14277 research outputs found
Sort by
k-fold cross-validation can significantly over-estimate true classification accuracy in common EEG-based passive BCI experimental designs: an empirical investigation
No full textIn passive brain computer interface (BCI) studies, a common approach is to collect
data from mental states of interest during relatively long trials and divide these trials
into shorter “epochs” to serve as individual samples in classification. While it is
known that using k-fold cross-validation (CV) in this scenario can result in unreliable
estimates of mental state separability (due to autocorrelation in the samples derived
from the same trial), k-fold CV is still commonly used and reported in passive BCI
studies. What is not known is the extent to which k-fold CV misrepresents true mental
state separability. This makes it difficult to interpret the results of studies that use
it. Furthermore, if the seriousness of the problem were clearly known, perhaps more
researchers would be aware that they should avoid it. In this work, a novel experiment
explored how the degree of correlation among samples within a class affects EEG-based
mental state classification accuracy estimated by k-fold CV. Results were compared
to a ground-truth (GT) accuracy and to “block-wise” CV, an alternative to k-fold
which is purported to alleviate the autocorrelation issues. Factors such as the degree
of true class separability and the feature set and classifier used were also explored.
The results show that, under some conditions, k-fold CV inflated the GT classification
accuracy by up to 25%. It is our recommendation that the number of samples derived
from the same trial should be reduced whenever possible in single-subject analysis,
and that both the k-fold and block-wise CV results are reported.Includes bibliographical references (pages 50-59
In the shadow of farmed salmons - fishing small pelagic fish for fish feed in Senegal: a critical analysis of a worlding practice
No full textSalmon aquaculture corporations sell or present their products as a panacea to improve general health and to fight global food insecurity. This discourse hides the necessity to catch small pelagic fish – mostly in developing countries – to produce fishmeal and fish oil required to feed farmed salmons. The fishmeal and fish oil industry is thus involved in a processing of worlding that shadows the existence of other worlds by reducing the ocean life to “marine ingredients.” By doing so, wealthy consumers in the Global North become disconnected from food production processes involved in the food they eat. This dissertation aims at challenging that disconnect by embedding fishmeal and fish oil production geographically and historically inquiring thus the temporal and spatial underpinnings of fishing small pelagic fish for feed production. The shoreline of Senegal and the website of the International Fishmeal and Fish Oil Organization (IFFO) trade organization will be the empirical grounding of this research. This dissertation aims at illuminating spaces of shadows of the fishmeal and fish oil industry and its worlding practices in three moves. First, the French colonial past of the Senegalese marine fisheries participates in contextualizing a continued colonial relation to fisheries resources in Senegal nowadays. Second, a preparatory field trip concomitantly with a documentary analysis structure the use of liminality as a concept to cast light on spaces of shadows where women and men fish workers process fish for a livelihood. Third, the discourse expressed by the IFFO on its website is analyzed as a worlding practice that commodifies the ocean and marine lives, reducing them to ingredients. From this research stands out that by connecting contemporary problems historically and spatially, ethical concerns in food production are revealed. More, a critical stand is made possible against a totalizing and hegemonic discourse of the fishmeal and fish oil industry by pointing out to spaces of shadows that are produced in their worlding practices. Against this process of shadowing, liminality becomes a lens for other worlds to start unfolding.Includes bibliographical references (pages 189-218
Diversity signal sets: implications for organizations and individuals
No full textOrganizations send multiple signals to communicate their support for diversity. Questions of how organizations communicate diversity and how such communication relates to organizational and individual outcomes have been of emerging interest to scholars and practitioners. Research on diversity signaling tends to focus on isolated expression-based signals, (i.e., statements made about diversity); however, less is known about evidence-based signals (i.e., diversity performance facts and statistics). I set out to examine different forms of expression-based signals and evidence-based signals and how such diversity signals are related to corresponding organizational performance outcomes and individual job-seeker outcomes of racial/ethnic minority group members. I argue that when organizations communicate diversity signal sets incorporating both expression-based and evidence-based signals positive organizational and individual outcomes will result. Drawing on Spence's (1971) signaling theory, I conduct three studies following the exploratory sequential mixed methods design. I find that organizations communicate multiple diversity signal sets, and such communication has positive implications for organizations and racioethnic minorities. I contribute to organizational research on diversity signaling by introducing the concept of diversity signal sets and by identifying the affective mechanisms behind the influence of multiple diversity signal sets on racioethnic minorities job-related outcomes.
Study One shows that US Fortune 500 firms signal four types of expression-based signals: equal opportunity (EO), business case (BC), value-and-integration (VI) and mixed, and three types of evidence-based signals: third-party bestowed diversity awards or recognition, employee demographics; and leadership demographics. These findings inform the selection of
variables in the subsequent studies – Studies Two and Three. Study Two examines the association between different types of expression-based signals and their association with evidence-based signals and organizational outcomes. Findings demonstrate that companies with no diversity signals on their corporate websites have negative financial (ROA and ROIC) and diversity-related performance (measured by external recognition/awards of diversity). I find the relationship between companies communicating VI type of expression-based signals and their diversity-related performance significantly positive.
Study Three explores organizational diversity signal sets and their influence on racioethnic minorities' job-related outcomes. I determine that diversity signal sets comprising of both expression-based and evidence-based signals significantly influence racioethnic minorities’ organizational attraction and leadership aspirations through different affective mechanisms (i.e., perceptions of inclusion and perceptions of affective trust). My research contributes to signaling theory by introducing diversity signal sets and by testing their organizational and individual implications. It opens future research avenues for investigation of diversity signal sets in global companies and their organizational and individual implications for multiple diversity groups in the workplace.Includes bibliographical references (pages 187-236
Split-beam sonars: a backscatter model and prototype data analysis
No full textSplit-beam echosounders provide estimation of fish size by directly measuring their
target strength. A model that can be used to evaluate and optimize split-beam sonar
performance is presented. The model has been used to generate beam patterns to
match prototype instruments and to simulate acoustic signals based on the scattering
of sound from particles in a three-dimensional domain. The sum of the contribution
of the scatterers is sampled at the transducer locations, resulting in simulated signals
used for data processing and sonar performance prediction. The model’s capabilities
were evaluated through comparisons with ASL Environmental Sciences’ AZFP-Split
prototype system. The field trials were conducted in Saanichton, BC, by lowering
a calibration target sphere to a range of 220 m in the acoustic beam. Both model
prediction and prototype system performance show accuracy of σ = 0.2 dB at 25 m
range. Potential future applications of the model include exploring methods of target
separation and improving accuracy when presented with complex target structures.Includes bibliographical references (pages 114-117
Integrated geospatial-geochemical catchment analysis of Newfoundland stream waters draining endmember silicate bedrock types (ultramafic-mafic rocks vs. granite)
No full textThe dissolved inorganic element chemistry of stream waters is a reflection of local bedrock composition, but also other catchments properties such as hydrology, surficial geology, vegetation/soil type, and elevation. Studies from small hydrological catchments with a singular or restricted range of bedrock type(s) are ideal for isolating geologic controls on water chemistry from other physicochemical and environmental controls. This study assesses dissolved major, trace, and rare earth element data from streams at two study sites within the same (boreal) climate zone of Newfoundland that each drain a silicate-endmember bedrock composition: (1) Bay of Islands in western Newfoundland draining an ophiolite sequence of mafic/ultramafic bedrock, and (2) Burin Peninsula draining a large Devonian carboniferous granite pluton (St. Lawrence granite) that hosts fluorite mineralization. Through integrating geochemical data with geospatially quantified catchment property data, this study has isolated the elements most strongly controlled by: bedrock geology (Mg, Cr, Ni in ultramafic rock streams; Cs, Rb in granite streams; contrasting REE abundances and Eu anomalies); weathering/soil/vegetation effects and associated flux of dissolved organic matter to streams (K, Co, Ce anomalies); and short temporal events or local geological features such as the fluorite mineralization (elevated element flux, high Y/Ho). Several of these more local water chemical features were traceable into larger catchment areas of Newfoundland major rivers. The low-abundance trace element data are the first to be reported for any Newfoundland surface water and have exposed this research approach as useful to trace boreal catchment dynamics through the lens of trace element geochemistry.Includes bibliographical reference
Design, modeling, and control of precision motion system for the wafer scanner systems
No full textOver the last few decades, academia and industry have focused on designing and developing
high-precision motion systems for various applications. These systems play
a crucial role in modern and future micro- and nanotechnologies, including scanning
probe microscopy and lithography machines in the semiconductor industry. The next
generation of high-precision motion systems, particularly in semiconductor manufacturing,
demands increased throughput (measured in wafers per hour) and improved
accuracy (in micrometers). To meet these requirements, enhancing force stability and
accuracy is essential for achieving faster acceleration in semiconductor lithography machines.
Recent research and assessments indicate that the reluctance actuator holds
promise as a driving mechanism for the next-generation precision motion system in
semiconductor lithography machines. The primary challenge with reluctance actuators
lies in managing the non-linear relationship between magnetic force and current, as well
as between magnetic force and the air gap. These nonlinearities pose significant issues
in design, control, and operation. Furthermore, there is a contemporary trend toward
adopting piezoelectric actuators to drive
exure-guided piezo stages in the short-stroke
(SS) domain. This rising preference highlights the industry's focus on achieving speed
and precision in the evolving landscape of wafer scanner systems. Piezoelectric-based
micopositioning stages are favoured in many applications due to their advantageous
features. However, piezoelectric actuators exhibit challenging nonlinear behavior, complicating
the modeling, control, and synchronization processes.
The dissertation introduces two significant contributions to enhancing the tracking
performance of reluctance-actuated motion stages and piezoelectric micropositioning.
First, it presents a design for a reluctance-actuated motion stage characterized by various
operational conditions. Next, three distinct control approaches are proposed to
linearize the dynamic behavior of the reluctance actuator and improve tracking performance
toward a desired reference signal. Notably, these novel control approaches
rely solely on position measurements, eliminating the need for
ux and force measurements
compared to existing literature-based control methods. The second contribution
involves integrating a uni-axial fine positioning piezo-actuated stage with an existing
precision motion system. This multi-stage design aims to enhance overall system precision.
Additionally, a feedforward compensator-based rate-dependent Prandtl-Ishlinskii
model has been developed to address hysteresis nonlinearities in piezoelectric actuators.
Finally, this dissertation contributes to the design of a tracking control system for a
specific class of non-minimum phase nonlinear systems with unknown uncertainties and
external disturbances. The proposed control strategy combines two key elements: an
output feedback control to Utilize stabilizing full-state feedback control and an extended
high-gain observer to Enhance robustness and disturbance rejection. As a case study,
we demonstrate that the state feedback controller simpli�es to a PID-like controller for
a relative degree-two system.Includes bibliographical references (pages 196-206
Resource allocation techniques for spectral and energy-efficient next generation wireless networks
No full textEfficient utilization of wireless resources is mandated to fulfill the requirements of the
sixth-generation (6G) wireless networks, such as high data rates, low latency, and ubiquitous
connectivity. The word "resource" implies quantities such as bandwidth, power,
and time. Efficiently allocating such limited resources is an effective means to enhance
the wireless systems’ performance. Specifically, resource allocation intends to assign limited
resources to users, maximizing the utilization of these resources, and attaining the
best system performance. In this line, in this dissertation, low-complexity and efficient
resource allocation strategies in networks assisted by various technologies, including nonorthogonal
multiple access (NOMA), reconfigurable intelligent surface (RIS), full-duplex
(FD), cell-free massive multiple-input multiple-output (CFmMIMO), and integrated sensing
and communication (ISAC) are developed and investigated. The first part of the
dissertation focuses on analyzing the outage and throughput performances, as well as
optimizing the sum rate for an FD NOMA-assisted cooperative spectrum-sharing network.
The second part develops novel user clustering and resource allocation algorithms
to boost the sum spectral efficiency of a CFmMIMO-NOMA system. Besides, novel lowcomplexity
resource allocation algorithms for optimizing the energy efficiency and total
transmit power of RIS-aided CF and RIS-enabled federated learning (FL) networks are
proposed. The third part examines the application of RIS and FD in ISAC networks
to improve the transmission rate and sensing performance. Finally, the last part draws
concluding remarks and discusses several topics for future investigation.Includes bibliographical reference
Offshore crew and cargo transfer limits: efficient utilization of safe time
No full textThere have recently been significant petroleum discoveries in the Flemish Pass region of offshore Newfoundland, and their development is an active topic of conversation and research. The Flemish Pass region is further offshore than the currently producing oil installations in the province, and it is in a region characterized by higher winds, higher waves, deeper water, pack ice, and reduced visibility. Typical crew transfer operations in the province are conducted with helicopters, but the Flemish Pass poses some unique challenges in terms of visibility and range for the current helicopter fleet. Marine crew transfer with the current FROG-6 capsule Hs limitation of 4.0 m could make operability drop as low as 10% in the winter months based on public-domain environmental data. As such, improved efficiency in marine crew transfer is sorely needed. This thesis poses the question: can the limits for crane-based transfer methods be improved, if the problem is studied in more detail? This thesis has considered the specific ship and the specific environmental conditions, in terms wind speed and wave height, heading, and peak period. A secondary question is: even if the limits for routine crew transfer remain unchanged, is it possible to safely and practically complete crew transfer operations in seastates above the typical limits for emergency situations? A potential flow analysis is conducted on 6 representative OSV hull forms using ShipMo3D. Time domain analyses are conducted to compare anticipated deck velocities with the operating limits of a commonly-used crew transfer capsule, the FROG-6. It is determined that there is significant variation between the wave height that causes a limit exceedance, depending on the size of the ship, its heading relative to the waves, and its loading condition (GM).
Through analysis of the time between limit exceedances, it is discovered that there is significant “safe time” in between limit exceedances, even in extremely rough sea-states. Filtering out the “safe times” with duration less than 10 minutes, there is still significant “useable time” in all the simulations, i.e. there are significant windows of time with benign vessel motions during which crew transfers can be conducted – even in extreme seas. This thesis concludes that: firstly, that operating limits for crew transfer could be reconfigured to account for differences in ship size/loading condition, and wave period/heading. Secondly, that the “safe time” in between limit exceedances could well be utilized for non-routine crew transfer operations, when supported by the appropriate procedures and motion monitoring/prediction systems.Includes bibliographical references (pages 129-131
The role and regulation of extracellular vesicles released by B cells in response to CD24
No full textExtracellular vesicles (EVs) are membrane-encapsulated nanosized particles that carry
bioactive cargo, including proteins, nucleic acids, and lipids. EVs are secreted by most living
cells, including B lymphocytes (B cells). B cells are antibody-producing immune cells that
develop in the bone marrow, where different cell surface receptors regulate their maturation. One
of the earliest surface proteins expressed in developing B cells is called CD24, a
glycophosphatidylinositol (GPI)-linked protein localized to lipid rafts on the cell plasma
membrane. Past research in the Christian lab showed that engagement of CD24 on the immature
murine WEHI-231 B lymphoma cell line could cause the release of bioactive EVs. Following in
the footsteps of the first discovery, this work employed a model system where donor cells
expressing palmitoylated GFP (WEHI-231-GFP) were co-cultured, after stimulation, with
recipient cells lacking either IgM (WEHI-303 murine B cells) or CD24 (CD24 knock-out mouse
bone marrow B cells). The study found that EVs traffic lipid and membrane proteins between B
cells in response to stimulation of either CD24 or IgM on the donor cells. Importantly, this study
found that EV-mediated transfer of CD24 or BCR may affect B cell development by inducing
apoptosis in recipient bystander cells. The following study aimed to determine how CD24
regulates the release of EVs. Bioinformatic analysis showed that CD24 expression is linked to the
PI3K/AKT and mTOR signaling pathways. Using chemical and genetic inhibition, I found that an
aSMase/PI3K/mTORC2/ROCK/actin pathway regulates EV release. Lastly, through live cell
imaging, the study confirmed that ROCK is required for inducing the membrane dynamics
required for EV release, presumably by regulating tethering of the actin cytoskeleton to the
plasma membrane. The information obtained from this study and other research indicate that EVs
induced by CD24 stimulation are ectosomes that budded off from the plasma membrane rather
than exosomes that originated from multivesicular bodies. Significantly, these data have
uncovered a novel pathway regulating ectosome release that has not been reported in any cell
types. This research topic provides a window into the diverse function of CD24 as well as
increasing our knowledge of how CD24 regulates EV release in cell-to-cell communication.Includes bibliographical reference