Pub Date : 2025-02-20DOI: 10.1109/TLT.2025.3544454
Sheng Bi;Zeyi Miao;Qizhi Min
The objective of question generation from knowledge graphs (KGQG) is to create coherent and answerable questions from a given subgraph and a specified answer entity. KGQG has garnered significant attention due to its pivotal role in enhancing online education. Encoder–decoder architectures have advanced traditional KGQG approaches. However, these approaches encounter challenges in achieving question diversity and grammatical accuracy. They often suffer from a disconnect between the phrasing of the question and the type of the answer entity, a phenomenon known as semantic drift. To address these challenges, we introduce LEMON, a knowledge-enhanced, type-constrained, and grammar-guided model for KGQG. LEMON enhances the input by integrating entity-related knowledge using heuristic rules, which fosters diversity in question generation. It employs a hierarchical global relation embedding with translation loss to align questions with entity types. In addition, it utilizes a graph-based module to aggregate type information from neighboring nodes. The LEMON model incorporates a type-constrained decoder to generate diverse expressions and improves grammatical accuracy through a syntactic and semantic reward function via reinforcement learning. Evaluations on benchmark datasets demonstrate LEMON's strong competitiveness. The study also examines the impact of question generation quality on question-answering systems, providing guidance for future research endeavors in this domain.
{"title":"LEMON: A Knowledge-Enhanced, Type-Constrained, and Grammar-Guided Model for Question Generation Over Knowledge Graphs","authors":"Sheng Bi;Zeyi Miao;Qizhi Min","doi":"10.1109/TLT.2025.3544454","DOIUrl":"https://doi.org/10.1109/TLT.2025.3544454","url":null,"abstract":"The objective of question generation from knowledge graphs (KGQG) is to create coherent and answerable questions from a given subgraph and a specified answer entity. KGQG has garnered significant attention due to its pivotal role in enhancing online education. Encoder–decoder architectures have advanced traditional KGQG approaches. However, these approaches encounter challenges in achieving question diversity and grammatical accuracy. They often suffer from a disconnect between the phrasing of the question and the type of the answer entity, a phenomenon known as semantic drift. To address these challenges, we introduce LEMON, a knowledge-enhanced, type-constrained, and grammar-guided model for KGQG. LEMON enhances the input by integrating entity-related knowledge using heuristic rules, which fosters diversity in question generation. It employs a hierarchical global relation embedding with translation loss to align questions with entity types. In addition, it utilizes a graph-based module to aggregate type information from neighboring nodes. The LEMON model incorporates a type-constrained decoder to generate diverse expressions and improves grammatical accuracy through a syntactic and semantic reward function via reinforcement learning. Evaluations on benchmark datasets demonstrate LEMON's strong competitiveness. The study also examines the impact of question generation quality on question-answering systems, providing guidance for future research endeavors in this domain.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"256-272"},"PeriodicalIF":2.9,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143645138","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
With the rapid advancement of generative artificial intelligence (GAI), its application in educational settings has increasingly become a focal point, particularly in enhancing students’ analytical capabilities. This study examines the effectiveness of the ChatGPT prompt framework in improving text analysis skills among students, specifically targeting readability, accuracy, completeness, logicality, and critical thinking. Conducted among high school students in Canada, the research assesses how GAI prompt frameworks significantly affect the quality of students’ analytical responses. Results showed significant improvements in all five aspects of readability, accuracy, completeness, logicality, and critical thinking, especially for students with no prior knowledge of the topic. However, enhancements in completeness and critical thinking were less pronounced, suggesting that while the ChatGPT framework substantially supports basic analytical skills, its effectiveness varies depending on the complexity of cognitive tasks and the extent of students’ existing knowledge. The study underscores the significant role that advanced GAI tools can play in modern educational environments, promoting deeper engagement with learning materials and enhancing students’ analytical abilities. It highlights the necessity of integrating these technologies to cater to diverse learning needs and cognitive challenges.
{"title":"Navigating the Textual Maze: Enhancing Textual Analytical Skills Through an Innovative GAI Prompt Framework","authors":"Xuefan Li;Tingsong Li;Minjuan Wang;Sining Tao;Xiaoxu Zhou;Xiaoqing Wei;Naiqing Guan","doi":"10.1109/TLT.2025.3539104","DOIUrl":"https://doi.org/10.1109/TLT.2025.3539104","url":null,"abstract":"With the rapid advancement of generative artificial intelligence (GAI), its application in educational settings has increasingly become a focal point, particularly in enhancing students’ analytical capabilities. This study examines the effectiveness of the ChatGPT prompt framework in improving text analysis skills among students, specifically targeting readability, accuracy, completeness, logicality, and critical thinking. Conducted among high school students in Canada, the research assesses how GAI prompt frameworks significantly affect the quality of students’ analytical responses. Results showed significant improvements in all five aspects of readability, accuracy, completeness, logicality, and critical thinking, especially for students with no prior knowledge of the topic. However, enhancements in completeness and critical thinking were less pronounced, suggesting that while the ChatGPT framework substantially supports basic analytical skills, its effectiveness varies depending on the complexity of cognitive tasks and the extent of students’ existing knowledge. The study underscores the significant role that advanced GAI tools can play in modern educational environments, promoting deeper engagement with learning materials and enhancing students’ analytical abilities. It highlights the necessity of integrating these technologies to cater to diverse learning needs and cognitive challenges.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"206-215"},"PeriodicalIF":2.9,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143521505","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-05DOI: 10.1109/TLT.2025.3539179
Kaitlyn Tracy;Ourania Spantidi
Virtual reality (VR) has emerged as a transformative educational tool, enabling immersive learning environments that promote student engagement and understanding of complex concepts. However, despite the growing adoption of VR in education, there remains a significant gap in research exploring how generative artificial intelligence (AI), such as generative pretrained transformer can further enhance these experiences by reducing cognitive load and improving learning outcomes. This study examines the impact of an AI-driven instructor assistant in VR classrooms on student engagement, cognitive load, knowledge retention, and performance. A total of 52 participants were divided into two groups experiencing a VR lesson on the bubble sort algorithm, one with only a prescripted virtual instructor (control group), and the other with the addition of an AI instructor assistant (experimental group). Statistical analysis of postlesson quizzes and cognitive load assessments was conducted using independent t-tests and analysis of variance (ANOVA), with the cognitive load being measured through a postexperiment questionnaire. The study results indicate that the experimental group reported significantly higher engagement compared to the control group. While the AI assistant did not significantly improve postlesson assessment scores, it enhanced conceptual knowledge transfer. The experimental group also demonstrated lower intrinsic cognitive load, suggesting the assistant reduced the perceived complexity of the material. Higher germane and general cognitive loads indicated that students were more invested in meaningful learning without feeling overwhelmed.
{"title":"Impact of GPT-Driven Teaching Assistants in VR Learning Environments","authors":"Kaitlyn Tracy;Ourania Spantidi","doi":"10.1109/TLT.2025.3539179","DOIUrl":"https://doi.org/10.1109/TLT.2025.3539179","url":null,"abstract":"Virtual reality (VR) has emerged as a transformative educational tool, enabling immersive learning environments that promote student engagement and understanding of complex concepts. However, despite the growing adoption of VR in education, there remains a significant gap in research exploring how generative artificial intelligence (AI), such as generative pretrained transformer can further enhance these experiences by reducing cognitive load and improving learning outcomes. This study examines the impact of an AI-driven instructor assistant in VR classrooms on student engagement, cognitive load, knowledge retention, and performance. A total of 52 participants were divided into two groups experiencing a VR lesson on the bubble sort algorithm, one with only a prescripted virtual instructor (control group), and the other with the addition of an AI instructor assistant (experimental group). Statistical analysis of postlesson quizzes and cognitive load assessments was conducted using independent t-tests and analysis of variance (ANOVA), with the cognitive load being measured through a postexperiment questionnaire. The study results indicate that the experimental group reported significantly higher engagement compared to the control group. While the AI assistant did not significantly improve postlesson assessment scores, it enhanced conceptual knowledge transfer. The experimental group also demonstrated lower intrinsic cognitive load, suggesting the assistant reduced the perceived complexity of the material. Higher germane and general cognitive loads indicated that students were more invested in meaningful learning without feeling overwhelmed.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"192-205"},"PeriodicalIF":2.9,"publicationDate":"2025-02-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143496548","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-03DOI: 10.1109/TLT.2025.3537618
Qi Lang;Minjuan Wang;Minghao Yin;Shuang Liang;Wenzhuo Song
Generative artificial intelligence (GAI) has demonstrated remarkable potential in both educational practice and research, particularly in areas, such as personalized learning, adaptive assessment, innovative teaching methods, and cross-cultural communication. However, it faces several significant challenges, including the comprehension of complex domain knowledge, technological accessibility, and the delineation of AI's role in education. Addressing these challenges necessitates collaborative efforts from educators and researchers. This article summarizes the state-of-the-art large language models (LLMs) developed by various technology companies, exploring their diverse applications and unique contributions to primary, higher, and vocational education. Furthermore, it reviews recent research from the past three years, focusing on the challenges and solutions associated with GAI in educational practice and research. The aim of the review is to provide novel insights for enhancing human–computer interaction in educational settings through the utilization of GAI. Statistical analysis reveals that the current application of LLMs in the education sector is predominantly centered on the ChatGPT series. A key focus for future research lies in effectively integrating a broader range of LLMs into educational tasks, with particular emphasis on the interaction between multimodal LLMs and educational scenarios.
{"title":"Transforming Education With Generative AI (GAI): Key Insights and Future Prospects","authors":"Qi Lang;Minjuan Wang;Minghao Yin;Shuang Liang;Wenzhuo Song","doi":"10.1109/TLT.2025.3537618","DOIUrl":"https://doi.org/10.1109/TLT.2025.3537618","url":null,"abstract":"Generative artificial intelligence (GAI) has demonstrated remarkable potential in both educational practice and research, particularly in areas, such as personalized learning, adaptive assessment, innovative teaching methods, and cross-cultural communication. However, it faces several significant challenges, including the comprehension of complex domain knowledge, technological accessibility, and the delineation of AI's role in education. Addressing these challenges necessitates collaborative efforts from educators and researchers. This article summarizes the state-of-the-art large language models (LLMs) developed by various technology companies, exploring their diverse applications and unique contributions to primary, higher, and vocational education. Furthermore, it reviews recent research from the past three years, focusing on the challenges and solutions associated with GAI in educational practice and research. The aim of the review is to provide novel insights for enhancing human–computer interaction in educational settings through the utilization of GAI. Statistical analysis reveals that the current application of LLMs in the education sector is predominantly centered on the ChatGPT series. A key focus for future research lies in effectively integrating a broader range of LLMs into educational tasks, with particular emphasis on the interaction between multimodal LLMs and educational scenarios.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"230-242"},"PeriodicalIF":2.9,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143611928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-02-03DOI: 10.1109/TLT.2025.3537802
Ahmad Chaddad;Yuchen Jiang
The concept of the Metaverse, viewed as the ultimate manifestation of the Internet, has gained significant attention due to rapid advances in technologies such as the Internet of Things (IoT) and blockchain. Acting as a bridge between the physical and virtual worlds, the Metaverse has the potential to offer remarkable experiences to its users. This study presents a comprehensive survey of Metaverse techniques, including artificial intelligence, blockchain, IoT, augmented reality, virtual reality, 5G, natural language processing, and digital twins. These Metaverse techniques lead to improved health outcomes and patient care, offering innovative treatments for complex conditions, and advancing medical education. We explore the benefits of the Metaverse by examining its effectiveness in supporting various medical applications and highlight potential research challenges and future trends for the medical Metaverse and education. Although the Metaverse is currently in its early stages, more efforts are required to enable its widespread adoption in the future.
{"title":"Integrating Technologies in the Metaverse for Enhanced Healthcare and Medical Education","authors":"Ahmad Chaddad;Yuchen Jiang","doi":"10.1109/TLT.2025.3537802","DOIUrl":"https://doi.org/10.1109/TLT.2025.3537802","url":null,"abstract":"The concept of the Metaverse, viewed as the ultimate manifestation of the Internet, has gained significant attention due to rapid advances in technologies such as the Internet of Things (IoT) and blockchain. Acting as a bridge between the physical and virtual worlds, the Metaverse has the potential to offer remarkable experiences to its users. This study presents a comprehensive survey of Metaverse techniques, including artificial intelligence, blockchain, IoT, augmented reality, virtual reality, 5G, natural language processing, and digital twins. These Metaverse techniques lead to improved health outcomes and patient care, offering innovative treatments for complex conditions, and advancing medical education. We explore the benefits of the Metaverse by examining its effectiveness in supporting various medical applications and highlight potential research challenges and future trends for the medical Metaverse and education. Although the Metaverse is currently in its early stages, more efforts are required to enable its widespread adoption in the future.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"216-229"},"PeriodicalIF":2.9,"publicationDate":"2025-02-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143521489","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The rapid changes in technology available for teaching and learning have led to a wide variety of potential tools that can be deployed to support a student's education experience. This article examines the learning interfaces for pedagogical virtual reality (VR) environments, including immersive VR (iVR). It also looks at how microlearning (ML) can be employed for instructional design at the sticking points of these interfaces. ML is an approach in which learning materials are provided in small bite-sized quantities and has been embraced as an ideal learning format for the modern learner. This study explores the research gap in ML literature regarding the ideal length of materials and modality when ML is employed for iVR. It does so through two experiments: in the first, students gave feedback on different interfaces for content and in the second, different lengths of text, video, and presentation style were tested for optimal user preference and comprehension. The findings show that preferences must be balanced against expected learning outcomes or desired level of engagement, but that fixed-point interfaces and longer texts may best be avoided. The study can be used to inform technology-enhanced learning delivery and can be used to guide policy regarding effective digital content, particularly within a VR environment.
{"title":"Microlearning in Immersive Virtual Reality: A User-Centered Analysis of Learning Interfaces","authors":"Amarpreet Gill;Derek Irwin;Linjing Sun;Dave Towey;Gege Zhang;Yanhui Zhang","doi":"10.1109/TLT.2025.3533360","DOIUrl":"https://doi.org/10.1109/TLT.2025.3533360","url":null,"abstract":"The rapid changes in technology available for teaching and learning have led to a wide variety of potential tools that can be deployed to support a student's education experience. This article examines the learning interfaces for pedagogical virtual reality (VR) environments, including immersive VR (iVR). It also looks at how microlearning (ML) can be employed for instructional design at the sticking points of these interfaces. ML is an approach in which learning materials are provided in small bite-sized quantities and has been embraced as an ideal learning format for the modern learner. This study explores the research gap in ML literature regarding the ideal length of materials and modality when ML is employed for iVR. It does so through two experiments: in the first, students gave feedback on different interfaces for content and in the second, different lengths of text, video, and presentation style were tested for optimal user preference and comprehension. The findings show that preferences must be balanced against expected learning outcomes or desired level of engagement, but that fixed-point interfaces and longer texts may best be avoided. The study can be used to inform technology-enhanced learning delivery and can be used to guide policy regarding effective digital content, particularly within a VR environment.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"165-178"},"PeriodicalIF":2.9,"publicationDate":"2025-01-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143464340","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-21DOI: 10.1109/TLT.2025.3532464
Guodong Yang;Yan Yan;Shaoqing Guo;Xiaodong Wei
In early education, reading difficulties can lead to negative outcomes. Augmented reality (AR) storybooks combine the benefits of e-books and print books, significantly aiding children's reading skills and gaining recognition from scholars and educators. However, the existing AR storybooks often overlook the design of interactive features, which may explain the inconsistent findings in research on their impact. This study aims to embed interactive tasks into AR storybooks and investigate their effects on children's reading engagement, story retelling, and reading comprehension. In total, 40 children aged eight to ten years were invited to participate in the reading activity. They were randomly assigned to an experimental group and a control group. The experimental group used AR storybooks that included interactive tasks, requiring them to complete various activities during reading. The control group used AR storybooks without interactive tasks, which provided multisensory experiences. Throughout the activity, researchers observed each child's reading engagement and completed a reading engagement assessment form. At the end of the activity, all children completed story retelling and reading comprehension tests. Finally, both groups of children participated in semistructured interviews for cross validation. The study found that children in the experimental group showed significantly higher levels of reading engagement, story retelling, and reading comprehension than children in the control group. While multimedia elements in AR storybooks can increase children's reading engagement, a large part of that engagement is driven by children's focus on AR elements. However, interactive tasks shift children's engagement more toward the story content. We also discovered that interactive tasks are a key factor in encouraging children to think actively and serve as an effective strategy for guiding them to focus on the main issues in the story. In addition, the strategy search decision feedback within the interactive tasks greatly aids children in understanding and remembering the story.
{"title":"The Impact of Embedding Interactive Tasks in Augmented Reality Storybooks on Children's Reading Engagement and Reading Comprehension","authors":"Guodong Yang;Yan Yan;Shaoqing Guo;Xiaodong Wei","doi":"10.1109/TLT.2025.3532464","DOIUrl":"https://doi.org/10.1109/TLT.2025.3532464","url":null,"abstract":"In early education, reading difficulties can lead to negative outcomes. Augmented reality (AR) storybooks combine the benefits of e-books and print books, significantly aiding children's reading skills and gaining recognition from scholars and educators. However, the existing AR storybooks often overlook the design of interactive features, which may explain the inconsistent findings in research on their impact. This study aims to embed interactive tasks into AR storybooks and investigate their effects on children's reading engagement, story retelling, and reading comprehension. In total, 40 children aged eight to ten years were invited to participate in the reading activity. They were randomly assigned to an experimental group and a control group. The experimental group used AR storybooks that included interactive tasks, requiring them to complete various activities during reading. The control group used AR storybooks without interactive tasks, which provided multisensory experiences. Throughout the activity, researchers observed each child's reading engagement and completed a reading engagement assessment form. At the end of the activity, all children completed story retelling and reading comprehension tests. Finally, both groups of children participated in semistructured interviews for cross validation. The study found that children in the experimental group showed significantly higher levels of reading engagement, story retelling, and reading comprehension than children in the control group. While multimedia elements in AR storybooks can increase children's reading engagement, a large part of that engagement is driven by children's focus on AR elements. However, interactive tasks shift children's engagement more toward the story content. We also discovered that interactive tasks are a key factor in encouraging children to think actively and serve as an effective strategy for guiding them to focus on the main issues in the story. In addition, the strategy search decision feedback within the interactive tasks greatly aids children in understanding and remembering the story.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"179-191"},"PeriodicalIF":2.9,"publicationDate":"2025-01-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143496529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-16DOI: 10.1109/TLT.2025.3530457
Nan Xie;Zhengxu Li;Haipeng Lu;Wei Pang;Jiayin Song;Beier Lu
Classroom engagement is a critical factor for evaluating students' learning outcomes and teachers' instructional strategies. Traditional methods for detecting classroom engagement, such as coding and questionnaires, are often limited by delays, subjectivity, and external interference. While some neural network models have been proposed to detect engagement using video data, they generally rely on fixed feature combinations, which fail to capture the logical connections and temporal dynamics of engagement.To address these challenges, this article introduces the MSC-Trans Engagement Detecting Network, a temporal multimodal data fusion framework that integrates a convolutional neural network (CNN) and a multilayer encoder–decoder structure. The proposed network includes two key components: first, a multilabel classifier based on ResNet and Transformer, which embeds labels into image features extracted by the CNN for high-precision classification through background inference, second, a temporal feature fusion module, which leverages an encoder–decoder structure to integrate multimodal features over time, enabling stable tracking of classroom engagement. Meanwhile, this open framework allows users to freely select feature combinations for temporal fusion based on specific scenarios and needs.The MSC-Trans Engagement Detecting Network was validated on the DAiSEE dataset, augmented with real classroom data. Experimental results demonstrate that the proposed method achieves state-of-the-art performance in continuous engagement tracking metrics, with flexible and scalable feature selection. This work offers a robust and effective approach for advancing engagement detection in educational settings.
{"title":"MSC-Trans: A Multi-Feature-Fusion Network With Encoding Structure for Student Engagement Detecting","authors":"Nan Xie;Zhengxu Li;Haipeng Lu;Wei Pang;Jiayin Song;Beier Lu","doi":"10.1109/TLT.2025.3530457","DOIUrl":"https://doi.org/10.1109/TLT.2025.3530457","url":null,"abstract":"Classroom engagement is a critical factor for evaluating students' learning outcomes and teachers' instructional strategies. Traditional methods for detecting classroom engagement, such as coding and questionnaires, are often limited by delays, subjectivity, and external interference. While some neural network models have been proposed to detect engagement using video data, they generally rely on fixed feature combinations, which fail to capture the logical connections and temporal dynamics of engagement.To address these challenges, this article introduces the MSC-Trans Engagement Detecting Network, a temporal multimodal data fusion framework that integrates a convolutional neural network (CNN) and a multilayer encoder–decoder structure. The proposed network includes two key components: first, a multilabel classifier based on ResNet and Transformer, which embeds labels into image features extracted by the CNN for high-precision classification through background inference, second, a temporal feature fusion module, which leverages an encoder–decoder structure to integrate multimodal features over time, enabling stable tracking of classroom engagement. Meanwhile, this open framework allows users to freely select feature combinations for temporal fusion based on specific scenarios and needs.The MSC-Trans Engagement Detecting Network was validated on the DAiSEE dataset, augmented with real classroom data. Experimental results demonstrate that the proposed method achieves state-of-the-art performance in continuous engagement tracking metrics, with flexible and scalable feature selection. This work offers a robust and effective approach for advancing engagement detection in educational settings.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"243-255"},"PeriodicalIF":2.9,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143645137","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-15DOI: 10.1109/TLT.2024.3513373
Minjuan Wang;John Chi-Kin Lee
{"title":"Editorial: Journey to the Future: Extended Reality and Intelligence Augmentation","authors":"Minjuan Wang;John Chi-Kin Lee","doi":"10.1109/TLT.2024.3513373","DOIUrl":"https://doi.org/10.1109/TLT.2024.3513373","url":null,"abstract":"","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"53-55"},"PeriodicalIF":2.9,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10841808","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142992955","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-01-15DOI: 10.1109/TLT.2025.3529994
Nicolas Pope;Juho Kahila;Henriikka Vartiainen;Matti Tedre
The rapid advancement of artificial intelligence and its increasing societal impacts have turned many computing educators' focus toward early education in machine learning (ML). Limited options for educational tools for teaching novice learners about the mechanisms of ML and data-driven systems presents a recognized challenge in K–12 computing education. In response, we introduce “GenAI Teachable Machine,” a visual, data-driven design platform aimed at introducing novice learners to fundamental ML concepts and workflows, particularly in the context of classifiers. Following the design science research (DSR) method, this study presents the prior recommendations, standards, codevelopment, and extensive field testing that resulted in a platform enabling young learners to express their own interest-driven ideas through codesigning and sharing personally meaningful apps. The platform improves on the design of Google's popular Teachable Machine 2 by its ability to create a standalone app by defining one or more actions to be triggered by each classifier result, and deploy that app to other devices. It also enables one to distribute the collection of training data among many users. In addition to the DSR process, this article presents findings from usability lab tests (N = 8) and 6-h classroom projects involving fourth and seventh grade children (N = 213). The results show that children who had no experience of ML were able to navigate through the workflow and turn their own ideas into concrete ML-based apps. The majority of children were able to reflect and present, in their own words, their working process using data-driven (design) thinking concepts and insights.
{"title":"Children's AI Design Platform for Making and Deploying ML-Driven Apps: Design, Testing, and Development","authors":"Nicolas Pope;Juho Kahila;Henriikka Vartiainen;Matti Tedre","doi":"10.1109/TLT.2025.3529994","DOIUrl":"https://doi.org/10.1109/TLT.2025.3529994","url":null,"abstract":"The rapid advancement of artificial intelligence and its increasing societal impacts have turned many computing educators' focus toward early education in machine learning (ML). Limited options for educational tools for teaching novice learners about the mechanisms of ML and data-driven systems presents a recognized challenge in K–12 computing education. In response, we introduce “GenAI Teachable Machine,” a visual, data-driven design platform aimed at introducing novice learners to fundamental ML concepts and workflows, particularly in the context of classifiers. Following the design science research (DSR) method, this study presents the prior recommendations, standards, codevelopment, and extensive field testing that resulted in a platform enabling young learners to express their own interest-driven ideas through codesigning and sharing personally meaningful apps. The platform improves on the design of Google's popular Teachable Machine 2 by its ability to create a standalone app by defining one or more actions to be triggered by each classifier result, and deploy that app to other devices. It also enables one to distribute the collection of training data among many users. In addition to the DSR process, this article presents findings from usability lab tests (<italic>N</i> = 8) and 6-h classroom projects involving fourth and seventh grade children (<italic>N</i> = 213). The results show that children who had no experience of ML were able to navigate through the workflow and turn their own ideas into concrete ML-based apps. The majority of children were able to reflect and present, in their own words, their working process using data-driven (design) thinking concepts and insights.","PeriodicalId":49191,"journal":{"name":"IEEE Transactions on Learning Technologies","volume":"18 ","pages":"130-144"},"PeriodicalIF":2.9,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10842355","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143361071","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"教育学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：