Explore the vast range of titles available.

Federated learning aims to collaboratively train a machine learning model with possibly geo-distributed workers, which is inherently communication constrained. To achieve communication efficiency, the conventional federated learning algorithms allow the worker to decrease the communication frequency by training the model locally for multiple times. Conventional federated learning architecture, inherited from the parameter server design, relies on highly centralised topologies and large nodes-to-server bandwidths, and convergence property relies on the stochastic gradient descent training in local, which usually causes the large end-to-end training latency in real-world federated learning scenarios. Thus, in this study, the authors propose the adaptive partial gradient aggregation method, a gradient partial level decentralised federated learning, to tackle this problem. In FedPGA, they propose a partial gradient exchange mechanism that makes full use of node-to-node bandwidth for speeding up the communication time. Besides, an adaptive model updating method further reduces the convergence rate by adaptive increasing the step size of the stable direction of gradient descent. The experimental results on various datasets demonstrate that the training time is reduced up to $14 \\times $14× compared to baselines without accuracy degrade.

This study analyses the development of socio-critical mathematical thinking (SCMT) in primary and secondary school students through problem-solving tasks based on real-life contexts. From a socio-critical perspective of mathematics education, based on Paulo Freire's emancipatory pedagogy and Ole Skovsmose's critical mathematics education, an intervention was designed in two public institutions in Armenia, Colombia. The strategy consisted of students identifying topics of interest to them, writing an opinion piece and supporting it with reliable statistical data, promoting reflection and critical analysis of social, political, economic and environmental phenomena. Thirty-six students participated, of whom 83.3% expressed, in interviews and written assignments, a change in their perception of the role of mathematics. The results show a transition from decontextualised exercises to the formulation of projects with a socio-critical approach, strengthening argumentation and democratic participation. It is concluded that the incorporation of real contexts and critical data analysis favour ECM and contribute to the formation of critical citizens committed to their environment.

This paper seeks to understand the impact of a training program on 19 pre-service primary school teachers’ perceptions towards educational robotics (ER). The training program is based on a reflective process of design and implementation of a learning scenario during the practicum, using a pre-experimental design. Quantitative data were collected through a questionnaire applied at three moments of the intervention: pre-intervention, post-intervention 1 (19 weeks after), and post-intervention 2 (37 weeks after). The results show that the features of the proposed training program positively influenced the pre-service teachers’ (PST) perceptions towards ER. Experiencing curricular integration of ER and participation in a reflective process of learning scenario design positively influenced their perceptions in post-intervention 1. After experiencing the integration of ER in the practicum class, PST adjusted their positive perceptions in post-intervention 2. PST also displayed a decrease in neutrality in their perceptions in post-intervention 1 and post-intervention 2. Given the limited sample, it’s not possible to generalize these results, however they have implications for initial teacher training programs dedicated to technology integration. PST must be allowed to confront their preconceived perceptions of integrating technology into teaching and learning processes with the reflective process of designing and implementing a lesson plan that integrates technology during the practicum.

Although user-generated microgames, defined as very simple games made by non-professionals on open platforms, are popular and appear to have considerable advantages in facilitating learning, further exploration is needed to establish their potential in instructional practices. The present study investigates the design, participation and experience of teaching and learning facilitated by user-generated microgames on an open educational platform. Through an exploratory experiment research method, four elementary school teachers designed and implemented microgame-based learning utilising these very small games on GeoGebra Classroom attended by 129 students. Data were gathered from lesson plans, classroom activity records and self-reflection questionnaires. This study revealed that teachers designed learning with various user-generated microgames and debriefing methods respecting learning content, but they shared comparatively similar scenarios by inserting microgame-based learning into the middle of the main session. The completion rate for the debriefing activity is minimum although the total joining times overshoot the number of students. Teachers found that user-generated microgames are acceptable to orchestrate short serious gaming sessions even though they are limited to one player with basic interfaces. Notwithstanding several disadvantages of these microgames recognised by students, such as missing learning instructions and inadequate interfaces, they so far enjoy learning by playing the games. The most critical implication of this study is to provide sufficient instructions and additional time for microgaming sessions in elementary schools to ensure sustainable completion of the briefing, playing and debriefing activities.

The digitalization and globalization of society and the corresponding impact on the rules of the labor market is shifting the education sector toward new pedagogical approaches that integrate wholly online methodologies. Sustainable Development Goal 4 advocates for inclusive and equitable quality education that promotes lifelong learning opportunities, and, as we have seen during the COVID-19 lockdown, online learning can play a key role. In a context where lifelong learning becomes crucial to maintaining graduates’ employability, the innovative teaching methodologies that promote employable competencies in online environments are especially desirable. With the purpose of improving the employability of students, this article analyses the impact of introducing the Authentic Learning Scenarios (ALS) paradigm in an online environment. We develop a quasi-experimental design. Based on the nine ALS criteria and their application to e-learning, we redesign a course in a business degree program. Data from 135 students were collected, with special focus on achieving general competences. We compare the perception of the competency profile attained between a group of students who took the course before incorporating the ALS paradigm and another group that took it once it had been redesigned. Results show that redesigning the course enables students to perceive the learning process as more authentic, as well as acquiring a more advanced competence profile. Besides this, it has been shown that technology can contribute to building cognitive authenticity in virtual classrooms, without the need for face-to-face internships, which are often not a feasible option for students of online programs.

The competence enhancement of college students is the overarching objective of curriculum construction and reform, which is crucial for the construction of first-class majors and disciplines. Principles of Genetic Engineering is an important professional course for the students of Biotechnology and Biomedical Engineering that are the national first-class construction majors at Guizhou Medical University in China. To increase the personalized self-directed and mutual learning opportunities for undergraduate and graduate students in biomedical science and engineering, an innovative teaching was designed to implement a vivid and interactive learning mode based on online and offline platform, smart teaching tools and scientific research achievements. After multiple rounds of teaching practice, continuous improvement and iterative updates, a 'three-stage and three-guidance' learning model was established to integrate pre-class, in-class, and post-class stages under the guidance of self-learning, question and research. In result, this curriculum was identified as the provincial 'Golden Course' and the average overall satisfaction rate for students reached 96.88%. Totally, this pattern can motivate students' intrinsic passion and interests in learning and enhance their scientific research thinking, innovation ability, teamwork and comprehensive skills in solving practical problems, supporting that the effective curriculum reform can promote the cultivation of high-quality talents and support the construction of first-class majors and disciplines.

Hybrid learning spaces may be described as physical environments enhanced by digital technologies, which enable learning scenarios involving both in-person and online participation. This article presents a hybrid learning space designed for higher education. The design of the space has been informed by Lefebvre’s design principles: (a) spatial practice enabling flexible usage scenarios, (b) representations of space conveying openness and adaptability, and (c) representational spaces supporting experiences of presence in both physical and digital form. The article describes design characteristics guiding the implementation of the hybrid learning space and explains corresponding design decisions, such as the use of a wall-sized projection. Further, the article introduces affordances and usage scenarios of the hybrid learning space developed. Moreover, an evaluation study of the hybrid learning space is conducted by means of a 360°-based virtual field trip (VFT). The VFT, led by an educator, serves as preparation for a field trip (FT) to a composting plant two weeks later. Participants of both VFT and FT (N = 11) completed a questionnaire addressing psychological constructs related to learning, including motivation, emotion, immersion, presence, and cognitive load. We report the results of the VFT alongside those of the FT as a baseline. Some notable differences, for example in social presence, suggest areas for further development of the hybrid learning space. Overall, the study characterises key features of hybrid learning spaces, identifies their contribution to high-quality teaching and provides inspirations for their further development.

The health crisis caused by COVID-19 compelled university teachers to adapt their learning scenarios to new technology-mediated contexts. This paper analyses teaching and learning experiences, strategies and lessons learned during the lockdown period at the Faculty of Education of the Universitat Autònoma de Barcelona (N=29 teachers, 227 students). The results reveal that participants experienced difficulties (lack of literacy in online pedagogies and work overload among lecturers; privation of physical presence and fluent communication among students). Teachers acquired knowledge around digital technologies and are predisposed to learn about innovative teaching methods supported by technologies. Students are dissatisfied with the learning experience, although they value the opportunities for flexible learning and saving time on commuting. Teaching strategies were less innovative and active than usual, and usually involved a combination of synchronous time for lectures and resolving problems, and self-study. Nevertheless, students valued more traditional teaching strategies (i.e. combinations of lectures and tutoring). The paper concludes that the teachers’ view of the use of digital technologies has improved, although training is needed to make effective use of such technologies for active learning and innovative approaches to teaching.

Academic data predictions are significantly important for improving the overall education system's effectiveness by providing early identification of weak students and personalized learning strategies. This paper proposes a deep learning model for the identification of weak and strong students using ensemble learning and multiparametric analysis. It combines several machine learning algorithms, including Naive Bayes, Support Vector Machines, Multi-Layer Perceptron, and Logistic Regression using an ensemble learning approach to enhance the model’s performance. Additionally, a custom 1D Convolutional Neural Network (CNN) is designed for classification. It utilizes multiparametric analysis to identify weak and strong students considering various parameters such as age, academic performance, location, and online learning behavior. The evaluation results indicate the performance of the proposed model has been improved in comparison to MLA FIS, SHMM, and DRL by 16.5%, 5.5%, and 2.4%, in terms of precision, 16.4%, 6.5%, and 3.5 % in terms of accuracy and 10.4%, 2.5% and 6.5% in terms of recall. These improvisations described that the model is efficient for multidomain feature extraction, ensemble classification, and high-variance feature selection, which result in a deeper understanding of student performance. Article HighlightsA deep learning model using ensemble learning and multiparametric analysis for identifying weak and strong students.The proposed deep learning model significantly improves identifying weak and strong students, boosting educational effectiveness.By combining various machine learning techniques with a custom CNN, the model enhances precision, accuracy, and recall.It offers a more detailed analysis of student performance, leveraging factors like age and online behavior for better personalized learning.

Adaptive decision‐making allows decision‐makers to plan long‐term coastal infrastructure under uncertain sea level rise projections. To date, economic assessments of adaptive decision‐making that take into account future learning about sea level rise uncertainty are rare and the existing ones have relied on simple quantification of future learning not validated against sea level science. To address this gap, we develop an economic adaptive decision‐making framework that takes into account future learning about sea level rise uncertainty and apply it to a coastal case study in Lübeck, Germany, to answer the question of how adaptation to sea level rise can be improved through adaptive adaptation pathways as opposed to non‐adaptive pathways. To address this question, we use a Markov decision process to formulate the stochastic optimization problem. We quantify future learning about sea level rise uncertainty through sea level rise learning scenarios based on and validated against the latest scenarios of the Intergovernmental Panel on Climate Change. Our case study results show that the city of Lübeck is currently under‐protected against storm surges and that immediate adaptation actions are advisable in the face of future sea level rise. We find that adaptive adaptation pathways, in contrast to non‐adaptive pathways, generate sea level rise thresholds for adaptation actions that are similar across climate change scenarios and can reduce expected costs up to 1.8%. Plain Language Summary Climate change is causing sea levels to rise as land ice melts under higher air temperatures. By 2100, sea levels are projected to rise between 28 and 101 cm, depending on the extent of future global warming. This will increase the risk of coastal flooding in the future and require adaptation actions. Planning for coastal adaptation to sea level rise is challenged by long‐lived protective infrastructure and uncertain projections of sea level rise. Adaptive decision‐making methods that specifically incorporate future learning about the uncertainty of sea level rise can address this challenge. For example, observing 30 cm of sea level rise in 2060 will lead to different projections from 2050 onwards and require different adaptation actions than observing 70 cm of sea level rise in 2060. To date, economic adaptation studies that account for future learning about sea level rise uncertainty are rare, and those that do exist have relied on simple methods. We develop a decision framework to address this research gap and apply it to the city of Lübeck on the Baltic Sea in Germany. Our results show that the city of Lübeck is currently under‐protected against storm surges and that immediate adaptation actions are advisable from an economic perspective. Key Points We develop adaptive adaptation pathways that incorporate learning about sea level rise uncertainty based on future observations Adaptive adaptation pathways generate sea level rise thresholds for adaptation actions that can be similar across climate change scenarios Adaptive adaptation pathways can reduce expected costs compared to non‐adaptive pathways by up to 1.8% in our study