Physical Sciences and Mathematics Commons^™

Long Range Battery-Free Wireless Power Transfer Testbed For Underground Mines Iot And Lpwan Devices, Anabi Hilary Kelechi, Samuel Frimpong, Sanjay Madria

Mining Engineering Faculty Research & Creative Works

Underground mines are susceptible to occasional roof falls and cave-ins, temporarily destroying the existing wireless communications and telemetry infrastructure. During this temporary outage, intermittent provision of electrical energy wirelessly to the already deployed low-power wireless area networks (LPWAN) and Internet of Things (IoT) devices assumes a fundamental requirement. In this article, we propose and design a long-range far-field radio frequency (RF) wireless power transfer (WPT) testbed to power LPWAN and IoT devices at 35 m in an underground mines facility. Class AB external power amplifier (PA) was introduced to achieve a long-distance RF WPT, in the 880 MHz band. Thus, …

A Critical Review Of Pharmaceutical Pollutants In Soil And Air: Ecotoxicological Impacts On Animal, Plant And Microbial Communities - Health Hazards And Waste Management, Md Faisal Amin, Md. Saydur Rahman

School of Integrative Biological & Chemical Sciences Faculty Publications

Pharmaceutical contamination in soil and air has become a critical environmental concern due to its widespread sources, complex behavior, and long-lasting ecological impacts. This review comprehensively explores the presence, fate, and effects of pharmaceutical compounds in soil, highlighting their entry routes, environmental persistence, and biological consequences. Pharmaceuticals infiltrate terrestrial and aquatic environments through various pathways, including agricultural application, pharmaceutical manufacturing waste, sewage sludge, hospital and household discharges, irrigation with contaminated water, and atmospheric deposition. A wide range of drug classes, such as antibiotics, analgesics, non-steroidal anti-inflammatory drugs, antidepressants, anticancer drugs, and hormones, have been identified in both soil and air …

Harnessing Digital Twin (Dt) Technology For Food Security And Climate Resilience In Sub-Saharan Africa (Ssa), Henri E.Z. Tonnang, Francis Chianu, Siyabusa Mkuhlani, Francis Muthoni, John Michael Humphries Choptiany, Franck B.N. Tonle, Bonoukpoe M. Sokame, Mercy Lung’Aho

All Peer-Reviewed Publications

Sub-Saharan Africa (SSA) faces chronic food insecurity despite possessing approximately 60% of the world’s uncultivated arable land. Field trials generate evidence but are costly and insufficiently scaled to address accelerating climate and demographic pressures. Digital twin (DT) technology, defined as the continuous, bidirectional virtual replication of physical systems using real-time data, supports monitoring, modelling, and optimisation of agrifood systems. To our knowledge, however, no published synthesis has examined DT agriculture research through the lens of SSA food systems or smallholder farming realities. A PRISMA-compliant systematic review was conducted across bibliographic databases using a pre-defined Boolean search and adapted PICOS eligibility …

Beyond Legacy Indices: A Critical Review Of The Modified Nemerow, Mean Enrichment Quotient, And Aggregated Multivariate Environmental Risk (Amer) Indices For Heavy Metal Pollution In Sediments, Md Shahnul Islam, Anjuman Ara

Department of Earth and Environmental Studies Faculty Scholarship and Creative Works

Background: Heavy metal pollution in sediments poses persistent ecological risks due to metal toxicity, bioaccumulation, and potential for remobilization. Legacy indices have been widely used for decades but suffer from fundamental flaws, including arbitrary background dependence, lack of grain-size normalization, and poor ecological relevance.

Objective: This review critically evaluates three emerging sediment pollution indices, the Modified Nemerow, Mean Enrichment Quotient (MEQ), and Aggregated Multivariate Environmental Risk (AMER) index, to assess their methodological advances and potential to replace legacy approaches.

Methods: A structured literature search was conducted in Web of Science, Scopus, and Google Scholar (2010–2026). Using an adapted Birch framework, …

The Future Of Nature-Based Recreation In Warming Tropical Cities, Perrine Hamel, Emma E. Ramsay, Shawnda A. Morrison, Su Li Heng, Winston T. L. Chow, Beatrice H. Ho, Pearl Min Sze Tan, Moshe Mandelmilch, Lancy Sim, Jason Kai Wei Lee

Research Collection College of Integrative Studies

Nature-based recreation promotes health in tropical cities but is increasingly threatened by rising heat. This review presents recent evidence on the issue of humid heat stress and outdoor recreation in tropical cities and outlines key adaptation strategies – addressing hazard, exposure, and vulnerability – to enable ‘heat-smart’ nature-based activities. Despite existing solutions, critical research gaps remain, especially in integrating social, physiological, and technological insights to better address humid heat stress in tropical urban environments.

A Systematic Review Of Audio Deepfake Detection Techniques For Digital Investigation, Mahra Alnaqbi, Richard Adeyemi Ikuesan

All Works

Deepfake technology has been driven by advanced machine learning and revolutionized multimedia creation by synthesizing hyper-realistic content. It includes images, videos, and audio. While its creative applications in entertainment and accessibility are significant, the technology also poses critical risks, especially in fraud, disinformation, and identity theft. Audio deepfakes are a subset of this phenomenon that replicate human voices with enhanced precision, mimicking tone, accent, and subtle vocal nuances. This has raised concerns in security-sensitive domains like voice authentication and forensic investigations. This systematic literature review (SLR) adopts PRISMA guidelines to explore the state-of-the-art in audio deepfake detection. It examines existing …

Image And Metadata-Driven Personality Inference For Career Recommendation: A Social Media-Based Ai Framework For Adolescents, Heba Ismail, Maryam Alhefeiti, Ashraf Khalil

All Works

This study presents a novel AI-based framework that leverages Instagram image and metadata analysis to infer Big Five personality traits and deliver personalized career recommendations for high school students in the UAE. Addressing the limitations of traditional recommender systems that rely on self-reported questionnaires or text, the proposed approach uses multimodal visual features—including profile metrics, HSV color patterns, semantic image labels, and texture analysis—to enable a non-intrusive, scalable personalization method. A pilot study involving data from 30 student accounts served as a proof of concept. Correlation analysis identified profile and HSV features as the most predictive, and four machine learning …

Fig-Gan: Fundus Image Generation Via Deep Learning Based Generative Adversarial Network For Amd Disease Diagnosis, Kailasa Thrishul, Ahed Abugabah, Amina Salhi, Manel Ayadi, Mohamed M. Sithik, D. Jayaprakash, A. Ahilan

All Works

Globally, age-related macular degeneration (AMD) remains a main cause of irreversible vision loss. Recently, deep learning models have primarily focused on classifying fundus images for early detection of AMD progression. However, existing models rarely address the generation of future progression-aware fundus images, particularly when complete real longitudinal follow-up scans are unavailable. This limitation makes it difficult to track retinal changes over time and highlights the need for generative models capable of producing realistic drusen-level structural variations. To address these issues, a novel deep learning-based FIG-GAN model is to generate synthetic future fundus images from baseline inputs. Multi-Attention U-Net (MAU-Net) is …

A Performance-Optimized V2v Task Offloading Framework For Real-Time Vehicular Communication, Tariq Qayyum, Asadullah Tariq, Ikbal Taleb, Mohamed Adel Serhani, Zouheir Trabelsi

All Works

As vehicular applications become increasingly complex, their computational demands often exceed the capabilities of individual vehicles. Vehicular Edge Computing (VEC) alleviates this limitation by enabling task delegation to nearby edge resources; however, high mobility, dynamic topology, and fluctuating vehicle density make real-time offloading decisions challenging. To address these issues, we propose a performance-optimized Vehicle-to-Vehicle (V2V) task offloading framework for dense and dynamic Vehicular Ad-hoc Networks (VANETs). The framework follows a two-stage design: (i) context-aware edge-node selection based on live topology capture via periodic beaconing, and (ii) cumulative score-based dynamic priority queuing at the selected edge node. The priority score jointly …

Smart Medical System Integrating Clinical Workflows For Robust Skin Cancer Detection Across Heterogeneous Pathologies, Ahed Abugabah, Prashant Kumar Shukla, Suchi Mishra, Abhishek Dwivedi

All Works

Skin cancer is among the most prevalent and life-threatening dermatological diseases worldwide, with melanoma responsible for a substantial proportion of skin cancer–related deaths due to delayed and unreliable diagnosis. Conventional clinical screening based on visual inspection and expert interpretation is inherently subjective and often affected by inter-observer variability, lesion heterogeneity, and imaging artifacts, highlighting the need for accurate and generalizable automated diagnostic systems. This study proposes a novel hybrid deep learning architecture for skin cancer classification that integrates an attention-guided autoencoder with a transformer-inspired global context modeling module, forming a unified and robust representation learning framework. The encoder–decoder structure is …

Temporally Rigorous And Traceable Predictive Maintenance Via Joint Labeler-Model Optimization, Maytha Al-Ali, Ahmad Alharbi

All Works

Predictive maintenance (PdM) is a critical enabler of intelligent asset management in Industry 4.0, yet many existing frameworks remain difficult to operationalize due to methodological fragmentation. Common limitations include sacrificing temporal realism and class granularity for computational expediency, decoupling labeling strategy design from model hyperparameter optimization, and insufficient support for reproducibility and deployment traceability; particularly in rare-failure regimes. To address these challenges, we propose a unified, end-to-end, and fully traceable PdM framework that jointly optimizes labeling and model parameters while enforcing strict temporal fidelity. The proposed pipeline co-optimizes the failure lookahead window () and LightGBM hyperparameters within a single Bayesian …

Understanding The Out-Of-Distribution Generalization Of Chain-Of-Thought Reasoning In Llms, Nuojing Liang, Xiaotong Yuan, Pan Zhou

Research Collection School Of Computing and Information Systems

Chain-of-Thought (CoT) reasoning substantially improves Large Language Models (LLMs), yet its out-of-distribution (OOD) generalization mechanism remains theoretically underexplored. We study CoT as compositional OOD generalization: the target task lacks complete reasoning trajectories or direct input-output pairs, while training provides only in-distribution (ID) atomic step data. We show that CoT generalizes by repeatedly applying a shared single-step predictor to compose reusable ID transitions. Theoretically, the OOD task risk is controlled by the sum of ID subtask risks, scaled by a depth-independent hard-decoding coefficient Bℓ b = O(k) that is independent of the reasoning depth T. A Rademachercomplexity analysis further gives an …

Drag As Evidence: Motion-Grounded Latent Recomposition For Drag-Based Editing, Xinyu Pu, Hongsong Wang, Jie Gui, Pan Zhou

Research Collection School Of Computing and Information Systems

Modern image editors excel at semantic manipulation and visual synthesis, yet remain limited in precise spatial control, motivating the development of drag-based editing. However, existing drag-based methods often struggle to balance drag accuracy with natural, plausible, and intent-aligned generation. We propose MoReDrag, a motion-grounded drag-based editing method. Our key insight is to treat pixel-space warping as coarse motion evidence, and to inject this evidence into the generative sampling trajectory. Specifically, MoRe-Drag performs region-aware latent recomposition over refinement, inpainting, and anchor regions, coupled with stage-adaptive conditioning that progressively shifts from motion-grounded structure formation to semantic refinement. We further support an instruction-free …

Llm-Enabled Applications Require System-Level Threat Monitoring, Yedi Zhang, Haoyu Wang, Xianglin Yang, Jin Song Dong, Jun Sun

Research Collection School Of Computing and Information Systems

LLM-enabled applications are rapidly reshaping the software ecosystem by using large language models as core reasoning components for complex task execution. This paradigm shift, however, introduces fundamentally new reliability challenges and significantly expands the security attack surface, due to the non-deterministic, learning-driven, and difficult-to-verify nature of LLM behavior. In light of these emerging and unavoidable safety challenges, we argue that such risks should be treated as expected operational conditions rather than exceptional events, necessitating a dedicated incident-response perspective. Consequently, the primary barrier to trustworthy deployment is not further improving model capability but establishing system-level threat monitoring mechanisms that can detect …

A Deep Learning Ensemble Framework For Multi-Subtype Renal Tumor Classification Using Contrast-Enhanced Ct, Hisham Abdeltawab, Ahmed Alksas, Mohamed Ghazal, Ashraf Khalil, Norah Saleh Alghamdi, Rasha T. Abouelkheir, Ahmed Elmahdy, Mohamed Abou El-Ghar, Sohail Contractor, Ayman El–Baz

All Works

Renal cell carcinoma (RCC) is considered the most aggressive and common form of renal cancer. Therefore, early detection is crucial to ensure appropriate and effective treatment planning. In our study, we propose a novel computer-aided diagnostic (CAD) approach which incorporates a deep learning ensemble to differentiate between five renal tumor subtypes, utilising the modality of contrast-enhanced computed tomography (CE-CT). The addressed renal lesions are malignant tumors (chromophobe RCC (chRCC), papillary RCC (pRCC), and clear cell RCC (ccRCC)) and benign tumors (renal oncocytoma (RO) and angiomyolipoma (AML)). Our study includes 280 patients who underwent renal biopsy, 112 patients were diagnosed with …

A Robust Approach For Olive Leaf Disease Detection In Uncontrolled Environments, Rima Grati, Khouloud Boukadi, Emna Ben Abdallah, Ahmed Seffah

All Works

Detecting diseases in olive leaves is crucial for maintaining tree health and ensuring stable olive production. Early signs of infection often appear on the leaves, making them a key indicator for timely disease detection and intervention. Traditionally, farmers rely on visual inspection or laboratory tests to diagnose plant diseases. However, recent advancements in deep learning (DL) have significantly improved the accuracy and efficiency of olive leaf disease diagnosis. Numerous studies in the literature have explored this task using CNN-based architectures and, more recently, Vision Transformers. While these models have shown promising performance on benchmark datasets, they are often trained and …

Ai-Driven Collision Risk Prediction And Debris Optimization In Low Earth Orbit, Emad Alkhazraji, Mohamed E. Yahia, R. E. Tolba, N. S. Alharthi

All Works

Debris congestion and collision in low Earth orbit have become a major point of concern due to the rapid expansion of activities in that region. In this work, an AI-driven framework is developed and deployed to predict, analyze, and optimize collision risk across the altitude inclination space. Local debris object population, mass density distribution, and conjunction derived risk are computed. Then, an AI surrogate model, viz., gradient-boosted decision trees, was trained to predict collision conjunction relative probabilities and risk. Thereafter, data analysis and explain ability techniques are utilized to uncover underlying risk drivers and reveal how altitude, inclination, mass loading, …

From Solitons To Chaos: Nonlinear Dynamics Of The Time-Fractional Modified Regularized Long-Wave Equation, Samsun Nahar Ananna, Chunhui Lu, Abdulla Al Mamun, Md Al Amin Meia, Nur Hasan Mahmud Shahen

School of Mathematical & Statistical Sciences Faculty Publications

This study investigates the nonlinear wave structures and complex dynamical behavior of the time-fractional modified regularized long-wave (mRLW) equation through an integrated analytical, numerical, and dynamical framework, with particular emphasis on connecting localized wave formation with perturbation-induced nonlinear dynamics. Within the adopted fractional framework, a Hirota-type bilinear approach is employed to construct one-wave, two-wave, three-wave, lump, one-lump–one-soliton, and one-lump–one-stripe structures, whose numerical consistency is subsequently examined using the Modified Variational Iteration (MVI) method through graphical comparisons and L 2 and L ∞ error norms. The periodically perturbed mRLW system is further investigated using phase portraits, Lyapunov exponents, bifurcation diagrams, Poincaré …

The Hidden Heatwave: Thermal Hazards In The New York City Subway System, Abdou Rachid Bah, Maimuna Muntaha, Lenny L. Metlitsky, Cesar Pascal Vasquez, Tyler Ayala, Isaac Morel Lopez, Shaquan A. Larose, Kingston Ditsch, Kip Nielsen, Shaunak Sharma, Reginald U. Avila-Ortega Blake, Hamidreza I. Norouzi

Publications and Research

Underground transit systems (Subway) are susceptible to frequent heat waves, which pose serious health and safety risks to commuters and transit workers. Yet, the patterns and extent of temperature changes within the subway system remain insufficiently understood. To address this gap, we conducted a study of thermal variability across 112 subway stations in New York City. Temperature and relative humidity were measured at both platform and street levels. Results show that during the summer, 70% of subway stations had platform heat index higher than street level heat index. We also developed a machine learning forecasting model to predict platform-level temperatures. …

A Deep Learning Based Three-Stage Fusion Neural Network With Composite Adaptive Momentum Optimizer For Bangla Cyberbully Identification, Mahin Montasir Afif, Abdullah Al Noman, Kazi Abdullah Jarif, Md Emamul Arefin Islam, Sathya Narayana Sharma K, Mahfujur Rahman, Dipta Gomes, Kazi Tanvir, Md Obaidur Rahaman, Syed Mohammed Shamsul Islam

Research outputs 2022 to 2026

Cyberbullying in Bangla social media is an increasingly critical issue, posing severe risks to mental health and online safety. Despite the growing need for automated detection systems, challenges such as limited linguistic resources and the nuanced characteristics of the Bangla language hinder model performance. In this study, we propose TriB-FNN, a novel and explainable three-stage fusion neural network, designed to effectively capture cyberbully in Bangla social media texts. The model integrates Convolutional Neural Networks (CNNs), Bidirectional Long Short-Term Memory (Bi-LSTM), and Gated Recurrent Units (GRUs) in a hierarchical pipeline. To optimize training, we employ a Composite Adaptive Momentum (CAM) Optimizer, …

Efficient Removal Of Methyl Orange Dye Via A Mnfe₂O₄/Go Nanocomposite With A Ctab Dual-Layer Surfactant Coating, Taher Karami, Soleiman Bahar, Reza Mostafazadeh

Research outputs 2022 to 2026

This study investigated the adsorption of methyl orange (MO) from aqueous solutions using a novel MnFe₂O₄/GO nanocomposite coated with cetyltrimethylammonium bromide (CTAB). The dual-layer surfactant modification facilitates both electrostatic and lipophilic interactions. This enhancement significantly improves dye removal efficiency. The adsorption process was monitored using spectrophotometry at 464 nm, with various characterization techniques confirming the structural and magnetic properties of the nanocomposite. The optimized parameters for maximum adsorption include a 2-minute ultrasonic dispersion, pH 6.8, and a surfactant-to-adsorbent ratio of 1, achieving a maximum adsorption capacity of 285.7 mg/g. The kinetic data followed a pseudo-second-order model, whereas the adsorption isotherm …

Suicide Ideation Detection Using Social Media Data And Ensemble Machine Learning Model, Erol Kina, Jin Ghoo Choi, Abid Ishaq, Rahman Shafique, Monica Gracia Villar, Eduardo Silva Alvarado, Isabel De La Torre Diez, Imran Ashraf

Research outputs 2022 to 2026

Identifying the emotional state of individuals has useful applications, particularly to reduce the risk of suicide. Users’ thoughts on social media platforms can be used to find cues on the emotional state of individuals. Clinical approaches to suicide ideation detection primarily rely on evaluation by psychologists, medical experts, etc., which is time-consuming and requires medical expertise. Machine learning approaches have shown potential in automating suicide detection. In this regard, this study presents a soft voting ensemble model (SVEM) by leveraging random forest, logistic regression, and stochastic gradient descent classifiers using soft voting. In addition, for the robust training of SVEM, …

Enhanced Carbon Burial In Seagrass Meadows Under Ocean Acidification Revealed By Carbon Dioxide Vents, Theodor Kindeberg, Teixidó, Steeve Comeau, Jean Pierre Gattuso, Beat Gasser, Alice Mirasole, Samir Alliouane, Ioannis Kalaitzakis, Denisa Berbece, Christopher Cornwall, Pere Masque

Research outputs 2022 to 2026

Seagrass meadows are natural carbon sinks, yet the effect of ocean acidification on their carbon burial capacity remains poorly understood. Here we investigated natural carbon dioxide vents in Ischia, Italy to assess how seawater pH influences carbon burial in an area dominated by the seagrass Posidonia oceanica. Organic carbon burial rates (mean ± standard error) between 1954 – 2021 were low under ambient conditions (1.5 ± 0.5 g m-2 yr-1) but increased sharply under acidified conditions (7 ± 1 g m-2 yr-1), reaching sevenfold higher values under extreme acidification (10 ± 3 g m-2 yr-1). Stable isotopes suggest that these …

Bridg-Ics: Ai-Grounded Knowledge Graphs For Intelligent Threat Analytics In Industry 5.0 Cyber-Physical Systems, Padmeswari Nandiya, Ahmad Mohsin, Ahmed Ibrahim, Iqbal H. Sarker, Helge Janicke

Research outputs 2022 to 2026

Industry 5.0’s increasing integration of IT and OT systems is transforming industrial operations but also expanding the cyber–physical attack surface. Industrial Control Systems (ICS) face escalating security challenges as traditional siloed defenses fail to provide coherent, cross-domain threat insights. We present BRIDG-ICS (BRIDge for Industrial Control Systems), an AI-enriched Knowledge Graph (KG) framework for context-aware threat analysis and quantitative assessment of cyber resilience in smart manufacturing environments. BRIDG-ICS fuses heterogeneous industrial and cybersecurity data into an integrated Industrial Security Knowledge Graph linking assets, vulnerabilities, and adversarial behaviors with probabilistic risk metrics (e.g., exploit likelihood, attack cost). This unified graph representation …

Peakon Solutions And Analytical Properties For The Camassa–Holm-Type Equations With Quadratic Nonlinearities, Yonghong Chen, Zhijun Qiao, Mingxuan Zhu

School of Mathematical & Statistical Sciences Faculty Publications

In this paper, we derive the multi-peakon dynamical system of a class of Camassa-Holm-type equations with quadratic nonlinearities. We also consider the analytical properties for the Cauchy problem. Firstly, we establish local well-posedness of solutions in Besov spaces and then provide the blow-up criteria. Subsequently, we impose appropriate sufficient conditions on the initial data to guarantee that the corresponding solution either exists globally or blows up in a finite time. Finally, we prove the ill-posedness in the Besov space B2,∞^3/2 by utilizing the non-traveling wave solutions.

Optimizing Multi-Layer/Multi-Physics Passive Adaptive Thermal Enclosures For Climate-Responsive Energy Regulation: A Numerical Study, Rajae Bousselham, Mingjiang Tao, Sergio Granados-Focil, Adriana Hera, Steven Van Dessel

Chemistry

Adaptive facades regulate heat transfer by adjusting their thermal behavior in response to environmental conditions. Passive adaptive systems, including stimuli-responsive thermal storage materials, tunable radiative coatings, and variable-conductivity layers, offer significant potential for material-based thermal regulation. However, their combined potential remains unexplored, as most studies examine individual mechanisms in isolation. As a result, trade-offs among energy savings, material usage, design complexity, and climate responsiveness are often underrepresented, overlooking coupled interactions in multilayer assemblies limiting their effectiveness. This study develops a transient three-dimensional numerical framework coupling heat transfer with multi-objective optimization to evaluate trade-offs between minimizing annual energy consumption and enclosure …

Defense-To-Attack: Bypassing Weak Defenses Enables Stronger Jailbreaks In Vision-Language Models, Yunhan Zhao, Xiang Zheng, Yige Li, Xingjun Ma

Research Collection School Of Computing and Information Systems

Despite their superb capabilities, Vision-Language Models (VLMs) have been shown to be vulnerable to jailbreak attacks. While recent jailbreaks have achieved notable progress, their effectiveness and efficiency can still be improved. In this work, we reveal an interesting phenomenon: incorporating weak defense cues into the attack pipeline can significantly enhance both the effectiveness and efficiency of jailbreaks on VLMs. Building on this insight, we propose Defense2Attack, a novel jailbreak method that bypasses the safety guardrails of VLMs by leveraging defensive patterns to guide jailbreak prompt construction. Specifically, Defense2Attack consists of three key components: (1) a visual optimizer that embeds universal …

Semantic-Structural Decoupling: Disentangling Semantic Attention From Structural Bias In The Attention Manifold, Pengkun Jiao, Bin Zhu, Jingjing Chen, Yu-Gang Jiang

Research Collection School Of Computing and Information Systems

The empirical success of attention mechanism in Multimodal Large Language Models (MLLMs) often obscures its inherent, subtle flaws. Specifically, MLLMs consistently exhibit disproportionate attention toward certain semantically uninformative visual tokens, a phenomenon termed "register" or "Visual Attention Sinks." While existing inference intervention methods attempt to identify these sink tokens and redistribute their attention weights, such approaches typically treat these tokens in isolation and suffer from computational inefficiency. Instead, we reframe this phenomenon as a generalized textual bias exerted over visual features that extends beyond isolated sink tokens. From this perspective, a pervasive structural bias leads to the dilution of the …

Analyzing Developer Discussions On Eu And Us Privacy Legislation Compliance In Github Repositories, Georgia M. Kapitsaki, Maria Papoutsoglou, Christoph Treude, Ioanna Theophilou

Research Collection School Of Computing and Information Systems

Context: Privacy legislation has impacted the way software systems are developed, prompting practitioners to update their implementations. Specifically, the EU General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA) have forced the community to focus on users’ data privacy. Objectives: Relying on the vast amount of data on developer issues available in GitHub repositories, our aim is to gather empirical evidence on the issues developers of Open Source Software discuss to comply with privacy legislation. Method: We examined such discussions by mining and analyzing 32,820 issues from GitHub repositories. We partially analyzed the dataset automatically to identify …

Spatialimaginer: Towards Adaptive Visual Imagination For Spatial Reasoning, Yian Li, Yang Jiao, Bin Zhu, Tianwen Qian, Shaoxiang Chen, Jingjing Chen, Yu-Gang Jiang

Research Collection School Of Computing and Information Systems

Spatial intelligence, which refers to the ability to reason about geometric and physical structure from visual observations, remains a core challenge for multimodal large language models. Despite promising performance, recent multimodal large language models (MLLMs) often exhibit fragile reasoning traces in spatial intelligence tasks that involve consistent spatial state recognition. We argue that these failures stem from a mismatch between the spatial recognition mechanism and the text-only reasoning behavior of these MLLMs. Effective spatial reasoning requires low-level geometric structure to be faithfully preserved and updated throughout the reasoning process, whereas textual representations tend to abstract away precisely these critical details. …