Article
Robotics
Weiyuan Li, Ruoxin Hong, Jiwei Shen, Liang Yuan, Yue Lu
Summary: Substantial progress has been made in embodied visual navigation based on RL, but the presence of interactable objects in real cluttered scenes poses a challenge for the ego-centric visual agent. To address this, the authors propose a transformer-based memory to empower agents with historical interactive information and utilize a surrogate objective to predict the next waypoint, facilitating representation learning and bootstrapping RL.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Computer Science, Interdisciplinary Applications
Isabel C. Gil-Garcia, Ana Fernandez-Guillamon, M. Socorro Garcia-Cascales, Angel Molina-Garcia
Summary: Traditional face-to-face learning is being replaced or combined by virtual and digital campuses, with a variety of online activities such as H5P becoming more interactive and creative teaching strategies. A comparison between traditional online and interactive H5P activities was conducted, showing that while there was no significant improvement in average grades, students found the interactive H5P activity easier to do and felt more motivated. Teachers also noted positive effects on student participation and motivation. These results highlight the importance of incorporating digital methods alongside traditional approaches to meet the diverse needs of students.
COMPUTER APPLICATIONS IN ENGINEERING EDUCATION
(2023)
Article
Psychology, Multidisciplinary
Silvia Molina Roldan, Jesus Marauri, Adriana Aubert, Ramon Flecha
Summary: Research indicates that students without special educational needs benefit from participating in interactive learning activities with students with special educational needs by learning to respect others, accept differences, develop patience, cultivate the ability to help others learn and behave better, and derive satisfaction from these interactions.
FRONTIERS IN PSYCHOLOGY
(2021)
Article
Construction & Building Technology
Phillip Stoffel, Laura Maier, Alexander Kuempel, Thomas Schreiber, Dirk Mueller
Summary: Advanced building control strategies like model predictive control and reinforcement learning, combined with forecasts for weather, occupancy, and energy prices, have the potential to reduce buildings' energy consumption and CO2 emissions. However, there is a lack of comparability among different control algorithms in the literature. This paper extensively evaluates six advanced control algorithms based on quantitative and qualitative key performance indicators.
ENERGY AND BUILDINGS
(2023)
Article
Education & Educational Research
Ching-Yi Wang, Cheng-Han Lin
Summary: COVID-19 has forced universities to switch to online teaching, posing challenges for design courses that emphasize on-site operation and face-to-face discussions. This study investigates the suitability and teaching strategies for distance teaching in design courses. The results show the importance of creating a communication environment and real interactive feeling, the suitability of hybrid design courses for distance teaching, the prominence of hard requirements in design practice courses, and the need to strengthen social presence in distance teaching.
INTERACTIVE LEARNING ENVIRONMENTS
(2023)
Review
Green & Sustainable Science & Technology
Ching-Yi Chang, Hui-Chun Chu
Summary: This study used bibliographic analysis to validate the research trends of technology-enhanced digital storytelling in education and educational research. Through analyzing top journals, frequently cited articles, keywords, research methods, and application domains, the study identified four thematic clusters and highlighted the dominant application domain of activity in DS research. The results provide recommendations for future research.
Article
Chemistry, Multidisciplinary
Pavel Ugwitz, Ondrej Kvarda, Zuzana Jurikova, Cenek Sasinka, Sascha Tamm
Summary: This paper discusses the methods, issues, and solutions of eye-tracking in virtual reality. It proposes a workflow and software architecture for the entire experimental scenario and provides examples of eye-tracking data collection and evaluation.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Artificial Intelligence
Doaa Mohamed Elbourhamy, Ali Hassan Najmi, Abdellah Ibrahim Mohammed Elfeky
Summary: Modern approaches in education technology, such as electronic books, infographics, and mobile applications, are being used to improve education quality and learning levels. Educational data mining is a popular method for predicting student performance and monitoring progress. A model was developed to inform students about their performance in a computer networks course, and the effectiveness of infographics for teaching was proven.
PEERJ COMPUTER SCIENCE
(2023)
Article
Chemistry, Analytical
Hung Son Nguyen, Francisco Cruz, Richard Dazeley
Summary: In this paper, the authors propose a method called Broad-Persistent Advising (BPA) that retains and reuses processed information to help trainers give more general advice and speed up the learning process in robotics. The results of testing the approach in two continuous robotic scenarios showed that the agent's learning speed increased by up to 37% in terms of reward points, while maintaining the same number of interactions required for the trainer compared to the DeepIRL approach.
Article
Computer Science, Artificial Intelligence
Ghodai Abdelrahman, Qing Wang
Summary: Teaching plays a crucial role in human learning, but current machine teaching methods often neglect the underlying learning concepts in a learning task by directly assessing individual training samples. In this paper, a novel method called Knowledge Augmented Data Teaching (KADT) is proposed to optimize the data teaching strategy by tracking the knowledge progress of a student model over multiple learning concepts. The experimental results show that the KADT method consistently outperforms state-of-the-art methods in various machine learning tasks, including knowledge tracing, sentiment analysis, movie recommendation, and image classification.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Green & Sustainable Science & Technology
Dianne Forbes, Dilani Gedera, Maggie Hartnett, Ashwini Datt, Cheryl Brown
Summary: Investigating students' lived experiences of online learning can provide valuable insights for effective teaching. This paper utilizes the lessons learned from pandemic teaching and learning to inform sustainable approaches. Through research on New Zealand university students' perspectives and experiences of online learning, this article offers insights into supporting students with sustainable strategies and the development of long-term learning strategies. The findings highlight the importance of support, communication, and engagement in addressing challenges and guiding future teaching and learning practices. The study also proposes institutional support for sustainable strategies for both students and staff.
Review
Psychology, Multidisciplinary
Leire Ugalde, Maite Santiago-Garabieta, Beatriz Villarejo-Carballido, Lidia Puigvert
Summary: Inclusive education plays a crucial role in promoting academic learning and cognitive development for children with SEN, and research supports the effectiveness of interactive learning environments in achieving these goals.
FRONTIERS IN PSYCHOLOGY
(2021)
Article
Education & Educational Research
Mohamed M. Mostafa
Summary: This study conducts a comprehensive bibliometric analysis of the journal Interactive Learning Environments (ILE) to examine impactful authors, citation patterns, collaboration networks, and emerging trends. Results show the most productive authors, sparse author collaboration, and a global divide between developed and developing nations. The multiple correspondence analysis (MCA) reveals the depth and breadth of the journal's focus.
INTERACTIVE LEARNING ENVIRONMENTS
(2022)
Article
Multidisciplinary Sciences
Phillip Kremer, Leonard Richter, Leander Melms, Claus F. F. Vogelmeier, Juergen R. R. Schaefer
Summary: The COVID-19 pandemic has brought challenges to the medical community worldwide. To minimize infection risk, patient-unrelated classes can be held digitally. This study presents a student-initiated, web-based teaching approach called "From symptom to diagnosis." Through the presentation of rare disease case reports in a symptom-focused manner, followed by in-depth discussions about differential diagnosis, the online seminar successfully improved students' diagnostic skills with high participant engagement.
Article
Robotics
Mayank Mittal, Calvin Yu, Qinxi Yu, Jingzhou Liu, Nikita Rudin, David Hoeller, Jia Lin Yuan, Ritvik Singh, Yunrong Guo, Hammad Mazhar, Ajay Mandlekar, Buck Babich, Gavriel State, Marco Hutter, Animesh Garg
Summary: We introduce Orbit, a modular framework for robot learning powered by Nvidia Isaac Sim. It allows easy creation of robotic environments with realistic scenes and precise body simulation. With benchmark tasks of different difficulty levels and support for diverse observations and action spaces, Orbit enables training reinforcement learning policies and collecting demonstration datasets efficiently. It comes with multiple robotic platforms, sensor modalities, motion generators, benchmark tasks, and learning library wrappers, aiming to support various research areas and foster interdisciplinary collaborations.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2023)
Article
Computer Science, Cybernetics
Tzu-Chi Yang, Meng Chang Chen, Sherry Y. Chen
UNIVERSAL ACCESS IN THE INFORMATION SOCIETY
(2020)
Article
Multidisciplinary Sciences
Yi-Ting Huang, Yeali S. Sun, Meng Chang Chen
Summary: This study develops a sequence-to-sequence neural network called TagSeq to analyze the behavior of malware. By investigating recorded sequences of Windows API calls and generating tags to label malicious behavior, it helps security analysts recognize the actions and intentions of malware.
Article
Education & Educational Research
Tzu-Chi Yang, Chung-Yuan Chang
Summary: This study proposes a data-driven approach that considers both institutional data and social media news for predicting students' career decisions. The results suggest that this approach achieved a higher performance in the prediction task, and it discusses how this approach can support students' career development in the university setting.
EDUCATION AND INFORMATION TECHNOLOGIES
(2023)
Article
Computer Science, Artificial Intelligence
Hsin-Chih Yang, Ming-Chuan Yang, Guo-Wei Wong, Meng Chang Chen
Summary: In this study, a self-attention-based neural network is used to predict the anomalies of fine particulate matter (PM2.5), and extreme value theory (EVT) is employed to solve the rarity issue of anomalous data. Experiments show that the proposed model achieves improvements of 478% in F1 score and 286% in Matthews correlation coefficient (MCC) compared to the fully connected network, and 229% in F1 and 148% in MCC compared to the typical transformer trained with the traditional loss function.
IEEE INTELLIGENT SYSTEMS
(2023)
Article
Multidisciplinary Sciences
George William Kibirige, Ming-Chuan Yang, Chao-Lin Liu, Meng Chang Chen
Summary: The paper proposes a new Remote Transported Pollutants (RTP) model that predicts local PM2.5 concentrations more accurately by integrating deep learning components and learning from various domains. Extensive experiments using real-world data demonstrate that the RTP model outperforms other models in predicting pollution events at different time ranges.
Article
Multidisciplinary Sciences
George William Kibirige, Chiao Cheng Huang, Chao Lin Liu, Meng Chang Chen
Summary: PM2.5 prediction is crucial for governments to establish effective policies in controlling atmospheric pollution and protecting public health. We propose a composite neural network model that incorporates aerosol optical depth, weather data, and ocean wind features collected from satellites, overcoming the limitations of traditional machine learning methods. Our findings show that the proposed architecture significantly improves overall performance compared to individual components and ensemble models. The monthly analysis also demonstrates its superiority in regions where land-sea breezes dominate PM2.5 accumulation.
SCIENTIFIC REPORTS
(2023)
Article
Computer Science, Artificial Intelligence
Ming-Chuan Yang, Meng Chang Chen
Summary: This work investigates the framework and statistical performance guarantee of composite neural network for solving complicated applications. It is composed of pre-trained and non-instantiated neural network models, connected in a rooted directed acyclic graph. The advantages of adopting pre-trained models as components are benefiting from domain experts' intelligence and diligence, and saving effort in data acquisition and model training. The study proposes the framework of a composite network and proves its superior performance compared to its pre-trained components in a high probability. Empirical evaluations of the PM2.5 prediction application show that the constructed composite neural network models outperform other machine learning models.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Interdisciplinary Applications
Tzu-Chi Yang, Zhi-Shen Lin
Summary: Computational thinking is essential in the current era and learning programming is the most effective way to develop it. Introducing computational thinking and programming at an early age is recommended. Graphic organizers serve as a bridge between students' existing knowledge and new learning, enhancing their learning process. The study found that using graphic organizers improved elementary school students' computational thinking, programming skills, and learning experiences.
COMPUTERS & EDUCATION
(2024)
Article
Education & Educational Research
Tzu-Chi Yang, Sherry Y. Chen
Summary: Individual differences among learners, particularly cognitive styles, are crucial for predicting learning behavior. This study examines the overlaps between two cognitive style dimensions, field dependence/independence and Holism/Serialism, in terms of online learning behavior using Lag Sequential Analysis. The results reveal significant overlaps in comprehensive/local and dynamic/fixed approaches.
INTERACTIVE LEARNING ENVIRONMENTS
(2023)
Article
Computer Science, Hardware & Architecture
Yi-Ting Huang, Chi Yu Lin, Ying-Ren Guo, Kai-Chieh Lo, Yeali S. Sun, Meng Chang Chen
Summary: This research proposes a malicious behavior analysis system based on the MITRE ATT&CK framework, which can effectively detect and respond to cyber threats and provides a mapping from malicious behaviors to ATT&CK techniques and API calls.
IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Faisal Ghaffar, George William Kibirige, Chih-Ya Shen, Meng Chang Chen
2020 IEEE GLOBAL COMMUNICATIONS CONFERENCE (GLOBECOM)
(2020)
Proceedings Paper
Computer Science, Information Systems
Ting-Wei Hsu, Chung-Chi Chen, Hen-Hsen Huang, Meng Chang Chen, Hsin-Hsi Chen
WWW'20: COMPANION PROCEEDINGS OF THE WEB CONFERENCE 2020
(2020)
Proceedings Paper
Computer Science, Information Systems
Yang Tzu-Chi
2020 IEEE 20TH INTERNATIONAL CONFERENCE ON ADVANCED LEARNING TECHNOLOGIES (ICALT 2020)
(2020)
Article
Computer Science, Theory & Methods
Shun-Wen Hsiao, Yeali S. Sun, Meng Chang Chen
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2020)
Article
Computer Science, Interdisciplinary Applications
Elise Ameloot, Tijs Rotsaert, Thomas Ameloot, Bart Rienties, Tammy Schellens
Summary: This study investigates the impact of using learning analytics to support students' autonomy and competence needs in a blended learning environment. The findings show that teachers' adaptation based on learning analytics positively influences students' satisfaction with the adapted learning environment. However, students' basic psychological needs vary depending on the face-to-face workshop subject. The study emphasizes the importance of thoughtful blended learning course design and provides recommendations for effective learning analytics utilization in university settings.
COMPUTERS & EDUCATION
(2024)
Article
Computer Science, Interdisciplinary Applications
Tzu-Chi Yang, Zhi-Shen Lin
Summary: Computational thinking is essential in the current era and learning programming is the most effective way to develop it. Introducing computational thinking and programming at an early age is recommended. Graphic organizers serve as a bridge between students' existing knowledge and new learning, enhancing their learning process. The study found that using graphic organizers improved elementary school students' computational thinking, programming skills, and learning experiences.
COMPUTERS & EDUCATION
(2024)
Article
Computer Science, Interdisciplinary Applications
Yuqin Yang, Kaicheng Yuan, Gaoxia Zhu, Lizhen Jiao
Summary: The study finds that the design of collaborative analytics-enhanced reflective assessment can promote conducive epistemic emotions to knowledge building among undergraduates, and enriches our understanding of the relationships between metacognition, epistemic emotions, and knowledge building practices.
COMPUTERS & EDUCATION
(2024)
Article
Computer Science, Interdisciplinary Applications
Vishal Kiran Kuvar, Jeremy N. Bailenson, Caitlin Mills
Summary: Recent research suggests that students' minds often wander off-task during learning, regardless of the learning modality. This study explores the potential of virtual reality (VR) to reduce task-unrelated thoughts (TUT) and finds that learning with VR leads to lower TUT and better performance.
COMPUTERS & EDUCATION
(2024)
Article
Computer Science, Interdisciplinary Applications
Yiming Liu, Jeremy Tzi Dong Ng, Xiao Hu, Zhengyang Ma, Xiaoyan Lai
Summary: This study examines teachers' usage intention and behavior towards the Game Learning Analytics (GLA) system in K-12 classrooms. The study found that personal, environmental, and technological factors influenced teachers' intention and behavior, and that technostress moderated the intention-behavior relationship. The study also identified the heterogeneity of GLA usage among teachers with different individual characteristics.
COMPUTERS & EDUCATION
(2024)
Article
Computer Science, Interdisciplinary Applications
Romina Cachia, Artur Pokropek, Nikoleta Giannoutsou
Summary: This article introduces a shortened version of the European Commission's SELFIE tool for measuring the digital capacity of schools. Two shorter measurement instruments, called midi-SELFIE and mini-SELFIE, are proposed based on the original tool. The validity and uses of these shortened versions are explored through various cases and compared to the complete instrument. The results suggest that the shortened versions of SELFIE are reliable alternatives for specific purposes.
COMPUTERS & EDUCATION
(2024)
