Learning Scenarios Research Articles

Unveiling learners’ intentions toward influencer-led education: an integration of qualitative and quantitative analysis

ABSTRACT In the rapidly evolving landscape of social media, the emergence of Influencer-led Education as a novel learning model presents a paradigm shift in the mobile learning era. To illuminate the underpinnings of learners’ inclination to embrace this model, this study adopts an integrated methodological framework, encompassing grounded theory and fuzzy-set Qualitative Comparative Analysis (fsQCA). Based on 17 in-depth interviews, an Action-Decision Framework of new model is established. Subsequently, a examination of 150 questionnaires through fsQCA uncovers diverse configurations of learners’ adoption intention of Influencer-led Education. The outcomes of the fsQCA disclose five configurations associated with high adoption intention and two configurations leading to low intention, highlighting heterogeneous representations among distinct learner cohorts within the decision framework. This research contributes to the expansion and deepening of theories pertaining to learners’ engagement with Influencer-led Education, while providing insights into the intricate mechanisms driving learning behaviors in the realm of novel learning scenarios.

Phishing Website Detection: An In‐Depth Investigation of Feature Selection and Deep Learning

ABSTRACTCloud and fog computing technologies benefit from integrating AI‐driven phishing detection as it enhances security, scalability, real‐time reaction, and privacy. Nowadays, there is a noticeable rise in illegal activity taking place online. One of the illicit cybersecurity practices is phishing, in which hackers trick consumers by pretending to be authentic websites and spoofing them to obtain sensitive user information. Phishing attacks, regrettably, have increased dramatically in recent years, according to research. Machine learning (ML) and deep learning (DL) techniques have shown encouraging progress in thwarting these attacks. Consequently, we employed DL and ML techniques to identify phishing websites in this study. This article presents four scenarios in both ML and DL models. Two are proposed in ML, while the others are employed in DL. The outcomes of four scenarios were contrasted to determine which algorithm performed better at distinguishing between legal and illicit websites. Many popular ML techniques were used, including K‐nearest neighbour, random forest (RF), decision trees, and SVMs. PCA and Importance Features are implemented in both ML scenarios to find the best features. RF successfully reached an accuracy of 97.82% using the Importance Feature technique. However, the PCA method failed to improve the performance of ML algorithms. As a result of ML‐based scenarios, 98 features are selected for the final deep learning scenarios. In DL‐based scenarios, algorithm architectures are essential to avoid overfitting and bias due to various hyperparameters. Thus, in the third scenario, our aim focuses on DL architecture design. Multilayer perceptron and convolutional neural networks (CNNs) are employed to detect phishing websites. Finally, our proposed 1D CNN model, using stratified k‐fold cross‐validation, outperformed the classical ML algorithm, achieving 98.94% accuracy and 0.99 AUC‐ROC score in detecting phishing websites.

Optimizing Robotic Manipulation With Decision-RWKV: A Recurrent Sequence Modeling Approach for Lifelong Learning

Abstract Models based on the transformer architecture have seen widespread application across fields such as natural language processing (NLP), computer vision, and robotics, with large language models (LLMs) like ChatGPT revolutionizing machine understanding of human language and demonstrating impressive memory capacity and reproduction capabilities. Traditional machine learning algorithms struggle with catastrophic forgetting, detrimental to the diverse and generalized abilities required for robotic deployment. This article investigates the receptance weighted key value (RWKV) framework, known for its advanced capabilities in efficient and effective sequence modeling, integration with the decision transformer (DT), and experience replay architectures. It focuses on potential performance enhancements in sequence decision-making and lifelong robotic learning tasks. We introduce the decision-RWKV (DRWKV) model and conduct extensive experiments using the D4RL database within the OpenAI Gym environment and on the D’Claw platform to assess the DRWKV model’s performance in single-task tests and lifelong learning scenarios, showing its ability to handle multiple subtasks efficiently. The code for all algorithms, training, and image rendering in this study is available online (open source).

ISOUP-SymRF: Symbolic feature ranking with random forests in online multi-target regression and multi-label classification

The task of feature ranking has received considerable attention across various predictive modelling tasks in the batch learning scenario, but not in the online learning setting. Available methods that estimate feature importances on data streams have so far predominantly focused on ranking the features for the tasks of classification and occasionally multi-label classification. We propose a novel online feature ranking method for online multi-target regression iSOUP-SymRF, which estimates feature importance scores based on the positions at which a feature appears in the trees of a random forest of iSOUP-Trees, and additionally extend it to task of online feature ranking for multi-label classification. By utilizing iSOUP-Trees, which can address multiple structured output prediction tasks on data streams, iSOUP-SymRF promises feature ranking across a variety of online structured output prediction tasks. We examine the ranking convergence of iSOUP-SymRF in terms of the methods’ parameters, the size of the ensemble and the number of selected features, as well as their stability under different random seeds. Furthermore, to show the utility of iSOUP-SymRF and its rankings we use them in conjunction with two state-of-the-art online multi-target regression and multi-label classification methods, iSOUP-Tree and AMRules, and analyze the impact of adding features according to the rankings obtained from iSOUP-SymRF.

Integrating behavioral and neurophysiological insights: High trait anxiety enhances observational fear learning.

Observational fear learning delineates the process by which individuals learn about potential threats through observing others' reactions. Prior research indicates that individuals with high trait anxiety (HTA) manifest pronounced fear responses in direct fear learning scenarios. However, the specific influence of trait anxiety on observational fear learning remains insufficiently explored. This study aimed to fill this gap by examining 64 university students, divided equally between those with HTA and low trait anxiety (LTA), selected from an initial pool of 483 participants. Participants were subjected to observational fear learning tasks, and their behavioral responses, physiological reactions, and brain activations were recorded. Results demonstrated that HTA participants exhibited differentiated skin conductance responses to threat and safety stimuli during the observational fear acquisition phase, notwithstanding prior assurances against shock delivery. Furthermore, during the direct test phase, HTA participants reported significantly elevated fear and shock expectancy ratings for both types of stimuli, in contrast to their LTA counterparts. Neuroimaging data, derived via functional near-infrared spectroscopy (fNIRS) revealed heightened medial prefrontal cortex activation in HTA participants when directly facing threats. This study systematically explores the influence of high trait anxiety on observational fear learning, uncovering that HTA individuals exhibit excessive fear responses. These findings highlight the critical role of trait anxiety as a significant risk factor in the development of anxiety disorders.

Uso de los recursos didácticos en la enseñanza de las ciencias experimentales química y biología: Una revisión de la literatura

In educational work, it is important to recognize the ingenuity of the teacher to face the challenges in the pedagogical task, seeking in this way to progressively innovate the teaching and learning experience. Therefore, the objective of the study was to examine the different didactic proposals used by teachers in the teaching of chemistry and biology; the methodology used was the structured literature review in the process of which 40 studies published during the last six years (2019-2024) were reviewed to undergo qualitative content analysis; the results can be useful for the academic community to know the potential of didactic resources from different learning scenarios in experimental sciences. The findings position the teacher in a constant acquisition and management of technological resources, gamified, experimental, others are related to the everyday or natural environment and others have served as support in the pedagogical process. To date, their implementation has not only enriched teaching, but has also radically transformed the educational experience, opening a range of possibilities for both teachers and students on their way to meaningful learning. In general, it can be concluded that there is still a growing demand for didactic resources to enrich the way science is taught and learned.

Transdisciplinary STEAM learning in improving students’ conceptual understanding of heat and temperature

ABSTRACT Background Students’ difficulty in understanding heat and temperature concepts hinders their ability to solve real-world problems. The positive impact of STEAM education on students’ creativity and problem-solving skills has been reported. However, limited research exists on its influence on students’ conceptual understanding in physics. Purpose This study examines the role of transdisciplinary STEAM learning in enhancing students’ conceptual understanding of heat and temperature, emphasizing its potential to improve the quality of physics education and prepare students for the 21st century. Sample The participants of the study were 76 senior high school students in Indonesia. Design and methods We implemented a mixed-methods sequential explanatory research design to evaluate the impact of transdisciplinary STEAM learning intervention. In this study, pre- and post-tests measured changes in conceptual understanding, and one-to-one interviews further explored students’ learning experiences. From the statistical analysis of the quantitative data, the results suggest that the transdisciplinary STEAM learning enhances students’ conceptual understanding, with a significant improvement observed after completing the lessons. Based on the qualitative thematic analysis, students experience that integrated learning in transdisciplinary STEAM does not only help them to better understand the concepts of heat and temperature, but it also improves their ability to address real-world issues related to these concepts. Conclusion Through the steps of design thinking, this study demonstrated how a real-world problem-solving scenario in transdisciplinary STEAM learning improved students’ conceptual understanding of heat and temperature.

Enhancing cognitive radio networks for education systems: A machine learning approach to optimized spectrum sensing in remote learning environments

The present study focuses on improving Cognitive Radio Networks (CRNs) based on applying machine learning to spectrum sensing in remote learning scenarios. Remote education requires connection dependability and continuity that can be affected by the scarcity of the amount of usable spectrum and suboptimal spectrum usage. The solution for the proposed problem utilizes deep learning approaches, namely CNN and LSTM networks, to enhance the spectrum detection probability (92% detection accuracy) and consequently reduce the number of false alarms (5% false alarm rate) to maximize spectrum utilization efficiency. By developing the cooperative spectrum sensing where many users share their data, the system makes detection more reliable and energy-saving (achieving 92% energy efficiency) which is crucial for sustaining stable connections in educational scenarios. This approach addresses critical challenges in remote education by ensuring scalability across diverse network conditions and maintaining performance on resource-constrained devices like tablets and IoT sensors. Combining CRNs with new technologies like IoT and 5G improves their capabilities and allows these networks to meet the constantly changing loads of distant educational systems. This approach presents another prospect to spectrum management dilemmas in that education delivery needs are met optimally from any STI irrespective of the availability of resources in the locale. The results show that together with machine learning, CRNs can be considered a viable path to improving the networks’ performance in the context of remote learning and advancing the future of education in the digital environment. This work also focuses on how machine learning has enabled the enhancement of CRNs for education and provides robust solutions that can meet the increasing needs of online learning.

When Meta-Learning Meets Online and Continual Learning: A Survey.

Over the past decade, deep neural networks have demonstrated significant success using the training scheme that involves mini-batch stochastic gradient descent on extensive datasets. Expanding upon this accomplishment, there has been a surge in research exploring the application of neural networks in other learning scenarios. One notable framework that has garnered significant attention is meta-learning. Often described as "learning to learn," meta-learning is a data-driven approach to optimize the learning algorithm. Other branches of interest are continual learning and online learning, both of which involve incrementally updating a model with streaming data. While these frameworks were initially developed independently, recent works have started investigating their combinations, proposing novel problem settings and learning algorithms. However, due to the elevated complexity and lack of unified terminology, discerning differences between the learning frameworks can be challenging even for experienced researchers. To facilitate a clear understanding, this paper provides a comprehensive survey that organizes various problem settings using consistent terminology and formal descriptions. By offering an overview of these learning paradigms, our work aims to foster further advancements in this promising area of research.

Exploring dilemma games in sustainable urban planning: A Cypriot case study on urban rooftop utilization for climate change

This paper describes the initial stages of a Cypriot case study on urban rooftop utilization for climate change, indicating how games can: foster collaboration; engage citizens; and produce policy incentives through real-world exploration and implications. The focus of the case study is a co-designed serious game developed within the context of the GREAT project (Games Realizing Effective and Affective Transformation) to be utilized as a facilitation tool between citizens and policy makers. Through learning and role-play scenarios, target groups can: voice their attitudes towards climate policies; gain valuable insights into specific initiatives; and enhance their acceptance of green infrastructure and sustainable urban planning. In this paper, we describe the initial co-design stages of the dilemma game, discuss the lessons learned, and provide future directions for research.

Quantum Few-Shot Image Classification.

Few-shot learning algorithms frequently exhibit suboptimal performance due to the limited availability of labeled data. This article presents a novel quantum few-shot image classification methodology aimed at enhancing the efficacy of few-shot learning algorithms at both the data and parameter levels. Initially, a quantum augmentation image representation technique is introduced, leveraging the local phase of quantum states to support few-shot learning algorithms at the data level. This approach enriches classical data while maintaining its intrinsic physical properties. Subsequently, a parameterized quantum circuit is employed to construct the classification model. This circuit, characterized by a reduced number of trainable parameters, shows increased resilience to overfitting, thereby offering a significant advantage at the parameter level for few-shot learning algorithms. The proposed approach is validated using three datasets, with experimental results indicating that it outperforms classical methods in few-shot learning scenarios while requiring fewer computational resources.

Crisis management in the operating room: A systematic review of simulation training to develop non-technical skills.

Crises in the operating room, often resulting from human factors, endangers patient safety. Simulation-based training to develop non-technical skills shows promise in managing these crises. This review examines the simulation techniques, targeted healthcare professionals, non-technical skills, crisis scenarios, and evaluation metrics used in operating room crisis management training. Systematic review. MEDLINE, APA PsycInfo and Web of Science databases were searched for peer-reviewed articles published between January 2004 and March 2024. This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The review protocol has been registered on the Open Science Framework (OSF) (https://osf.io/7bsc8). The inclusion criteria were as follows: (1) The study population comprised healthcare and medical professionals or students; (2) the intervention involved a simulated learning or training experience; (3) the outcomes focused on non-technical skills or crew resource management; (4) the training setting was the operating room (simulated or real); and (5) the learning scenarios depicted a crisis or an adverse event. This systematic review identified 29 eligible articles. The findings highlight the predominance of high-fidelity simulations, primarily targeting medical staff rather than nurses or other healthcare professionals. Training focused on communication, teamwork, situation awareness, problem solving, and decision making, with scenarios mostly addressing patient deterioration. Assessments reached up to Kirkpatrick's Level 3, demonstrating a positive training impact through learners' reactions and learning metrics rather than behavior and organizational results. Despite their effectiveness, current training practices exhibit limitations. Incorporating nurses and other paramedical staff in interprofessional training, as well as emphasizing team-related scenarios and evaluating behavioral changes in practice, could enhance training effectiveness. This has implications for interprofessional healthcare education and skills transfer to real-world settings, ultimately improving patient safety.

A Software Engineering Approach for Conceptualising an Online Learning Scenario for a Deductive Approach

A Software Engineering Approach for Conceptualising an Online Learning Scenario for a Deductive Approach

Development of integrated science teaching module Pamekasan's local potential for training students' collaboration skills

Collaboration skills are one of the demands of the 21st century that are needed to face the development of the times. In fact, collaboration skills are still not widely applied in schools, including at SDIT Al Irsyad Al Islamiyah. The school has not integrated collaboration skills for several reasons, namely not having a deep understanding of collaboration skills and having constraints in conditioning students in groups. Therefore, an interesting learning scenario is needed that can combine practical activities, local potential, and collaboration skills. The development of a teaching module that integrates Pamekasan's local potential is carried out in order to help students practice their collaboration skills so as to produce a valid and effective teaching module. In this study, the model used was 4D (without dissemination) with interviews, questionnaires, and observations as data collection techniques. The results of the study showed that the teaching module was declared valid with a teaching module validation percentage of 96%, material validation of 83%, language validation of 83%, and validation of learning practitioners of 97%. The teaching module was also declared effective because it was reviewed from the increase in observation results at the third meeting which showed collaboration with a score of 84%, student learning outcomes were categorized as very good, namely 93%, and teacher responses obtained a score of 100% with a very good category.

Delineating the effective use of self-supervised learning in single-cell genomics

Self-supervised learning (SSL) has emerged as a powerful method for extracting meaningful representations from vast, unlabelled datasets, transforming computer vision and natural language processing. In single-cell genomics (SCG), representation learning offers insights into the complex biological data, especially with emerging foundation models. However, identifying scenarios in SCG where SSL outperforms traditional learning methods remains a nuanced challenge. Furthermore, selecting the most effective pretext tasks within the SSL framework for SCG is a critical yet unresolved question. Here we address this gap by adapting and benchmarking SSL methods in SCG, including masked autoencoders with multiple masking strategies and contrastive learning methods. Models trained on over 20 million cells were examined across multiple downstream tasks, including cell-type prediction, gene-expression reconstruction, cross-modality prediction and data integration. Our empirical analyses underscore the nuanced role of SSL, namely, in transfer learning scenarios leveraging auxiliary data or analysing unseen datasets. Masked autoencoders excel over contrastive methods in SCG, diverging from computer vision trends. Moreover, our findings reveal the notable capabilities of SSL in zero-shot settings and its potential in cross-modality prediction and data integration. In summary, we study SSL methods in SCG on fully connected networks and benchmark their utility across key representation learning scenarios.

An Improved Method for Enhancing the Accuracy and Speed of Dynamic Object Detection Based on YOLOv8s.

Accurate detection and tracking of dynamic objects are critical for enabling skill demonstration and effective skill generalization in robotic skill learning and application scenarios. To further improve the detection accuracy and tracking speed of the YOLOv8s model in dynamic object tracking tasks, this paper proposes a method to enhance both detection precision and speed based on YOLOv8s architecture. Specifically, a Focused Linear Attention mechanism is introduced into the YOLOv8s backbone network to enhance dynamic object detection accuracy, while the Ghost module is incorporated into the neck network to improve the model's tracking speed for dynamic objects. By mapping the motion of dynamic objects across frames, the proposed method achieves accurate trajectory tracking. This paper provides a detailed explanation of the improvements made to YOLOv8s for enhancing detection accuracy and speed in dynamic object detection tasks. Comparative experiments on the MS-COCO dataset and the custom dataset demonstrate that the proposed method has a clear advantage in terms of detection accuracy and processing speed. The dynamic object detection experiments further validate the effectiveness of the proposed method for detecting and tracking objects at different speeds. The proposed method offers a valuable reference for the field of dynamic object detection, providing actionable insights for applications such as robotic skill learning, generalization, and artificial intelligence-driven robotics.

Effectiveness of Block Programming and Quarky Robots to Improve Computational Abilities Thinking through the STEMC Approach

The results of PISA 2022 show that Indonesian students are still weak in science literacy, mathematics, and technology-based problem solving, which are the basis for developing computational thinking (CT) skills. CT includes four primary indicators: decomposition to break down significant problems into smaller parts, pattern recognition to find similarities in data, abstraction to filter out irrelevant information, and algorithms to design systematic steps in solving problems. There are many ways to train CT, so this study uses block programming and the Quarky robot. This approach was chosen because it is visual and interactive and makes it easier to apply CT concepts practically, making it suitable for building 21st-century skills. The study was conducted in two high schools in Banda Aceh, involving 20 students from school A and 26 students from school B. Students were divided into two study groups in each school. Learning activities were designed based on STEMC in the form of Student Worksheets (LKPD), which include interactive learning scenarios, block programming challenges, and exploration of the Quarky robot's functions to solve real problems. The activities were arranged in stages, from a basic introduction to applying CT concepts in solving complex problems. The results showed a significant increase in students' CT abilities, especially in the algorithm indicator. Although both schools progressed, school B recorded higher growth, with better pre-test and post-test results than school A. This shows that block programming-based learning and Quarky robots effectively improve CT skills, which is essential in 21st-century education

Towards Automated Chicken Monitoring: Dataset and Machine Learning Methods for Visual, Noninvasive Reidentification

The chicken is the world’s most farmed animal. In this work, we introduce the Chicks4FreeID dataset, the first publicly available dataset focused on the reidentification of individual chickens. We begin by providing a comprehensive overview of the existing animal reidentification datasets. Next, we conduct closed-set reidentification experiments on the introduced dataset, using transformer-based feature extractors in combination with two different classifiers. We evaluate performance across domain transfer, supervised, and one-shot learning scenarios. The results demonstrate that transfer learning is particularly effective with limited data, and training from scratch is not necessarily advantageous even when sufficient data are available. Among the evaluated models, the vision transformer paired with a linear classifier achieves the highest performance, with a mean average precision of 97.0%, a top-1 accuracy of 95.1%, and a top-5 accuracy of 100.0%. Our evaluation suggests that the vision transformer architecture produces higher-quality embedding clusters than the Swin transformer architecture. All data and code are publicly shared under a CC BY 4.0 license.

Private Data Protection with Machine Unlearning in Contrastive Learning Networks

The security of AI models poses significant challenges, as sensitive user information can potentially be inferred from the models, leading to privacy breaches. To address this, machine unlearning methods aim to remove specific data from a trained model, effectively eliminating the training traces of those data. However, most existing approaches focus primarily on supervised learning scenarios, leaving the unlearning of contrastive learning models underexplored. This paper proposes a novel fine-tuning-based unlearning method tailored for contrastive learning models. The approach introduces a third-party dataset to ensure that the model’s outputs for forgotten data align with those of the third-party dataset, thereby removing identifiable training traces. A comprehensive loss function is designed, encompassing three objectives: preserving model accuracy, constraining gradients to make forgotten and third-party data indistinguishable, and reducing model confidence on the third-party dataset. The experimental results demonstrate the effectiveness of the proposed method. Membership inference attacks conducted before and after unlearning show that the forgotten data’s prediction distribution becomes indistinguishable from that of the third-party data, validating the success of the unlearning process. Moreover, the proposed method achieves this with minimal performance degradation, making it suitable for practical applications in privacy-preserving AI.

Leveraging Multiple Adversarial Perturbation Distances for Enhanced Membership Inference Attack in Federated Learning

In recent years, federated learning (FL) has gained significant attention for its ability to protect data privacy during distributed training. However, it also introduces new privacy leakage risks. Membership inference attacks (MIAs), which aim to determine whether a specific sample is part of the training dataset, pose a significant threat to federated learning. Existing research on membership inference attacks in federated learning has primarily focused on leveraging intrinsic model parameters or manipulating the training process. However, the widespread adoption of privacy-preserving frameworks in federated learning has significantly diminished the effectiveness of traditional attack methods. To overcome this limitation, this paper aims to explore an efficient Membership Inference Attack algorithm tailored for encrypted federated learning scenarios, providing new perspectives for optimizing privacy-preserving technologies. Specifically, this paper proposes a novel Membership Inference Attack algorithm based on multiple adversarial perturbation distances (MAPD_MIA) by leveraging the asymmetry in adversarial perturbation distributions near decision boundaries between member and non-member samples. By analyzing these asymmetric perturbation characteristics, the algorithm achieves accurate membership identification. Experimental results demonstrate that the proposed algorithm achieves accuracy rates of 63.0%, 68.7%, and 59.5%, and precision rates of 59.0%, 65.9%, and 55.8% on CIFAR10, CIFAR100, and MNIST datasets, respectively, outperforming three mainstream Membership Inference Attack methods. Furthermore, the algorithm exhibits robust attack performance against two common defense mechanisms, MemGuard and DP-SGD. This study provides new benchmarks and methodologies for evaluating membership privacy leakage risks in federated learning scenarios.