Learning System Research Articles

Causal representation learning through higher-level information extraction

The large gap between the generalization level of state-of-the-art machine learning and human learning systems calls for the development of artificial intelligence (AI) models that are truly inspired by human cognition. In tasks related to image analysis, searching for pixel-level regularities has reached a power of information extraction still far from what humans capture with image-based observations. This leads to poor generalization when even small shifts occur at the level of the observations. We explore a perspective on this problem that is directed to learning the generative process with causality-related foundations, using models capable of combining symbolic manipulation, probabilistic reasoning and pattern recognition abilities. We briefly review and explore connections of research from machine learning, cognitive science and related fields of human behavior to support our perspective for the direction to more robust and human-like artificial learning systems.

Developing Role of Artificial Intelligence in Radiology in the UK

Artificial intelligence (AI) is revolutionising radiological diagnosis in the UK, promising to enhance the accuracy, efficiency, and accessibility of healthcare. The integration of AI into radiology is particularly timely, as the National Health Service (NHS) faces increasing demand for imaging services, coupled with a shortage of radiologists. AI technologies, including deep learning algorithms and machine learning systems, are being developed to assist in interpreting complex medical images such as X-rays, CT scans, and MRIs. One of the key benefits of AI in radiology is its ability to quickly and accurately detect abnormalities. For instance, AI algorithms can identify early signs of diseases like cancer, strokes, and fractures, often with a precision that rivals or exceeds human expertise. This has the potential to significantly reduce diagnostic errors, expedite treatment plans, and improve patient outcomes. For example, AI tools are already in use in the UK to flag lung nodules on CT scans, assisting radiologists in early cancer detection. AI also offers efficiency gains. By automating routine tasks, such as identifying normal scans or prioritizing urgent cases, AI can help streamline workflows, reduce waiting times, and alleviate the burden on overworked radiologists. This is critical, as delays in diagnosis can have serious consequences for patient care. However, the widespread adoption of AI in radiology is not without challenges. Concerns about data privacy, algorithmic transparency, and the potential for over-reliance on AI must be carefully managed. It is crucial to strike a balance where AI complements, rather than replaces, the expertise of radiologists. Ultimately, AI's role in radiological diagnosis in the UK is poised to grow, offering a future where healthcare is not only faster and more accurate but also more equitable for patients across the country.

The impact of the pandemic on psychophysical well-being and quality of learning in the growth of adolescents (aged 11–13): a systematic review of the literature with a PRISMA method

IntroductionThis study examines the social impact and psychological effects of the COVID-19 pandemic on the growth of adolescents students, aged 11–13, on well-being and learning.MethodsSaid review, therefore, will consider all the report of scientific works published since 2020 on the following platforms: SCOPUS, PsycoInfo, and Web of Science. Specifically, the research has selected all studies concerning students aged 11–13. Furthermore, the authors have restricted the scope of the study to two fundamental aspects of the above-mentioned subjects, such as their physical and psychological well-being and didactic learning following the COVID-19 pandemic. The justification for choosing such themes lies in the fact that the SARS-CoV-2 pandemic and the ensuing measures adopted to curb its spread have had -and very likely will continue having- sizable repercussions in the life of each individual, especially in students considered, whose psychological and physical well-being has been considerably affected.ResultsIn particular, during the peak of the pandemic and its subsequent phases, the various measures issued to limit contagion have completely compromised and disregarded the rights of children and youngsters. Suddenly, leisure, sport, play, relations, and socialization disappeared from their lives. Besides, in the early stages of the pandemic, remote teaching replaced face-to-face teaching, later to become an intermittent system in the red- and orange-labelled areas of the country. Consequently, such measures have severely limited opportunities for learning, growth, curiosity and creating relationships that are typical of both a school and outside environment. Concerning the well-being of the subjects considered, a review of the target literature indicates a wide range of psychological and physical disorders, such as malaise, eating and sleep disorders, addiction to games, internet, alcohol, and drugs. For what regards, instead, the realm of learning, most studies highlight a diminished school performance as the main implication of remote teaching. Furthermore, a remarkable number of studies reveals substantial social inequalities; specifically, students coming from middle class families have managed to maintain higher quality educational standards in a critical context such as the pandemic, while those students coming from socially disadvantaged families have had fewer learning opportunities both in terms of time and range of experiences (due to lack of electronic devices, Internet connectivity, homework, continuity of after-school activities). In summary, if the school system pre-COVID-19 already presented significant limitations in its ability to reduce existing social inequalities, school closures during lockdowns pose enormous challenges in developing effective policies to compensate learning losses and inequalities.DiscussionTherefore, based on the results shown, the necessity arises of an emergency plan with social and educational objectives to restore those social, emotional, and cognitive abilities that are compromised. In general, it is the responsibility of countries and governments to guarantee the right to education in these difficult times, while avoiding any type of inequalities and establishing a digital learning system, that allows to face situations such as those generated by the Coronavirus, thus preventing those situations of serious anxiety and stress that have affected most subjects.

Algorithmic Impact Assessments at Scale: Practitioners’ Challenges and Needs

Algorithmic Impact Assessments (AIAs) are often suggested as a tool to help identify and evaluate actual and potential harms of algorithmic systems. While the existing literature on AIAs provides a valuable foundation, critical understanding gaps remain, including the lived experiences of practitioners who implement assessments and a lack of standardization across industry. Such gaps pose significant risks to the usefulness of assessments in the responsible development of algorithmic systems. By conducting 107 assessments with practitioners who build personalization, recommendation, and subscription systems at a large online audio-streaming platform and 8 semi-structured stakeholder interviews, we attempt to bridge this gap by identifying practitioners’ challenges when applying AIAs that might hinder their effectiveness. The paper analyzes whether harms described in the literature related to machine learning and recommendation systems are similar to the concerns practitioners have. We find that the challenges practitioners encounter can be grouped into three categories: technical and methods, infrastructure and operations, and resourcing and prioritization. We also describe ways for teams to more effectively mitigate concerns. This paper helps bridge gaps between the theory and practice of AIAs, advances understanding of the potential harms of algorithmic systems, and informs assessment practices to serve their intended purpose.

Hybrid multiple instance learning network for weakly supervised medical image classification and localization

Weakly supervised medical image analysis is of great significance for computer-aided diagnosis due to the difficulty in obtaining accurately labeled medical data. In this paper, we proposed a new Multi-instance Learning (MIL) framework called HybridMIL integrating CNN Convolutional Neural Networks (CNN) and Broad Learning Systems (BLS). Our HybridMIL can overcome several challenging issues over existing MIL methods based on either CNN or BLS alone: (i) Multiple levels (i.e., different resolutions) of feature information can be simultaneously extracted through a newly proposed instance-level feature enhancement (IFE) module; (ii) Global-level semantic information contained in the deep layers can be better represented under the global-level semantic enhancement (GSE) module; (iii) Hybrid feature fusion (HFF) module is newly designed to effectively fuse and align the multi-level outputs of IFE and global-level semantic information of GSE for subsequent classification and localization tasks. The proposed HybridMIL is evaluated on various public medical and MIL benchmark datasets. The results indicate that HybridMIL surpasses other recent MIL models in terms of classification and localization performance by up to 8.5% and 9.0%, respectively. Lastly, we demonstrate the highly competitive performance of HybridMIL in general MIL problems, going beyond weakly supervised medical image analysis.

Relationship between artificial intelligence and legal education: A bibliometric analysis

This study aims to explore past research trends and identify key future directions at the intersection of artificial intelligence and legal education. The study’s data were gathered from the Scopus database, comprising 68 selected documents spanning from 1999 to 2024. The research methodology involves the use of VOSviewer software for bibliometric analysis. The results reveal that research on artificial intelligence and legal education, while still limited, has been undertaken in various countries, focusing on five primary research directions, including: (1) Improving technical education systems in colleges and universities through educational technology and modern legal learning systems; (2) Application of artificial intelligence and algorithms in the legal field; (3) Applying computational theory and e-learning technology in legal education; (4) Legal education and legal knowledge; (5) Digital transformation in the field of legal training. Based on the research results, five future research directions on this topic are also proposed, including: (1) Application of artificial intelligence in analyzing and predicting legal trends; (2) Artificial intelligence in personalizing the legal learning experience; (3) Artificial intelligence and legal professional ethics; (4) Development of artificial intelligence tools supporting legal teaching and research; and (5) Integration of artificial intelligence into online learning systems for legal education.

Integrating Metal-Phenolic Networks-Mediated Separation and Machine Learning-Aided Surface-Enhanced Raman Spectroscopy for Accurate Nanoplastics Quantification and Classification.

Increasing accumulation of nanoplastics across ecosystems poses a significant threat to both terrestrial and aquatic life. Surface-enhanced Raman scattering (SERS) is an emerging technique used for nanoplastics detection. However, the identification and classification of nanoplastics using SERS faces challenges regarding sensitivity and accuracy as nanoplastics are sparsely dispersed in the environment. Metal-phenolic networks (MPNs) have the potential to rapidly concentrate and separate various types and sizes of nanoplastics. SERS combined with machine learning may improve prediction accuracy. Herein, we report the integration of MPNs-mediated separation with machine learning-aided SERS methods for the accurate classification and high-precision quantification of nanoplastics, which is tailored to include the complete region of characteristic peaks across diverse nanoplastics in contrast to the traditional manual analysis of SERS spectra on a singular characteristic peak. Our customized machine learning system (e.g., outlier detection, classification, quantification) allows for the identification of detectable nanoplastics (accuracy 81.84%), accurate classification (accuracy > 97%), and sensitive quantification of various types of nanoplastics (polystyrene (PS), poly(methyl methacrylate) (PMMA), polyethylene (PE), and poly(lactic acid) (PLA)) down to ultralow concentrations (0.1 ppm) as well as accurate classification (accuracy > 92%) of nanoplastic mixtures at a subppm level. The effectiveness of this approach is substantiated by its ability to discern between different nanoplastic mixtures and detect nanoplastic samples in natural water systems.

Learning in radiation oncology: 12-month experience with a new incident learning system.

Safety and quality improvement are essential to clinical practice in radiation therapy as planning and treatment increase in complexity and sophistication. An incident learning system (ILS) is a safety and quality improvement tool that can aid risk mitigation to improve patient safety and quality of care. The aim of this study was to quantify the impact of implementing a new e-ILS, Learning In Radiation ONcology (LIRON), on reporting and safety culture within a local health district (LHD). The ILS (LIRON) was implemented in 2020 with the intent of tracking actual incidents, near misses and procedural non-compliances for analysis of root causes and contributing factors. A survey was conducted after 12 months of LIRON use, and distributed to radiation oncologists, radiation therapists and radiation oncology medical physicists within the LHD. Results were compared with the responses to a pre-ILS implementation survey, to review changes in staff perceptions of safety culture, barriers to reporting and ILS understanding. Survey response rates were similar at baseline and at the 12-month follow-up, 64% and 63%, respectively. Findings showed increased ILS participation (49-71%), increased perception of no barriers to reporting (34-43%) and increased encouragement to report (37-43%). Greater confidence in the department's ability to learn from the ILS was evident (24-46%). Initial findings of LIRON implementation show positive impact but warrant further long-term review for greater understanding of its impact on staff perceptions, safety culture and improving departmental processes.

An Intelligent Reinforcement Actuating Sensing Learning System Based on the Neuromorphic-Induced Electrodynamics Model

An Intelligent Reinforcement Actuating Sensing Learning System Based on the Neuromorphic-Induced Electrodynamics Model

Efficient Visual-Aware Fashion Recommendation Using Compressed Node Features and Graph-Based Learning

In fashion e-commerce, predicting item compatibility using visual features remains a significant challenge. Current recommendation systems often struggle to incorporate high-dimensional visual data into graph-based learning models effectively. This limitation presents a substantial opportunity to enhance the precision and effectiveness of fashion recommendations. In this paper, we present the Visual-aware Graph Convolutional Network (VAGCN). This novel framework helps improve how visual features can be incorporated into graph-based learning systems for fashion item compatibility predictions. The VAGCN framework employs a deep-stacked autoencoder to convert the input image’s high-dimensional raw CNN visual features into more manageable low-dimensional representations. In addition to improving feature representation, the GCN can also reason more intelligently about predictions, which would not be possible without this compression. The GCN encoder processes nodes in the graph to capture structural and feature correlation. Following the GCN encoder, the refined embeddings are input to a multi-layer perceptron (MLP) to calculate compatibility scores. The approach extends to using neighborhood information only during the testing phase to help with training efficiency and generalizability in practical scenarios, a key characteristic of our model. By leveraging its ability to capture latent visual features and neighborhood-based learning, VAGCN thoroughly investigates item compatibility across various categories. This method significantly improves predictive accuracy, consistently outperforming existing benchmarks. These contributions tackle significant scalability and computational efficiency challenges, showcasing the potential transformation of recommendation systems through enhanced feature representation, paving the way for further innovations in the fashion domain.

Factors influencing global citizenship competencies: A structural equation model of adult learners in Taiwan

Compared to formal education, adult education is one of the most crucial systems for social education and lifelong learning. This paper examines learners' global citizenship competencies and analyzes the influence of demographic variables. Moreover, this article employs structural equation modeling by developing a survey instrument to measure global citizenship competence. The researchers sampled learners from community colleges in Taiwan, distributing 1,400 paper questionnaires, and receiving 1,280 responses. Research findings are: (1) Adult learners' performance in global citizenship competencies is uneven. Participation in multiculturalism, economic phenomena, and political organizations needs to be more advanced. (2) Adult learners' global citizenship competencies were significantly affected by demographic variables such as educational attainment, occupation, monthly personal income, and in an urban or rural area. (3) Results showed that the structural equation modeling is appropriate and has good explanatory power. Implications and suggestions are given at the end.

Byzantine fault tolerance in distributed machine learning: a survey

ABSTRACT Byzantine Fault Tolerance (BFT) is crucial for ensuring the resilience of Distributed Machine Learning (DML) systems during training under adversarial conditions. Among the rising corpus of research on BFT in DML, there is no comprehensive classification of techniques or broad analysis of different approaches. This paper provides an in-depth survey of recent advancements in BFT for DML, with a focus on first-order optimisation methods, particularly, the popular one Stochastic Gradient Descent (SGD) during the training phase. We offer a novel classification of BFT approaches based on characteristics such as the communication process, optimisation method, and topology setting. This classification aims to enhance the understanding of various BFT methods and guide future research in addressing open challenges in the field. This work provides the foundations for developing robust BFT systems, using a variety of optimisation methods to strengthen resilience.

Efficient Neural Hybrid System Learning and Interpretable Transition System Abstraction for Dynamical Systems

Efficient Neural Hybrid System Learning and Interpretable Transition System Abstraction for Dynamical Systems

Virtual Teacher-Aided Learning System Based on Voice Operated Character Animation

Virtual Teacher-Aided Learning System Based on Voice Operated Character Animation

Learning to search promising regions by space partitioning for evolutionary methods

To alleviate the premature, many evolutionary computation algorithms try to balance the exploitation and exploration by controlling the population diversity. However, randomly diversifying a population cannot always guarantee that an algorithm exploits or explores promising regions. To address this issue, a general framework is proposed in this paper for learning promising regions that are made up of subspaces to guide where to exploit and explore by two reinforcement learning systems. The learning mechanism is as follows: (1) To enhance the efficiency of exploitation, an exploitative reinforcement learning system is constructed to estimate the exploitative potential values of subspaces. Accordingly, basins of attraction are approximated by clustering subspaces and historical solutions are selected within the same basin of attraction to generate new solutions. (2) To efficiently explore the solution space, an explorative reinforcement learning system is established to estimate the explorative potential values of subspaces. Accordingly, algorithms are guided to explore subspaces with higher explorative potential values, promoting the discovery of unexploited promising basins of attraction. The framework is implemented into three conventional evolutionary algorithms, and the mechanism and effectiveness of the implemented algorithms are investigated by comprehensive experimental studies. The experimental results show that the proposed algorithms have competitive performances over the other twelve popular evolutionary algorithms.

Human Computer Interaction Strategies for Effective Digital Learning Experiences: From Classroom to Screen

The rapid advancement of technology has transformed traditional classroom settings into complex digital learning environments, emphasizing the need for effective Human-Computer Interaction (HCI) strategies to enhance online learning experiences. This research explores the intersection of HCI and digital learning, focusing on how HCI strategies can improve user engagement, comprehension, and learning outcomes. Through a review of existing literature, this study identifies common challenges faced by learners and educators, such as usability issues, technical difficulties, and inadequate localization. It also highlights effective HCI strategies, including user- centered design, interactive multimedia features, and feedback mechanisms, which are crucial for creating engaging and personalized learning experiences. Findings indicate that personalized learning systems and interactive content significantly enhance performance and achievement by addressing individual learning needs and improving content retention. Best practices derived from user feedback suggest adopting a user-centered design approach, integrating interactive elements, focusing on usability, and continuously updating learning tools. This research underscores the importance of aligning HCI principles with educational technology to foster a more effective and inclusive digital learning environment. Keywords - Human-Computer Interaction (HCI), Digital Learning Environments, User Experience (UX),Interactive Design, Personalized Learning, Educational Technology.

The effects of an augmented reality lens imaging learning system on students’ science achievement, learning motivation, and inquiry skills in physics inquiry activities

The effects of an augmented reality lens imaging learning system on students’ science achievement, learning motivation, and inquiry skills in physics inquiry activities

Artificial intelligence, machine learning and GIS in environmental engineering: current trends

The use of computational tools, such as Artificial Intelligence, Machine Learning or Geographic Information Systems, has had a significant impact on the knowledge generated on environmental issues in recent years. The number of publications shown in this preliminary review of the IEEE Xplore Digital Library database shows its broad applicability and scientific relevance. Search filters and keywords such as water, air, soil, climate change, energy and waste were used. The data obtained was processed to visualize the proportion of use in key topics of environmental interest, giving scientific guidance towards the sites of greatest applicability, as well as towards areas that deserve to be reinforced. In this way, this work aims to promote the development of collaborative work solutions in the related areas of environmental and computational engineering.

Application of Machine Learning in Education: Recent Trends Challenges and Future Perspective

In recent times, Machine learning (ML) is one of the most valuable fields of artificial intelligence (AI) that is transforming education. The application of ML in education provides a promising benefit both to the scientists and researchers and this is the focus of this study. This paper reviews recent trends and advancements of ML in education focusing on areas such as personalisation of learning, predictive analytics, plagiarism detection, intelligent tutoring systems, gamification of learning and recommendation systems. After conducting the literature review we found out the current benefits and challenges of ML in education. The paper also provides insight into the applications and provide the recommendations to address the challenges of ML in the field of education.

Implementation challenges of electronic blood transfusion safety systems: Lessons from an international, multi-site comparative case study.

Severe transfusion reactions resulting from errors in matching the correct blood with the correct patient are considered never events. Despite the relative technical simplicity of barcode scanning for patient-blood bag matching, the adoption and universal application of this safety measure are by no means universal. This study highlights the logistical and institutional challenges associated with spreading, scaling up, and sustaining such IT-supported safety measures in healthcare. We report findings from a 5-year, prospective, multi-site case study conducted across one hospital in England and three hospitals in the Netherlands. Ethnographic methods, including interviews and observations, were used at each site to investigate the implementation of barcode scanning-supported safety pathways for blood transfusions. Significant variation was observed across the sites in the adoption and implementation of barcode scanning-supported safety pathways. Despite the potential for reducing transfusion errors, the introduction of this innovation was met with varying levels of success in different settings. This study highlights the critical role of inter-hospital learning and flexible system design in successfully implementing barcode scanning-supported safety pathways for blood transfusions. A more structured, national-level network for knowledge sharing could enhance the spread and sustainability of such innovations across healthcare settings.