Hong Kong University of Science and Technology
Hong Kong University of Science and Technology Institutional RepositoryNot a member yet
162821 research outputs found
Sort by
Spatio-temporal data fusion framework based on large language model for enhanced prediction of electric vehicle charging demand in smart grid management
No full textAccurate prediction of electric vehicle (EV) charging demand is pivotal for effective smart grid management and renewable energy integration. However, predicting spatio-temporal EV charging patterns remains challenging due to complex data fusion requirements arising from heterogeneous temporal, spatial, and contextual features, as well as difficulties in effectively integrating multiple modeling approaches. This paper introduces EV-STLLM, a novel spatio-temporal data fusion framework based on Large Language Model explicitly designed for accurate short-term EV charging demand forecasting through innovative integration of data-level and model-level fusion techniques. At the data level, a multi-source embedding module is developed to seamlessly fuse temporal features (e.g., time slots, weekdays), spatial heterogeneity (e.g., geographical location), and contextual charging behaviors into a unified representation via embedding convolutional network. At the model level, a large language model (LLM) is employed to capture global spatiotemporal dependencies, enhanced with Low-Rank Adaptation (LoRA) for parameter-efficient fine-tuning, substantially reducing computational costs while maintaining prediction robustness. Using a comprehensive real-world dataset comprising over 830,000 EV charging records across 16 districts and 331 subdistricts in Beijing, we validate EV-STLLM across multiple forecasting scenarios (district and subdistrict levels, one-step and two-step ahead predictions). Extensive comparative evaluations demonstrate that EV-STLLM consistently outperforms classical, graph-based, and deep learning baselines. Specifically, in one-step ahead district-level forecasting, EV-STLLM achieves up to a 15.41% reduction in MAE and a 53.51% reduction in MAPE compared to the leading baseline, underscoring its potential to significantly enhance data-driven smart grid operations.</p
Nurses’ knowledge, attitudes, and role perception in medication administration: do hospital context and nurses’ level of professional experience make a difference?
No full textBackground: Medication administration is a fundamental component of nursing practices, with direct implications for patient safety, health outcomes, and healthcare quality. Despite its importance, however, the literature lacks a comprehensive understanding of how individual and systemic factors are associated in medication administration, thus prompting this study. Methods: A web-based survey was administered to 3,829 nursing staff members across seven hospitals in a Hong Kong cluster. The survey aimed to examine the associations between hospital context, professional ranking, and working experience on nurses’ knowledge, attitudes, and role perception in medication administration. The final count of usable responses was 1,393, yielding a response rate of 36.4%. Results: The study’s findings suggest that nurses’ knowledge, attitudes, and role perception in the medication administration process vary significantly according to the hospital context and the nurses’ work experience and professional rank. Limitations: This study relies on self-reported data and a sample limited to Hong Kong, which may affect generalizability. The lack of formal psychometric validation is also an important methodological consideration. Conclusions: This study highlights the importance of considering individual and systemic factors, including hospital context and nurses’ professional ranking and nursing experience, in strategies aimed at enhancing medication administration and reducing errors. The findings also reveal the need for further research to untangle the complex relationships between these factors and their influence on patient safety and healthcare quality.</p
Assessing the Oxidative Potential of Atmospheric Particulate Matter by Dithiothreitol (DTT) Assay: Mechanistic Insight, Instrument Development and Applications of Real-Time Monitoring
No full textAmbient particulate matter (PM) pollution, particularly fine particulate matter (PM2.5), poses a significant threat to public health, contributing to millions of premature deaths and a substantial burden of disease globally. The complex composition of PM2.5, including transition metals and redox-active organics, catalyzes the formation of reactive oxygen species (ROS) in human lung fluid, leading to oxidative stress and various health impacts. Oxidative potential (OP), defined as the capacity of PM to consume antioxidants and generate ROS, has emerged as a crucial metric for assessing PM toxicity. This thesis presents comprehensive investigations into the OP of PM as assessed by the dithiothreitol (DTT) assay, with a focus on elucidating the underlying mechanisms, developing innovative monitoring techniques, and exploring the applications of real-time monitoring for effective air quality management. This thesis consists of four major parts: (1) Mechanistic Insights: This work explores the nonlinear concentration-response relationships associated with transition metals, particularly copper (Cu) and manganese (Mn), within the DTT assay framework. We elucidate a quasi-Michaelis–Menten mechanism that quantitatively describes the interactions between these metals and OP, contributing to a deeper understanding of how these chemicals influence OP. (2) Innovative Monitoring Instrumentation: We develop a novel online monitoring instrument that employs a particle-into-liquid sampler (PILS) combined with advanced light absorption measurement techniques. This system facilitates continuous, real-time assessments of PM2.5 OP, effectively addressing the limitations of traditional filter-based methods and capturing transient pollution events. (3) Impact of Firework Emissions: This study investigates the acute impact of firework emissions on PM2.5 OP during the Chinese New Year, revealing a dramatic surge in OP values. It highlights the nonlinear response characteristics associated with redox-active metals while examining the role of organic compounds in driving these changes. (4) Long-Term Monitoring: This study examines the OP of PM2.5 in Hong Kong through continuous hourly monitoring over 17 months, providing a comprehensive dataset to enhance understanding of PM2.5 toxicity. Our findings reveal significant seasonal variations in OP and PM2.5 levels, with specific pollution episodes used as case studies. This research highlights the importance of OP as a key metric for assessing PM2.5 toxicity and emphasizes the need to integrate OP measurements into air quality assessments. Overall, this thesis enhances the understanding of PM2.5 OP and its health implications while advancing real-time monitoring techniques that can inform effective air quality management and public health strategies.</p
Inferring Fitness Parameters from Genetic Time-Series Data: Statistical Analysis and Algorithms
No full textFitness describes differences in reproductive success and is a central quantity governing evolutionary change. Genetic time-series data, such as mutant allele frequency trajectories, capture temporal changes in allele frequencies and allow the inference of fitness parameters. Recent technological developments have made it possible to collect increasing amounts of genetic time-series data, thereby motivating the development of many estimators for fitness parameter inference. Despite these efforts, a rigorous statistical understanding of the performance of these estimators and their dependence on system parameters remains limited. In addition, fitness effects may be non-additive, and jointly inferring additive and non-additive components is difficult when only limited measurements are available. This thesis addresses these challenges using the marginal path likelihood framework, which connects stochastic evolutionary models with observed genetic time-series data. Closed-form estimators are derived within this framework to analyze fitness parameter inference and to develop estimation algorithms. The first part of this thesis establishes a statistical framework that quantifies how randomness arising from limited sampling and genetic drift affects estimator performance. By decomposing these sources of randomness, the analysis identifies key quantities that determine estimator precision, clarifies their independent and relative effects, and characterizes how temporal integration mitigates them through distinct scaling behaviors. The second part develops a fully interpretable method for jointly estimating additive and non-additive fitness effects under complex evolutionary models, using only single mutant allele frequency trajectories. Together, these studies strengthen the theoretical understanding of fitness parameter inference from genetic time-series data, providing guidance for practical analysis, evolutionary modeling, and experimental design.</p
Recycling of perovskite solar cells
No full textPerovskite solar cells (PSCs) offer high efficiency and low-cost manufacturing but face challenges of lead management and limited operational lifetimes. This work reviews the material, device, and process characteristics that enable efficient recycling of PSCs. We summarize technoeconomic analysis and life cycle assessments that demonstrate substantial reductions in cost and environmental impacts through multi-round material recovery and compare recycling pathways across device architectures and functional layers. We further discuss practical barriers to scaling laboratory methods to industrial systems, including solvent management and regulatory compliance, and highlight emerging strategies such as design for recycling, automation, and data-driven process optimization. These insights illustrate how closed-loop recycling can support the sustainable deployment of PSC technology and advance a circular photovoltaic economy.</p
Alternate rotating layered circulation in response to extrinsic and intrinsic forcing in the Japan sea
No full textThe three-dimensional circulation in the Japan Sea (JS) plays an important role in its water mass and biogeochemical substances exchange with neighboring oceans. However, characterizing the spatiotemporal circulation pattern in the JS, and diagnosing its complex forcing mechanism between intrinsic flow-topography interaction and extrinsic flux through the straits connected with adjacent seas remain challenge. Combined observations with numerical modeling and a novel Stokes-based layer-integrated vorticity equation (LIVE) dynamics, we discovered a three-layer circulation with alternating cyclonic, anti-cyclonic, and cyclonic circulation in the upper (0–150 m), middle (150–250 m), and bottom (>250 m) layers in the JS, respectively. The strong cyclonic and weak anti-cyclonic circulations in the upper and middle layers show similar seasonal phase: the domain-integrated vorticity anomaly is positive during winter and negative from summer to early autumn. In contrast, cyclonic circulation in the bottom layer remains relatively stable throughout the year. We diagnosed that besides vorticity input from wind stress curl in the upper layer, the lateral planetary vorticity fluxes from inflow/outflow through the straits surrounding the JS lead to vortex stretching in all layers and extrinsically control the structure of the layered circulation. The joint effects of baroclinicity and relief (JEBAR) arising from flow–topography interaction is an intrinsic dynamic response to the extrinsic forcing and dynamically shapes the layered circulation. Based on Stokes circulation theorem, this study characterizes the layered circulation pattern, and based on LIVE dynamics, effectively identifies intrinsic and extrinsic forcing mechanisms for the layered circulation in the JS and other marginal seas.</p
How virtual experience reshapes commuters’ MaaS subscription and mode choice: insights from an economic experiment
No full textAs a noteworthy example of subscription-based service in transportation, Mobility-as-a-Service (MaaS) provides seamless and integrated multimodal travel solutions through bundles, encouraging travelers to transition from private modes to sustainable travel options. While previous studies have primarily focused on the impact of MaaS bundles on mode preferences, the complicated and extensive MaaS-induced behavioral changes and their evolving impact have been overlooked. This study addresses this gap by investigating changes in users’ subscriptions and travel choices with accumulated virtual experience of MaaS bundle usage. Combining stated choice experiments and experimental economics, we conduct a four-part multimodal travel experiment targeting commuters, offering an engaging environment where participants make sequential decisions comprising MaaS bundle subscriptions and travel mode choices. Dynamic discrete choice models are formulated to calibrate participants’ dynamic decision-making processes under MaaS bundle subscriptions and behavioral changes over multiple virtual periods. The results indicate that the virtual experience of subscribing to a particular bundle would motivate them to subscribe to the same bundle again in subsequent periods. When MaaS subscribers make mode choices, their behavior is not simply making trade-offs between travel time and cost. Rather, they tend to consider the future use of their bundles fully, and they are more inclined to make travel decisions based on available bundle discounts. The impact of subscriptions is most pronounced in promoting ride-sourcing trips, followed by multimodal and single-mode public transportation options. These findings offer initial insights into the impact of MaaS subscriptions in reshaping traveler’ subscription and travel choices over a relatively longer period.</p
Phase engineering of nanomaterials: from fundamentals to application frontiers
No full textPhase, which refers to the specific atomic arrangement, is one of the key parameters to determine the physicochemical properties and functions of nanomaterials. Recently, phase engineering of nanomaterials (PEN) has emerged as a promising research direction in materials science, since precise control over atomic arrangements enables the synthesis of nanomaterials with unconventional phases that are different from their thermodynamically stable counterparts, resulting in unique physicochemical properties. Therefore, PEN provides a new strategy for developing novel functional nanomaterials to enhance their performance in various applications. This review focuses on PEN strategies for preparing novel noble metals and transition metal dichalcogenides (TMDs) with unconventional phases. It provides a comprehensive summary of crucial synthetic methods, such as direct synthesis and phase transformation, demonstrates their phase-dependent properties and catalytic performance, and highlights the significant impact of phase on the functions and applications of nanomaterials. Finally, we discuss the challenges and future directions for PEN, including in-depth studies on synthetic mechanisms, effective strategies to improve the stability of unconventional-phase nanomaterials, and innovative AI-aided structural design. These efforts aim to provide theoretical and technical guidance on both fundamental research and practical applications in the field of PEN.</p
Updated clinical practice guidelines for the management of adult diffuse gliomas
No full textIt has been five years since the last version of the clinical practice guidelines for the management of adult diffuse gliomas was published by the Asian Glioma Genome Atlas (AGGA). Significant progress and revisions have occurred in the diagnosis and treatment of adult diffuse gliomas in recent years. In response to these updates, the joint guideline committee of the Chinese Glioma Cooperative Group (CGCG), the Society for Neuro-Oncology of China (SNO-China), and the Chinese Brain Cancer Association (CBCA) has revised the clinical practice guidelines. This updated guideline emphasizes molecular and pathological diagnostics, as well as the primary treatment modalities of surgery, radiotherapy, chemotherapy, and targeted therapy. Additionally, we have incorporated findings from recent clinical trials of new therapies to align with cutting-edge treatment strategies. This guideline is designed to serve as a practical resource for all professionals involved in managing adult diffuse glioma patients, while also providing valuable information for insurance companies and other institutions responsible for regulating cancer care costs in China and beyond.</p
Abrupt temperature shifts alter DOM composition and electron transfer capacity: Unraveling the impact of climate extremes on greenhouse gas emissions in Yangtze River floodplain sediments
No full textFrequently abrupt temperature shifts driven by global climate change significantly disrupt carbon sequestration and CO2 emissions in wetland ecosystems. The condition has raised an important question to clarify ambiguity in the regulatory mechanism of greenhouse gas (GHG) emissions from some of the world's ecologically important large river floodplain wetland systems. This study investigated how abrupt temperature changes could modulate GHG emissions from surface sediments of the China's Yangtze River floodplain wetlands (YRW). A microcosm experiment was conducted in YRW with four temperature treatments: control (20 °C), abrupt cooling (8 °C), abrupt heating (32 °C), and abrupt heating followed by cooling (32 °C to 8 °C) by simulating the latest daily temperature fluctuations (≥12 °C) in the region. Dissolved organic matter (DOM) composition was analyzed in each treatment using the three-dimensional fluorescence spectroscopy followed by the measurement of concentrations of dissolved organic carbon (DOC), total carbon (TC), and total nitrogen (TN) using a TOC analyzer. DOM redox properties were then evaluated through mediated electrochemical techniques. Our results showed that the abrupt heating increased CO2 emissions, while heating followed by cooling suppressed CO2 emissions but significantly increased CH4 emissions. These responses were clearly linked to changes in the carbon concentration and DOM composition in YRW. Specifically, the abrupt heating may have enriched microbial-derived fulvic acid and increased DOC levels, enhancing the electron-accepting capacity (EAC) of DOM by 40–60 %. Fluorescence spectroscopy measurement showed a 12–30 % increase in microbial-derived DOM components under heating. These results demonstrate that temperature-induced alterations in the DOM structure and redox reactivity in large river floodplain wetland sediments strongly regulate GHG fluxes of both CO2 and CH4, providing an important insight into carbon cycling and help developing improved wetland management strategies under climatic warming in YRW.</p