make materials, code, and data open. The final part of this tutorial will go over some mitigation strategies when releasing datasets and materials so other researchers may easily reproduce them. Participants in this tutorial will learn what the practices of open science are, how to use them in their own research, and how to use the Open Science Framework. The website (https://aied2023-tutorial.howtoopenscience.com/ ) and associated resources can be found on an Open Science Framework project (https://doi.org/10.17605/osf.io/yd9kr ). © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Student performance prediction continues to be a focus of research in educational data mining due to its many potential benefits. While teachers’ assessment reports are a crucial part of the educational process, they have not been commonly used in performance prediction. We propose a model that uses similarity learning as an embedding-enhancing technique. Results outperform earlier research with an average accuracy of 73% for detecting strong performance. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Evidence-based learning strategies, such as the testing effect, might help address the achievement gap. However, exploiting the testing effect depends on having a set of instructional activities with fine-grained tagging. While instructors might find questions in textbooks, they often lack fine-grained tagging, and data labeling is laborious. Despite much research on text classification, to our best knowledge, state-of-the-art question classifiers are mostly based on extensive models (i.e., BERT) and English text. Respectively, those are incompatible with the resource-constrained devices (e.g., mobile) and languages (e.g., Portuguese) of many underprivileged countries in the global south. Therefore, we developed a question classifier on top of DistilBERT, a version of BERT compatible with resource-constrained applications, using grid search and hold-out. Based on a corpus of 1045 coding questions written in Brazilian Portuguese, we found a model that achieved a near-perfect performance on unseen data, similar to last-generation results using BERT for English text. Thus, we present a step towards equitable education by i) providing underprivileged Portuguese-speaking countries with the support that enables opportunities already available for first-world countries and ii) demonstrating the feasibility of creating resource-constrained applications compatible with state-of-the-art AIED systems. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Educational philosophies have slowly focused on differentiated and self-learning systems that respectively emphasize tailoring instruction to meet individual needs and gathering, processing, and retaining knowledge without the help of another person in recent years. In this regard, creating opportunities to further self-learning resources has become increasingly important. Some people started to prefer these over traditional learning practices for various reasons, such as the difficulty of transportation in metropolises or fitting their timetables to in-person lessons. The creation of such platforms or opportunities for physical education, however, proves to be more difficult as individuals require continuous and precise feedback regarding the usage of their bodies. Accordingly, we have developed an augmented reality application that presents a platform for dance that focuses on differentiated and self-learning principles with accurate feedback. We built the augmented reality (AR) app prototype using the Swift programming language and used the MoveNet pose detection model along with our own neural network to capture the body position. Our proposal could prove a valuable addition to learning physical activities assisted by AI systems since the application of AI technologies to dance and physical education could be improved by further investigation and research. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

This paper provides a comprehensive overview of using the Kit-build concept map in a flipped classroom setting to aid in the learning process. The Kit-build concept map is an effective tool for helping learners create concept maps during pre-class preparation, which enhances their understanding of basic concepts and helps them acquire fundamental knowledge to be applied during lectures. Additionally, this paper explores the various achievements of using the Kit-build concept map, including improved student engagement and a deeper understanding of the material covered. There are still several challenges that need to be addressed in the future, such as ensuring that the tool is accessible to all learners and used effectively to achieve learning outcomes. This paper argues that the Kit-build concept map is a valuable tool for both students and educators and that its continued use and refinement can lead to significant improvements in the learning process. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Artificial Intelligence in Education (AIED) is a driving force to improve education. Nevertheless, policymakers from the Global South fear that AI will increase the digital divide and reduce the opportunities for students in these regions to thrive. To address this problem, we analyzed the past 30 years of data on four aspects of the digital divide. Then, based on these findings and a series of discussions with stakeholders (e.g., policymakers), we proposed the concept of AIED Unplugged. An approach to creating AI-based educational technologies that do not require changes in current school settings (e.g., infrastructure), do not rely on stable internet access, and do not ask for digital skills to use them. We applied this concept to redesign an education policy in Brazil to help students improve their writing skills. Our results show a reduction in time, cost and complexity to running the policy, and a positive impact on more than 500,000 students in 7,000 schools in the country. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

The extent to which individualized and adaptive learning can be supported by recommender systems is increasingly being discussed in the field of adult and continuing education (ACE). Aspects of accessibility and customization of learning platforms play just as much a role as the added value of AI from a pedagogical perspective. This paper addresses the question of how recommender systems can be used to support self-directed learning of adult learners with heterogenous prerequisites and learning needs. Building on the initial situation of the target group as well as assumptions of learning opportunity-use models, an idea of AI and humans acting in partnership is designed and at the same time made to the object of investigation. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Renovation of education requires students to engage in a newer style of learning, or collaborative learning and teachers to implement lesson study of that learning. Such learning, however, creates a situation wherein multiple student groups simultaneously engage in dialogues in a noisy classroom. Thus, we developed a “Learning Recorder” system, which has a 360-degree video camera to record students when set in the center of the group, and sends the voice data to an Automatic Speech Recognition system for immediate transcription. This paper explains how this system has evolved from our practical experiences with AI-powered lesson study and how it helped even a novice teacher look back and learn from the student dialogue. We propose that an information appliance like the Learning Recorder draws teachers’ attention to student learning, solicits multiple interpretations, and brings about collaborative learning among teachers. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.