17876 research outputs found
Sort by
BERT-Based Intrusion Detection System for RF Jamming Attacks in Vehicular Network
Full text linkAs vehicular networks continue to evolve toward increased connectivity and autonomy, they become more vulnerable to cybersecurity threats, particularly Radio Frequency (RF) jamming attacks that can severely disrupt communication systems. This thesis presents a comprehensive study on the application of transformer-based models for RF jamming detection in intelligent transportation systems. Specifically, we evaluate the performance of four pre-trained transformer architectures—BERT, RoBERTa, DistilBERT, and ALBERT—fine-tuned for the multi-class classification of interference, smart jamming, and constant jamming attacks under varying vehicular speeds (25 m/s and 15 m/s). These models are systematically compared against traditional machine learning baselines, including K-Nearest Neighbors (KNN), Random Forest (RaFo), and Long Short-Term Memory (LSTM) networks.
Our results show that transformer models significantly outperform the baselines in terms of classification accuracy and robustness under mobility-induced signal variation. At high speed (25 m/s), BERT achieved the best performance with an accuracy of 96.25%, while at low speed (15 m/s)—a more challenging scenario due to reduced temporal dynamics—DistilBERT led with an accuracy of 91.00%. In contrast, baseline models like KNN and RaFo experienced substantial performance drops, falling to 82.27% and 80.04%, respectively.
In addition to detection performance, we conducted an in-depth analysis of putational efficiency metrics, including training time, inference speed, FLOPs, and GPU memory usage. DistilBERT emerged as the most resource-efficient model, training 59% faster and using 31% less memory than BERT, while maintaining comparable accuracy. RoBERTa demonstrated strong accuracy but incurred the highest computational cost, and ALBERT, though compact in size, showed only moderate performance gains.
Confusion matrix analysis revealed that all transformer models classified interfer-
ence scenarios with near-perfect accuracy. However, distinguishing between smart and constant jamming attacks remained challenging, especially at lower speeds. Importantly, none of the transformer models misclassified attack traffic as benign, indicating strong reliability from a cybersecurity standpoint.
This study demonstrates that transformer-based architectures offer a viable and effective approach to RF jamming detection in vehicular networks, balancing high detection performance with practical deployment efficiency. Future work will explore lightweight fine-tuning techniques and real-world deployment scenarios
Captain Benjamin Watson fonds, 1812-1816, n.d.
Full text linkCaptain Benjamin Watson lived in Newport, Rhode Island, and served with the 25th Infantry during the War of 1812.The fonds consists mostly of correspondence written by Rhode Island Captain Benjamin Watson while he was serving in the military during the War of 1812. Most of the letters are written to his wife Frances. A few are addressed to his brother James. One letter describes in detail the Battle of Stoney Creek, while several others were written from Fort George during the time that the Americans occupied the Fort. Another letter mentions the Battle of Lundy’s Lane. There are also several documents of court proceedings
H.T. (Henry) Fisher fonds, 1949-1976, n.d.
Full text linkH.T. (Henry) Fisher was the director of Long-Range Planning at the Ontario Paper Company Ltd. Fisher was from the United Kingdom but came to Canada with his family around 1948. Initially he worked for the Company in Baie Comeau, and later in Thorold, Ontario.
On February 29, 1912 the Ontario Paper Company Ltd. was incorporated under the leadership of Robert McCormick, editor and publisher of the Chicago Tribune, as a subsidiary of the newspaper. Four months later construction began in Thorold, Ontario, on the banks of the Welland Canal. This location provided an abundance of power, water, and water transportation. The mill was one of the most advanced of its time, using electricity instead of water power. The mill was also the first of its kind as it combined pulp and paper making instead of separating the two operations. A new mill was built in 1980 and the Company was sold in 1996.The fonds contains material created and acquired by Henry Fisher during the time he was employed by the Ontario Paper Company Ltd. Fisher was the director of Long-range planning at the Company. Most of the material consists of reports that involve evaluating different operating scenarios for mills, including the mill in Thorold, Ontario, and Baie Comeau, Quebec. There are also two large photographs of paper mills
From Silence to Support? Examining How Five Ontario Universities Represent their GBV Services for International Students
Full text linkThis Major Research Paper explores how selected Ontario universities address the unique challenges and experiences of international students concerning gender-based violence (GBV). I begin this project with a literature review to critically analyze existing research on international students’ experiences with GBV in universities through an overview of the current knowledge in this field. I then employ a thematic analysis to evaluate the GBV prevention, education and support strategies outlined in the annual reports and websites of five universities in Ontario. My findings indicate that international students encounter significant barriers when seeking help for GBV, including institutional neglect of international students, a lack of education and support specifically designed for these students, and a host of related problems resulting from their experience of precarity related to citizenship, cultural and linguistic dislocation, and immigration-related vulnerabilities. My analysis suggests these systemic forms of exclusion are rooted in Western-centric, white supremacist ideologies, which enable racial and gender-based violence, and silence international students. Based on these findings, the paper proposes actionable recommendations for enhancing service systems, informing policy development, and guiding future research. Overall, this study addresses a critical gap in the literature by examining the disconnect between university support structures and the needs of international students and it advocates for policy and systemic reforms, underscoring the necessity of intersectional research and institutional accountability to address international students’ vulnerability to GBV
Enhancing Robustness of Graph Neural Networks Against Evasive Backdoor Attacks
Full text linkAs Graph Neural Networks (GNNs) are increasingly applied in critical domains such as financial services, healthcare, and entertainment, ensuring their security against adversarial manipulation has become a key concern. Backdoor attacks pose a significant security risk, as they can introduce subtle vulnerabilities during training that may later compromise model predictions in deployment. This thesis investigates methods for defending against such attacks through rigorous investigations, architectural enhancements, and improvements in model resilience. We started by systematically evaluating the most advanced attacks and defensive strategies on GNNs, showing that while combining several defenses greatly improves overall security against state-of-the-art attacks, no single mechanism now in use offers total safety. We then examine advanced GNN architectures, specifically higher-order graph neural networks (HOGNNs), and demonstrate their inherent suitability for mitigating a range of attacks without the need for additional security layers. Based on these findings, we introduce a novel HOGNN-based model that, inspired by state-of-the-art backdoor defenses, incorporates a built-in defense through a cosine similarity-driven subgraph extraction policy that eliminates semantically inconsistent edges. The proposed method significantly outperforms existing approaches in reducing the effectiveness of backdoor attacks, as confirmed by extensive experimental validation. Building on this foundation, we further enhance the model with an additional built-in defense based on clustering-based outlier detection, which filters anomalous nodes and achieves strong resistance to backdoor attacks while maintaining high accuracy across standard benchmarks. Altogether, this thesis advances both the fundamental understanding and the practical development of GNN architectures that are intrinsically secure and well-suited for reliable deployment in sensitive real-world environments
J.H. McCombs Nursery price list, ca. 1920
No full textJesse Henry McCombs (1876-1949) was the son of Jeremiah McCombs and Ellen Bellinger. Both Jeremiah and Jesse were nurserymen and may have been related to other McCombs men in the Fonthill area in the same profession. The Jesse McCombs family resided on North Pelham Street in Fonthill and ran a nursery business on Highway 20, east of Rice Road. At some point in his career Jesse McCombs managed Highland Nurseries. It is estimated that McCombs was in the nursery business from about 1915 to at least 1949. A 1915 advertisement in The Canadian Horticulturalist provides offerings for J.H. McComb’s Union Nurseries. McCombs is a direct descendent of UE Loyalist Timothy McCombs.The record is a price list of offerings by the J. H. McCombs Nursery in Fonthill, Ont
Adolescent Dating Violence: Associations with Community Violence and Dark Triad Traits
Full text linkAdolescent dating violence (ADV) is a serious, prevalent, public health issue, influenced by both environmental and individual-level factors. My study explores the associations between exposure to community violence, Dark Triad personality traits (psychopathy, narcissism, and Machiavellianism) and perpetration and victimization of dating violence during adolescence. Considering the interrelated nature of personality and environmental risk factors, pathways from community violence exposure to dating violence via Dark Triad traits were tested.
The sample consisted of 110 adolescents (51 boys) between the ages of 14 to 18, from a high school in Southern Ontario. Participants completed self-report measures assessing dating violence, community violence exposure and Dark Triad traits. Results revealed that community violence exposure was associated with greater dating violence perpetration and victimization. Psychopathy was positively associated with community violence exposure and dating violence perpetration. There was an indirect effect from community violence exposure to perpetration via psychopathy. Narcissism was positively associated with community violence exposure and dating violence (girls). Machiavellianism was not significantly associated with community violence exposure or dating violence variables. These findings suggest that environmental (i.e., community violence exposure) and individual-level (i.e., Dark Triad personality traits) factors contribute to violence among adolescents. Possible explanations for these associations and implications are discussed
Sport Policy in Newfoundland and Labrador: A Case Study
Full text linkProvincial and territorial governments are an under-researched area of the Canadian sport system, despite their key role in the structure and function of regional sport policy. Canadian sport policy research is well-documented at the national level (Comeau, 2013; Thibault & Harvey, 2013) and to a lesser extent with provincial/territorial sport organizations (Edwards et al., 2009; Rich et al., 2022), yet there is little research that has examined the role of provincial governments (Rich et al., 2024). There have been studies conducted that show the importance of understanding the impact of regional contexts on sport policy (Edwards & Leadbetter, 2016; Lachance & Parent, 2023); however, to my knowledge, little research has explored how these regional contexts might contribute to institutional complexity (see Borgen-Flood & Rich, 2025). The purpose of this study was to examine how the Government of Newfoundland and Labrador addressed sport within their policy between the years 1999 and 2024 and how both institutional and regional identity influenced sport policy in Newfoundland and Labrador. This study was guided by the following research questions: 1) how have elected governments in Newfoundland and Labrador addressed sport in their policy? And 2) how had identity (institutional and regional) influenced sport policy in Newfoundland and Labrador? Data was collected using document analysis and semi-structured interviews with key informants within Newfoundland and Labrador’s regional sport system. Jedlicka and colleagues’ (2022) critical policy analysis framework was used to analyze the document and interview transcripts. Institutional complexity and identity (institutional and regional) were used as the theoretical lens for this study. The findings revealed that sport policy was addressed differently across the different administrations in power over the 25 years of focus for this study. The findings also revealed that identity contributed to complexity through the conflicting political priorities associated with the province’s regional and institutional identities. Overall, this study contributes to an understanding of sport policy in Newfoundland and Labrador while also making theoretical contributions to the literature on institutional complexity
Supervised image classification of Mars 2020 PIXL Optical Fiducial System multispectral data and PIXL-analogous petrological characterization of Mars Sample Return Jezero crater floor analogue rocks
Full text linkNASA’s Mars 2020 (M2020) and NASA/ESA’s proposed Mars Sample Return (MSR) missions seek to characterize the surface of Jezero crater, Mars, in search of ancient biosignatures. The Perseverance rover carries a suite of instruments that includes the Planetary Instrument for X-ray Lithochemistry (PIXL) X-ray fluorescence (XRF) spectrometer, and the Sampling and Caching Subsystem (SCS), which collects rock samples for potential return to Earth in a future MSR mission. The overarching objective of the two studies presented in this thesis is to investigate the igneous petrology of Jezero crater’s Séítah (wehrlitic olivine cumulates) and Máaz (basaltic to trachy-andesitic lavas) formations using data acquired in situ by PIXL and from analogous terrestrial rocks. The first study presents a supervised image classification methodology for estimating the modal mineralogy of PIXL rock targets by extrapolating XRF data to a broader area using imagery acquired by PIXL’s Micro Context Camera (MCC). The objective of this investigation is to assess the efficacy of the methodology when applied to PIXL targets from Séítah and Máaz. The methodology was found to perform accurately when applied to coarse-grained rocks, as their texture provides a larger population of training data. Conversely, fine-grained rocks are classified far less accurately because their lack of training data makes them more susceptible to sources of noise in PIXL and MCC data. The second study characterizes terrestrial rocks collected from Rùm, Scotland and Hart Mountain, Oregon as MSR analogues for Jezero’s igneous Séítah and Máaz formations. The objective of this study was to assess the igneous textures, mineralogy, and geochemistry these rocks, and compare them to their martian counterparts, with the goal of establishing them as compelling analogues. The analogues from Rùm possess a similar cumulate texture and ultramafic mineral assemblage to Séítah samples, while the Hart Mountain samples have a finer-grained, plagioclase, augite, and olivine-bearing that resembles the extrusive rocks of Máaz. While geochemical differences exist between the analogues and their martian counterparts, these are explained by inherent bulk chemistry dissimilarities between the Earth and Mars, implying that the Rùm and Hart Mountain samples are accurate terrestrial representations of Séítah and Máaz
On the Optical Properties of the Misfit Layer Compound (PbSe)1+δ(NbSe2)n
No full textMisfit layer compounds (MLCs) are layered materials that consist of two different sublattices with a mismatch in at least one of their lattice parameters. This incommensurate mismatch along with charge transfer between the constituent sublattices gives rise to interesting properties and can potentially be tuned for practical applications. These systems can be understood using the framework of rigid band model and the changes in band filling depending on the extent of interlayer charge transfer.
This thesis presents the optical properties of (PbSe)1+δ(NbSe2)n and their evolution as the number of NbSe2 layers, n, are varied. NbSe2 is a known transition metal dichalcogenide superconductor and PbSe is a semiconductor with a rocksalt structure. NbSe2 and n=1-3 samples were synthesized via chemical vapour transport and characterized via X-ray diffraction, energy dispersive X-ray spectroscopy, and atomic force microscopy. Further characterization from resistivity and magnetization measurements showed a decrease in the superconducting transition temperature, Tc, with decreasing n. Reflectivity measurements were analyzed using the Drude-Lorentz model and reveal a shift in the plasma frequency to lower energies as n decreases along with the broadening of their interband transitions when compared to pure NbSe2. This behaviour is consistent with increased band filling due to charge transfer from PbSe to NbSe2. Optical conductivity functions were obtained using Kramers-Kronig analysis confirm decreased conductivity for the MLCs and follow a similar pattern found in the resistivity measurements. Finally, frequency-dependent scattering rate and effective mass were obtained from the extended Drude model. Results show a region of purely Drude response in the scattering rate for the MLCs and a trend of decreasing mass enhancement with decreasing n at low wavenumber