11237 research outputs found
Sort by
Developing Personalised Learning Support for the Business Forecasting Curriculum: The Forecasting Intelligent Tutoring System
No full textIn forecasting research, the focus has largely been on decision support systems for enhancing performance, with fewer studies in learning support systems. As a remedy, Intelligent Tutoring Systems (ITSs) offer an innovative solution in that they provide one-on-one online computer-based learning support affording student modelling, adaptive pedagogical response, and performance tracking. This study provides a detailed description of the design and development of the first Forecasting Intelligent Tutoring System, aptly coined FITS, designed to assist students in developing an understanding of time series forecasting using classical time series decomposition. The system’s impact on learning is assessed through a pilot evaluation study, and its usefulness in understanding how students learn is illustrated through the exploration and statistical analysis of a small sample of student models. Practical reflections on the system’s development are also provided to better understand how such systems can facilitate and improve forecasting performance through training.</p
Vulnerability detection using BERT based LLM model with transparency obligation practice towards trustworthy AI
No full textVulnerabilities in the source code are one of the main causes of potential threats in software-intensive systems. There are a large number of vulnerabilities published each day, and effective vulnerability detection is critical to identifying and mitigating these vulnerabilities. AI has emerged as a promising solution to enhance vulnerability detection, offering the ability to analyse vast amounts of data and identify patterns indicative of potential threats. However, AI-based methods often face several challenges, specifically when dealing with large datasets and understanding the specific context of the problem. Large Language Model (LLM) is now widely considered to tackle more complex tasks and handle large datasets, which also exhibits limitations in terms of explaining the model outcome and existing works focus on providing overview of explainability and transparency. This research introduces a novel transparency obligation practice for vulnerability detection using BERT based LLMs. We address the black-box nature of LLMs by employing XAI techniques, unique combination of SHAP, LIME, heat map. We propose an architecture that combines the BERT model with transparency obligation practices, which ensures the assurance of transparency throughout the entire LLM life cycle. An experiment is performed with a large source code dataset to demonstrate the applicability of the proposed approach. The result shows higher accuracy of 91.8 % for the vulnerability detection and model explainability outcome is highly influenced by “vulnerable”, “function”, "mysql_tmpdir_list", “strmov” tokens using both SHAP and LIME framework. Heatmap of attention weights, highlights the local token interactions that aid in understanding the model's decision points.</p
Temporal gene signature of myofibroblast transformation in Peyronie’s disease: first insights into the molecular mechanisms of irreversibility
No full textTransformation of resident fibroblasts to pro-fibrotic myofibroblasts in the tunica albuginea is a critical step in the pathophysiology of Peyronie’s disease (PD). We have previously shown that myofibroblasts do not revert to fibroblast phenotype and suggested a point of no return at 36 hr after induction of the transformation. However, the molecular mechanisms that drive this proposed irreversibility are not known.Aim: Identify molecular pathways that drive the irreversibility of myofibroblast transformation by analysing the expression of the genes involved in the process in a temporal fashion.Methods: Human primary fibroblasts obtained from tunica albuginea of patients with Peyronie’s disease were transformed to myofibroblasts using TGF-β1. The mRNA of the cells was collected at 0, 24, 36, 48, and 72 hrs after stimulation with TGF-β1 and then analysed using Nanostring nCounter Fibrosis panel. The gene expression results were analysed using Reactome pathway analysis database and ANNi, a deep learning-based inference algorithm based on a swarm approach.Outcomes: A time course of changes in gene expression during transformation of PD-derived fibroblasts to myofibroblasts.Results: The temporal analysis of the gene expression revealed that the majority of the changes at gene expression level happened within the first 24 hours and remained so throughout the 72 hours period. At 36 hours, significant changes were observed in genes involved in MAPK-Hedgehog signalling pathways.Clinical Translation: This study highlights the importance of early intervention in clinical management of PD and future potential of new drugs targeting the point of no return.Strengths & Limitations: The use of human primary cells and confirmation of results with further RNA analysis are the strengths of this study. The study is limited to 760 genes rather than the whole transcriptome.Conclusion: This is the first analysis of temporal gene expression associated with the regulation of the transformation of resident fibroblasts to pro-fibrotic myofibroblasts in PD. Further research is warranted to investigate the role of MAPK-Hedgehog signalling pathways in reversibility of PD.</p
Anomaly-Based Intrusion Detection System for the Internet of Medical Things
No full textThe use of the Internet of Things (IoT) in the health sector, known as the Internet of Medical Things (IoMT), allows for personalized and convenient (e)-health services for patients. However, there are concerns about security and privacy as unethical hackers can compromise these network systems with malware. To address these concerns, we proposed using hyperparameter-optimized Machine and Deep Learning models to build more robust security solutions. We used a representative Anomaly Intrusion Detection System (AIDS) dataset to train six state-of-the-art Machine Learning (ML) and Deep Learning (DL) architectures, with the Synthetic Minority Oversampling Technique (SMOTE) algorithm used to handle class imbalance in the training dataset. Our hyperparameter optimization using the Random search algorithm resulted in accurate classification of normal cases for all six models, with Random Forest (RF) and K-Nearest Neighbors (KNN) performing the best in terms of accuracy. The attention-based hybrid Convolutional Neural Network and Long Short-Term Memory (CNN-LSTM) model was the second-best performer, while the hybrid CNN-LSTM model performed the worst. However, there was no single best model in classifying all attack labels, as each model performed differently in terms of different metrics.</p
brainlife.io: a decentralized and open-source cloud platform to support neuroscience research
No full textNeuroscience is advancing standardization and tool development to support rigor and transparency. Consequently, data pipeline complexity has increased, hindering FAIR (findable, accessible, interoperable and reusable) access. brainlife.io was developed to democratize neuroimaging research. The platform provides data standardization, management, visualization and processing and automatically tracks the provenance history of thousands of data objects. Here, brainlife.io is described and evaluated for validity, reliability, reproducibility, replicability and scientific utility using four data modalities and 3,200 participants.</p
Adaptive multi-view subspace learning based on distributed optimization
No full textAs the rapid development of Internet of Things (IoT), the data is collected from different sensors and stored in distributed devices, these data can be regarded as the multi-view data. There are currently numerous clustering algorithms designed to handle multi-view data. However, most of these algorithms still suffer from the following problems: They are designed to operate directly on raw data, which preserves excessive redundant information and increases the computational burden for subsequent tasks. They primarily focus on pairwise relationships between views, neglecting the intricate high-order connections among multiple views. The prior information of singular values is not taken into account in multi-view. Different views are considered to have equal contributions for clustering. To efficiently address the above problems, adaptive multi-view subspace learning based on distributed optimization (AMSLDO) is proposed in this paper. Specifically, the original multi-view data is projected to a low-dimensional space for subspace representation, and multiple representation matrices are stacked in a tensor with weighted tensor nuclear norm to obtain high-order correlations and discover the prior information of singular values. Furthermore, adaptive graph learning automatically assigns weights to obtain a consensus graph. Meanwhile, the samples are partitioned into the ideal number of clusters through Laplacian rank constraint. An efficient distributed optimization algorithm based on the Alternating Direction Method of Multipliers (ADMM) framework is designed to solve the proposed model. Extensive experiments are conducted on six datasets, demonstrating the superiority of the proposed model compared with eleven state-of-art methods.</p
Designers as co-authors: exploring visual strategies in the creation of hybrid novels
No full textThis practice-based research study investigates the creation of hybrid novels from the designer’s perspective as co-author. The research outcome comprises the 247-page hybrid novel Monster and a 40,000 – word exegetic thesis.The hybrid novel is a literary category defined by Sadokierski (2010) in which word and image can combine to form a reading experience that is neither purely verbal nor purely visual. The limited research on this genre has focused on visual and literary analysis. Less attention has been given to authorial role of designers as visual content providers, the exploration and development of creative tools, and the narrative role of visual elements.The study addresses three major research questionsWhat is the role of the designer as co-author in a hybrid novel?What communication processes does this require between designers and writers?What visual methods are available to designers in developing the narrative and expressive scope of a hybrid novel?These questions were investigated through practice-based research and reflective practice. Co-design strategies were utilised and the resulting process validated these methods for development as frameworks to enable designers, writers, and publishers to communicate and collaborate.This research demonstrates the designer’s authorship through both the visual design content and design management in the development of the narratives. This affirms the importance of visual elements such as white space, paragraph and character styles, and the visual representation of time, for the communication of the story.The research contributes to practical knowledge in identifying the visual strategies by which designers originate content and affirm their co-authorship. This study also provides a framework for communication and collaboration between multiple authors. This framework also complements the research on how creators communicate with each other when using co-design strategies. The research makes innovative contributions to theoretical knowledge in applying concepts from design thinking and co-design to the context of the co-authored hybrid novel.</p
Augmented Reality Book Design for Teaching and Learning Architectural Heritage: Educational Heritage in Hong Kong Central and Western District
No full textThe teaching and learning of architectural heritage can play a vital role in engaging students in envisioning the past and nurturing their cultural identity. However, this endeavour often faces challenges stemming from limited access to heritage sites and varying levels of student interest in the subject matter. This project introduces an augmented reality (AR) book design that seamlessly integrates pop-up three-dimensional models with a dedicated mobile app, effectively illustrating the architectural designs of educational heritages in Hong Kong Central and Western District. An evaluative study was conducted to assess the usability of the AR book for architectural heritage education, which employed a mixed-methods design comprising questionnaire survey with heritage education undergraduate students (N = 80) and semi-structured interviews with a subset of the participants. The results revealed a favourable response to the use of AR technology for virtual representations of cultural heritage, alongside participants’ positive attitudes towards the virtual learning experience facilitated by the AR book. The findings of this study underscore the feasibility and potential benefits of integrating AR technology into architectural heritage education. This integration can offer digitally-mediated learning experiences that actively engage young learners in the exploration and preservation of cultural heritage.</p
Making activities for the competency development of school-age children
No full textContributions: This study examined the effectiveness of making activities in fostering the competency development of school-age children engaged in a making program. The findings suggest that community-based makerspaces can provide autonomous and informal learning experiences, contributing to the development of competences in school-age children. When integrated with formal learning in schools, these experiences can facilitate a well-rounded education that nurtures 21st century skills in the younger generation.
Background: The making program, hosted by community youth centers in Hong Kong, comprised a series of five workshops. These workshops provided guidance throughout the creative processes, encouraging participants to invent artefacts under the theme of ‘smart design for living’.
Research Questions: What generic skills and other attributes can school-age children develop through making activities? What factors influence their development of generic skills and other attributes? What disparities emerged between their community-based and school-based making experiences?
Methodology: The study utilized a mixed-method approach, encompassing of a pre- and post-test questionnaire survey involving school-age children who took part in the making workshops (N = 232), as well as semi-structured interviews with a subset of the participants (N = 25).
Findings: Survey results revealed significant enhancements in participants’ information technology skills, communication skills and divergent thinking, along with a favorable acceptance of the making tools. Pertinent topics related to competency development, including age-related effects, computer accessibility, and mobile device ownership, were examined and discussed within the context of the study.</p
Landscapes of Variability: An Aesthetics of Variation
No full textLandscape, as most spatial formations, can be understood as a multitude of social, ecological and cultural interdependences that hosts the fundamental relations between people and their environment. At the same time, social presence is a necessary requirement for the concept of de-signed landscape to exist, since it is the result of the active participation of people in its configuration and production. While all the above have been true for architectural and landscape design throughout modernity, the COVID 19 health crisis gave them a new meaning and brought them to the forefront. "e pandemic unsettled the balance of our ways of living and threw us within an unknown reality that we entered violently and without warning. In that context, digital technologies seemed to propose the only viable alternative that would allow for the continuation of our social activities. ‘Virtual landscapes’ became part of this array of digital technologies that seemingly came to our aid. In light of Andre’s Leroi-Gourhan approach that every technology is a social construction, that concept could be seen as a refection on the current socio-political status quo: a transition from an ecological-cultural perception of our environment to an aesthetic-digital one. Despite all the above, the post-pandemic experience of this contemporary perception on an ap-plied level, where it concerned the production, configuration and experience of landscapes, provided a rude awakening which diminished the pre-existing euphoria for the identity, the physiognomy and the capabilities it promised, let alone for its implementation. Considering the above, the hypothesis of virtual landscapes as an element of variability might be an alternative. In other words, considering virtual landscapes as an extension or augmentation of the actual ones in ways that could generate multiple possibilities. Landscape therefore becomes augmented and extended virtually through variations that produce multiplicity and variability. "rough that process, landscape can become a cultural host where its aesthetics could transform our social perceptibility. By shaping a multiple context of virtual social presence in the public realm, we can raise awareness and provide guidance by the construction of personal, social and cultural meaning and knowledge of a place, as urban agents of sustainable change.</p
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇