DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
Disease burden and treatment‑seeking behaviour of adults with atopic dermatitis in singapore: an online cross‑sectional survey
No full textIntroduction: Atopic dermatitis (AD) is a chronic systemic inflammatory skin disease with a notably high prevalence in Singapore. Despite available treatments, a significant proportion of patients remain untreated, highlighting a critical need to understand treatment-seeking behaviours and address the multi-faceted disease burden. Methods: An online survey was conducted among 344 adult patients and caregivers answering on behalf of patients to obtain data on clinical impact and quality-of life (QoL), current treatment goals, management, financial impact and treatment-seeking behaviours. This study analysed the differences between patients with different AD severity using data initially collected in Excel and processed in SPSS. Results: AD patients in Singapore face challenges like self-consciousness due to appearance (38%), treatment costs (36%) and the need for additional skincare (34%), with severe AD patients significantly more affected by these issues. Key symptoms like skin dryness (61%), itchiness (56%) and red/scaly skin (48%) worsen with disease severity. AD's impact on patients intensifies with severity, with 100% of severe AD patients rating their condition as 'very serious', correlating with a higher Dermatology Life Quality Index (DLQI) score. Management strategies for AD flare-ups include lifestyle changes (53%) and home remedies (48%). Financial burden is considerable, averaging US dollars (USD) 1368 per month, with 82% perceiving it as 'extreme', affecting treatment adherence. Conclusion: The study underscores the significant burden and impact on QoL faced by adult AD patients in Singapore. It highlights the necessity for targeted research on economic impacts and treatment behaviours in specific groups and the urgent need for effective interventions to enhance QoL, particularly for those with severe AD.Published versionThis study and journal submission fees were funded by Pfzer without direct involvement in or infuence on the research methodology and results
Integrated project allocation and examiner assignment system for final year projects part 5
No full textManaging Final Year Projects (FYP) at scale—across hundreds of students, supervisors, and
examiners—demands more than spreadsheets and siloed systems. As the complexity of project
allocations grows, so does the need for a robust, integrated solution.
The allocation of projects and examiners for the FYP process at the College of Computing and
Data Science (CCDS) has traditionally involved multiple standalone systems to handle
supervision assignment, project matching, and examiner scheduling. This fragmentation often
led to inefficiencies and duplicated administrative effort. To overcome these challenges, our
team developed FYP-APP—a unified platform that consolidates the Supervision Computation
Algorithm (SCA), Student Selection System (SSS), and the FYP Examiner Allocation System
(FEAS).
This report highlights my key contributions to the development and deployment of the FYPAPP,
particularly in finalizing the FEAS component. Additionally, I implemented JWT-based
authentication to safeguard API endpoints and ensure secure user access. These contributions
were critical to preparing the platform for deployment by May 2025, setting a strong foundation
for a more efficient, scalable, and secure FYP management experience for future cohortsBachelor's degre
Distributed observer-based resilient optimization control for desss under false data injection attacks
No full textWhen designing cooperative control strategies for distributed energy storage systems (DESSs), traditional convex optimization algorithms often lead to sub-optimization, and overcurrent issues To address these challenges and achieve power loss minimization across multiple DESSs, this paper proposes an automatic constraint convex optimization algorithm (ACCOA). ACCOA enhances traditional methods by incorporating a penalty function into the evaluation process, ensuring automatic satisfaction of inequality constraints during algorithm iterations. The proposed ACCOA facilitates communication solely between neighboring DESSs, offering scalability and robustness advantages. However, in the optimization process, distributed current allocation becomes susceptible to false data injection (FDI) attacks. To mitigate this vulnerability, a distributed extended observer (DESO) is developed to detect and counteract attack signals, thereby enhancing resilience against FDI attacks in secondary controllers. This research culminates in a resilient-enhanced control framework for cooperative optimization among DESSs, validated through simulations and OPAL-RT experiments to demonstrate the effectiveness of ACCOA-DESO in a multi-DESS system.Submitted/Accepted versio
Roblo-XXX: a frame analysis of sexual content in Roblox
No full textRoblox is an online platform that enables users of all ages to mingle, create, and collaboratively play User Generated Games (UGGs). It has been criticised for the presence of sexual content, resulting in public outcry dominated by sensationalist accounts of the situation that attribute it to child predators. However, these accounts often lack crucial context supplied by player perspectives.
This project builds upon past literature on video games and sexual gameplay, utilising Goffman’s frame analysis to explore how social media users make sense of sexual content on Roblox. Through coding of over 1300 posts, 6 active frames were identified that guide the views of different users around sexual content on Roblox, illuminating how simultaneously active frames can produce unique, contrasting and even contradictory perspectives within shared realities. This study thereby provides insights that may better inform adequate and effective regulatory measures surrounding the issue on the platform.Bachelor's degre
Empirical study on multi-unet diffusion models
No full textWe explore augmenting a pre-trained diffusion model using Low-Rank
Adaptation (LoRA) to rapidly train specialised Unet models for specific
diffusion timesteps. By leveraging the trained model’s existing understanding
of the data distribution, this approach enables efficient specialisation. We
investigate the interactions between these multi-specialised Unet models,
evaluate their performance, and explore various strategies for integrating their
capabilities. The aim of this project is to explore if this multi-model adaptation
approach can provide an efficient method for enhancing model performance
by building upon an already pre-trained diffusion model.Bachelor's degre
Understanding IBS with XAI
No full textIrritable Bowel Syndrome (IBS) is a prevalent gastrointestinal disorder with a
complex etiology, where alterations in the gut microbiome and metabolome are
increasingly recognised as playing a significant role. This study aimed to identify
robust IBS-associated microbial and metabolic signatures and explore their
interactions through a comprehensive multi-omics analysis of the PRJNA812699
cohort, coupled with cross-cohort validation. We analysed 16S rRNA
metagenomics, metatranscriptomics, and metabolomics data individually and in an
integrated multi-omics dataset using Random Forest classifiers for IBS prediction
and feature importance analysis to identify key biomarkers. Explainable AI (XAI)
methods, including feature relationship analysis via Random Forest Regressor and
network graph visualisation, were employed to understand the complex interactions
between these features. Our analysis of the primary cohort revealed that
metabolomics data exhibited strong predictive power for IBS, and multi-omics
integration significantly enhanced classification accuracy. Cross-cohort validation
across three independent IBS cohorts and a larger cross-cohort dataset identified
recurring bacterial taxa, including Prevotella copri and genera like Bacteriodes and
Streptococcus, as potential generalizable IBS markers. Feature relationship analysis
highlighted key interactions, such as the central role of Odoribacter splanchnicus at
the species level and Streptococcus at the genus level in the cross-cohort network.
These findings provide a multifaceted view of the IBS gut microbiome,
underscoring the value of multi-omics integration and cross-cohort validation in
identifying robust biomarkers and understanding the intricate biological networks
underlying IBS pathophysiology. This research contributes to the growing body of
knowledge on IBS, offering potential avenues for future diagnostic and therapeutic
development.Bachelor's degre
Exemplar selection for in-context learning under complex scenarios
No full textLarge Language Models (LLMs) exhibit remarkable few-shot capabilities, where the selec-
tion of exemplars for In-Context Learning (ICL) has been shown to significantly impact
downstream performance. However, prior research largely focuses on either input-side se-
mantic similarity or directly leveraging unmodified sentence embeddings, assuming they
already encode the necessary task information, which may be inadequate for tasks like
semantic parsing requiring deeper language understanding. In contrast, we introduce
a two-fold strategy that enriches LLM representations with deeper linguistic properties
and structural signals tailored for semantic parsing. First, we use tree-edit-distance–based
similarity to form contrastive pairs, finetuning a BERT model that integrates both seman-
tic and structural information of the target parse. Second, we perform an intermediate-
layer intervention by injecting linguistic property information into the hidden states,
subsequently letting higher layers further refine these enhanced representations. This
injection process builds upon recent findings that LLMs often internalize more latent
knowledge than is reflected by their outputs, and that suitable representation-level in-
terventions can better activate such knowledge for specific tasks. Empirical results show
our augmented representations more effectively capture both the semantic and structural
similarity pertinent to semantic parsing, outperforming conventional exemplar selection
baselines. Our approach thus highlights the value of integrating structural signals and
explicit hidden-state interventions in creating more linguistically informed embeddings,
ultimately benefiting in-context learning for semantic parsing tasks requiring rich lan-
guage understanding.Bachelor's degre
Embracing life's dualities: an interpretative phenomenological analysis (IPA) investigation on positive meaning making and positive coping in parents of autistic young adults in Singapore
No full textIn his 2025 Budget Speech, Prime Minister Lawrence Wong highlighted Singapore’s urgent need to support parents to Autistic young adults as “parents worry about what happens post-18, after their children graduate from the structured environment of a Special Education (or SPED) school,” (Ministry of Finance, 2025). This is what is referred to as the “post-18 cliff effect”. This transition, although riddled with challenges and uncertainty, is also a time where parents discover their inner strength. Yet, the current literature does not capture the lasting positive experiences and contributions of parents in journeying with their Autistic young adult. To address this gap, the current qualitative study explores how parents to Autistic young adults construct positive meaning around their journey and the positive coping strategies they engaged in. Semi-structured interviews were conducted with nine parents in Singapore, using Seligman’s (2011) PERMA model as a guiding framework. Data analysis applied the constructivist approach outlined in Interpretative Phenomenological Analysis (IPA), resulting in the conception of the Parental Self-Efficacy and 4Meaning pathways (P-SE4M) model. The P-SE4M model illustrates four meaning-centric themes: (1) Eudaimonic Meaning, (2) Meaning Steadies the Tides, (3) Meaning Through Mindfulness, and (4) Melding Meaningful Bonds. Additionally, the P-SE4M model articulates the reciprocal pathways between parental self-efficacy and the four themes. The study’s findings highlight the value of incorporating culturally grounded meaning-centred strategies into policies supporting this group of parents in Singapore. One potential avenue is the integration of these strategies within the Autism Enabling Masterplan.Bachelor's degre
Development of EEG-BCI based cognition training game
No full textAs cognitive impairment becomes increasingly common in aging populations, the need for greater effective training interventions grows as well. This study explores the impact of real time neurofeedback with adaptive mechanisms in a brain-computer interface (BCI) cognitive training game, specifically designed for the elderly population. By comparing baseline training without real-time neuro-loop, to EEG-based neurofeedback training, we assessed the attention and memory improvements of participants through RMS attention scores, match scores and final performance metrics. Results do indicate positive enhanced attentional control and better memory recall from the adaptive neurofeedback training. The findings suggest that real-time adjustments to in-game mechanisms do optimize engagement and learning efficiency, preventing cognitive overload all while sustaining focus. While promising, this study does acknowledge a need for larger sample sizes, prolonged training sessions and further validation to determine the longer term effectiveness of this approach. Nonetheless, integrating neurofeedback-driven personalization into cognitive training may provide immense help in mitigating cognitive decline and enhancing the quality of life for the elderly.Bachelor's degre
From vision to reality: unpacking the smart nation as a digital order
No full textAs Singapore advances its Smart Nation vision, digital technologies are increasingly embedded in society. Through interviews with experts who have worked on the Smart Nation project and analysis of Smart Nation reports, speeches and information presented in the Smart Nation CityScape exhibition, this study examines how the state perceives digital technology and how this translates into the materialisation of smart urbanism and reinforcement of the digital order in Singapore, responding to the call for more smart urbanism scholarship in non-Western contexts and local studies which have yet to incorporate Smart Nation 2.0 data. Using the concepts of technological determinism by Smith and Marx (1994), technological frames from the social construction of technology and technology as social structure by Sclove (1995), our study finds that the state adopts a soft technological determinist stance which translates into a technology-driven and decentralised smart urbanism approach that reinforces the reconfiguration of Singapore through digital technology.Bachelor's degre