21684 research outputs found
Sort by
Exploring factors that influence psychological wellbeing after childhood adversity
Get PDFAbstract available at each chapter
Parental bereavement following perinatal loss: a qualitative exploration of support experiences and the role of milk donation
Get PDFAbstract available at each chapter
Alternative amphiphilic antifouling strategies using fluorous moieties
Get PDFNonspecific adsorption of chemical species onto sensing surfaces, commonly referred to as fouling, poses a significant challenge in biosensing, particularly for label-free detection methods. Fouling can compromise both the accuracy and selectivity of these techniques by interfering with the sensor’s ability to distinguish target analytes from background species.
To address this issue, the development of antifouling surface chemistries has emerged as a promising strategy to reduce or eliminate unwanted interactions. However, many existing antifouling agents fail to provide consistent protection against a broad spectrum of proteins. In this thesis, a series of antifouling molecules incorporating a pentafluorobenzyl moiety were designed to create amphiphilic surfaces as a less harmful alternative to traditional polyfluoroalkyl substances. These molecules were successfully functionalized onto gold nanostructures to form self-assembled monolayers.
Localized surface plasmon resonance, a label-free analytical technique, was employed to evaluate the antifouling performance of the functionalized surfaces. All tested surfaces demonstrated improved resistance to fouling by bovine serum albumin compared to unmodified gold. Additionally, a mixed monolayer incorporating both antifouling molecules and anti-BSA antibodies was fabricated to serve as a biosensing interface. This mixed surface was designed to selectively detect target analytes via antibody recognition while repelling nonspecific protein adsorption through the antifouling component. Although the experimental data were inconclusive, the PFB-based mixed surface exhibited antifouling behaviour comparable to that of a control surface functionalized with polyethylene glycol
Privacy-preserving Federated Learning based on Differential Privacy
Get PDFAlthough FL is claimed to guarantee privacy protection, semi-honest servers and local clients can still reconstruct sensitive information from the gradients. Therefore, to enhance privacy protection, differential privacy (DP) is widely adopted in FL by randomizing the gradients before transmitting them to other parties. Nevertheless, randomizing gradients inevitably degrades FL performance in terms of lower accuracy and higher communication overhead. To solve this problem, this thesis focuses on exploring methods to enhance privacy protection and improve the overall utility of DP-based FL (DPFL) frameworks.
This thesis begins with the research question on improving the degraded accuracy performance and reducing communication overhead for centralized DPFL while maintaining a strong privacy protection guarantee. Two different frameworks are proposed to tackle this question. The first framework combines local DP (LDP) and central DP (CDP) to prevent both central servers and clients from recovering private information by adding noise to the local gradients before uploading and to the aggregated gradients on the server side before broadcasting, respectively. To improve the overall utility of the proposed DPFL, a novel sparse mechanism is adopted on the local gradients before adding noise and a global momentum gradient descent is introduced on the server side and the client side. For the second framework, a novel LDP-based FL framework with two performance improvement modifications is proposed. One modification is to calculate the difference between noisy and original gradients, and add the difference to the objective function to be minimized. The other modification is to calculate the expectation of the loss created by noise, which is also incorporated into the objective function to be optimized. For both modifications, the privacy protection levels are the same as those for plain DPFL since no modifications to DP settings have been made. This thesis presents the necessary convergence analysis for the proposed framework under convex and non-convex settings. A series of simulations is conducted to validate both frameworks’ effectiveness in terms of higher accuracy and lower communication costs. Specifically, the first framework can outperform other DPFL frameworks while saving 90% of communication costs since sparse mechanism can improve the performance under DP noise.
The second framework can save up to 40% of communication and training rounds while achieving better accuracy than plain LDP-based FL. The second research topic in this thesis is to investigate the impact of DP on privacy protection across various DP noise and clipping settings. To address this, an evaluation method for privacy leakage in the FL is proposed by utilizing reconstruction attacks to analyze the difference between the original images and reconstructed ones. Furthermore, this thesis studies the accumulative privacy loss under two different reconstruction attack settings and demonstrates that anonymizing local clients can decrease the probability of privacy leakage. Next, the effects of different clipping methods on privacy protection are analyzed. Simulations are conducted to characterize the trade-off between privacy protection and learning accuracy and demonstrate that there is an optimal DP setting to provide the desired privacy guarantee. The summarized theoretical findings and simulation results in this work can be utilized to guide heterogeneous DP settings for DPFL.
The third research topic of this thesis explores privacy enhancement and accuracy improvement in decentralized DPFL. To address these challenges, a novel anonymous decentralized DPFL framework is proposed. Specifically, two decentralized DPFL methods based on the gossip and fake-centralized manners are first introduced, where the training clients selection rate (TCSR) in each round for both methods and the model exchange rate (MER) in the gossip method are researched. To enhance privacy protection, an anonymous mechanism, is proposed where all clients are unaware of whom they are communicating with and cannot determine whether they are communicating with the same client across several rounds. Next, the required noise scale is derived in terms of the DP settings, TCSR and MER. Subsequently, the convergence bound for the proposed framework is provided, which suggests that an optimal number of clients for is needed to achieve the best convergence performance. Finally, a series of simulations is conducted to evaluate the performance. The simulation results show that the proposed framework only has a small degradation in accuracy compared to the non-private FL and validate our theoretical results.
In conclusion, this thesis provides insight into increasing overall utility and enhancing privacy protection in DPFL. The convergence and privacy analysis of the proposed frameworks provides a basis for future research focusing on further improving the performance of DPFL. Moreover, the proposed privacy leakage evaluation method can provide a more intuitive understanding of privacy loss, which can be utilized to improve accuracy and promote privacy audits for regulatory compliance and user assurance
Millimetre-wave and terahertz signal generation based on monolithic integrated dual-wavelength semiconductor DFB lasers
Get PDFThe high-frequency RF signals generated using the optical beat frequency method exhibit significant design flexibility in terms of frequency. With the advancement of photonic integration technology, these devices achieve a high level of integration, enabling the generation and modulation of signals to be completed chip level devices. Typically, the beat frequency generation process requires two independent single-wavelength lasers, necessitating additional design considerations to ensure the polarization and phase relationship between the two devices. Additionally, at least two power sources are needed to pump the two devices separately, which means that external noise is more likely to affect the phase correlation between the two independent optical signals.
This thesis presents the design, fabrication, and characterisation of a series of monolithically integrated dual-wavelength lasers. The main feature is that two lasing wavelengths are simultaneously emitted within a single resonant cavity, resulting in the two lasing modes naturally having the same polarisation. Lasers and their arrays which achieved beat frequencies range from 50 GHz to 1 THz were fabricated based on three different grating structures: two π phase shifts with equivalent chirped gratings, sidewall grating with lateral modulation of grating coupling coefficient, and superimposed sampled Bragg gratings. For devices operating at terahertz frequencies, the frequency interval can be precisely controlled. With a grating pattern precision of 0.5 nm, the theoretical minimum beat frequency interval can reach 1.1 GHz.
Furthermore, by injecting an external optical frequency comb’s sub-harmonic to the integrated electro-absorption modulator, the phases of the two lasing modes are locked. The experiment recorded two frequencies, 67.75 GHz and 136 GHz, both of which exhibited a linewidth of 1 Hz in the locked state. The phase noise test results indicate that the phase noise of the generated signal adheres to the inherent attenuation law of the frequency comb. This confirms that the injection locking of a dual-wavelength laser can successfully multiply the frequency of an external RF signal without introducing additional noise.
Experimental results confirm that the monolithically integrated dual-wavelength laser possesses a wide design range for beat frequencies. Moreover, after locking, the generated millimetre-wave and THz signals exhibit extremely narrow linewidths and very low phase noise, making the system highly stable and precise for high-frequency applications. This device has significant potential to become a crucial component in microwave/THz photonic integration
Microbial communities and off-grid wastewater treatment
Get PDFWater and sanitation have improved significantly in the past century due to advanced wastewater treatment. Centralised wastewater treatment systems deal with large amount of wastewater in small footprint are always preferred in urban areas. In contrast, decentralised wastewater treatments are more common in rural areas. Among these, septic tanks are one of the most common off-grid solutions, however, their performance various and hardly meets any wastewater discharge standards. This thesis presents a study on the performance and microbial community of small-scale constructed wetland systems treating wastewater from septic tank. Additionally, a lab-based batch experiment to study the change of microbial community and antimicrobial resistance (AMR) during acute amoxicillin exposure.
Three small-scale constructed wetland systems were built and operated in Mexico. Two identical household systems (WS1 and WS2) collected wastewater from single households. They consisted of an equalization tank and a biodigester (AR), followed by a horizontal constructed wetland (CW1), a vertical constructed wetland (CW2) and a holding tank (HT) for the reuse of treated wastewater in gardening. WS3, while similar to WS1 and WS2, received toilet wastewater from a school building and included an additional aeration tank between AR and CW1. All three systems were monitored from the initial three months of the operation, with WS1 and WS2 undergoing an extended monitoring of ten months. Two household systems (WS1 and WS2) were receiving wastewater with high chemical oxygen demand (COD) concentration (average 1131.4mg/L in WS1 and 1894.7mg/L in WS2), NO2 - concentration (average 51.1mg/L in WS1 and 76.7mg/L in WS2) and NO3 - (average 106.4mg/L in WS1 and 214.8mg/L in WS2). Despite higher influent concentrations than full-scale constructed wetland, our constructed wetland system removed at least 73.4% COD, 79.9% NO2 - and 78.8% NO3 - in average. School system (WS3) received wastewater with much lower COD (153.0-587.0mg/L), NO2 - (3.0-54.4mg/L) and NO3 - (13.7-62.9mg/L) concentrations than the household systems, whilst achieved similar removal rates to household systems (58.8-89.9% of COD, 39.3-58.8% of NO2 - and 29.7- 80.6% of NO3 - ).Removal of COD, NO2 - , NO3 - in all three systems were comparable to long-term operational full-scale constructed wetland systems in Mexico from the first month. It took 10 months to achieve a stable removal of COD, NO2 - , NO3 - in both household systems. Phosphate accumulated in the systems in long-term and NH4 + was not removed. Microbial communities were analysed through 16S rRNA sequencing, where AR and HT (influent and effluent of constructed wetlands) were analysed over the monitoring time with additional analysis of within CW1 and CW2 of each system in the last seven months. Microbial community composition in WS3 shared less similarity with household systems, where despite influent was different in two household systems the microbial community compositions in WS1 and WS2 were similar to each other. Bacteroidota was the most dominant phylum in the first three months, while Proteobacteria became the most dominant phylum in the long-term study. Only eight ASVs from the first three months were shared with all 46 most abundant ASVs in long-term study, which suggests the change of microbial community happened overtime. Microbial communities took longer (14 months) to stabilise than pollutant removal, and despite different microbial communities were found in these two systems they achieved comparable pollutant removal rates.
Laboratory preliminary experiments aimed to develop a rapid approach to detect AMR in mixed microbial community from wastewater were undertaken. The growth of anaerobic microbial communities under aerobic condition was monitored by optical density (OD) to mimic transition of wastewater from a septic tank to constructed wetland. A gradient of amoxicillin concentrations (0- 32.0mg/L) was tested, revealing that the growth rate of the microbial community varied across three distinct ranges with increased amoxicillin concentrations in which beneficial range (0-1.5mg/L), detrimental range (1.5- 4.5mg/L) and no further change range (greater than 4.5mg/L). Meanwhile the length of lag phase increased solely with rising amoxicillin concentrations. A further experiment with selected amoxicillin concentrations examined the change of microbial communities through 16S rRNA sequencing and the expression of antimicrobial resistance genes (ARGs) via high throughput quantitative polymerase chain reaction (HT-qPCR), respectively. A shift of the microbial community composition was observed after amoxicillin exposure, where relative abundance of Paenibacillus azoreducens increased while Bacillus cereus group decreased in all amoxicillin exposed microbial communities in both DNA and RNA. However, data were insufficient to conclude that amoxicillin exposure was the only factor for this shift. ARG expression increased with amoxicillin concentration, however ARG abundance was not available due to poor data quality. The results suggest that the length of lag phase may serve as a potential measure for assessing AMR in mixed microbial community, although further work is needed to validate this method.
This study highlights that small-scale constructed wetland systems are capable of removing pollutants from wastewater after septic tank, which could be a potential solution for off-grid sanitation. The potential risk of spreading pathogens and AMR associated with small-scale constructed wetland remain unknown. Further research is needed to understand and assess those risks in the future
Magnetoelectric antennas for wireless power and data transfer at microwave frequencies in medical devices
No full textAbstract not currently available
Youth building the future: The political economy of the adoption of a youth policy in Mexico
Get PDFThis thesis presents a single qualitative case study that examines the main material and ideational factors behind the adoption of the youth policy ‘Jóvenes Construyendo el Futuro‘ (JCF) (Youth Building the Future) in Mexico in 2019. The JCF is an eclectic, broad-based youth policy that combines elements of conditional cash transfer, work-based training and social inclusion that, by 2023, had supported more than 2.3 million young people ‘not in education, employment or training’ (NEET) aged 18 to 29. Historically, most youth policies in Mexico and Latin America have been small-scale ‘upskilling’ interventions. Therefore, most research typically focuses on implementation or labour, social, or economic impacts, overlooking the adoption phase. For these reasons, both because of its scale and its underlying rationale, the JCF appears as a paradigmatic policy.
This thesis employs a combined analytical approach of Cultural Political Economy (CPE) and the Policy Cycle Model (PCM). Methodologically, the study is built on the triangulation of interviews with key informants from the public, private, and social sectors and the analysis of a variety of primary and secondary policy documents. The data was analysed using iterative coding within Thematic Analysis (TA). The findings reveal how several intertwined material and ideational factors explain the adoption of this youth policy. A stern anti-neoliberal stance provided the framework for the leftist MORENA government that won the presidency in 2018 to reinterpret the problems that plagued millions of NEET youth as a direct consequence of the neoliberal development model (1980s - 2018) portrayed as the cause of high levels of poverty, informal employment and lack of opportunities. This problematisation led to the design of an intervention with elements of universality and lax conditionality for the target population, which is uncommon in social policies in Mexico. Despite inevitable disagreements with the private sector in defining the rules of the programme, the overwhelming electoral legitimacy and political-administrative leverage resulted in an expeditious adoption of the JCF.
This thesis contributes various conceptual, empirical, and theoretical insights to current debates on the relationship between development models and the type of welfare institutions that can support the construction of new social bases in contexts of high labour informality, among others, all this amidst significant political-ideological shifts
A sequencing toolbox for Dengue virus: towards accessible genomic surveillance in the Philippines
No full textAbstract not currently available
The landscape of long non-coding RNA in cutaneous squamous cell carcinoma progression
No full textAbstract not currently available
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! 👇