16168 research outputs found
Sort by
Understanding cultural relationships: Whānau, whanaungatanga and Māori student attainment of university entrance in a mainstream secondary school in Aotearoa, New Zealand
Get PDFThis study seeks to understand the importance of cultural relationships in supporting Māori student achievement of University Entrance. This research is based on the stories of five female ākonga Māori, all of whom completed five years of secondary education, and their whānau. It looks deeply into their relational experiences of whanaungatanga and whānautanga with their school, and the impact this had on their academic achievement of NCEA Level 3 and University Entrance.
The results highlight the importance of culturally grounded transformative praxis and the risk of attempting to incorporate culturally located principles such as whānau and whanaungatanga into a schooling context, while still operating within historical hegemonic frameworks
Electromyography based gesture decoding employing few-shot learning, transfer learning, and training from scratch
Get PDFOver the last decade several machine learning (ML) based data-driven approaches have been used for Electromyography (EMG) based control of prosthetic hands. However, the performance of EMG-based frameworks can be affected by: i) the onset of fatigue due to long data collection sessions, ii) musculoskeletal differences between individuals, and iii) sensor position drifting between different sessions with the same user. To evaluate these aspects, in this work, we compare the performance of EMG-based hand gesture decoding models developed using three approaches. This comparison allows for future works in EMG-based Human-Machine Interfaces development to make more informed ML decisions. First, we trained from scratch a Transformer-based architecture, called Temporal Multi-Channel Vision Transformer (TMC-ViT). For our second approach, we utilized a pre-trained and fine-tuned TMC-ViT model (a transfer learning approach). Finally, for our third approach, we developed a Prototypical Network (a few-shot learning approach). The models are trained in a subject-specific and subject-generic manner for eight subjects and validated employing the 10-fold cross-validation procedure. This study shows that training a deep learning decoding model from scratch in a subject-specific manner leads to higher decoding accuracies when a larger dataset is available. For smaller datasets, subject-generic models, or inter-session models, the few-shot learning approach produces more robust results with better performance, and is more suited to applications where long data collection scenarios are not possible, or where multiple users are intended for the interface. Our findings show that the few-shot learning approach can outperform training a model from scratch in different scenarios
Notional
No full textEveryone is unique, especially their face. Even twins have more or less different faces. Faces are the most critical way for people to remember each other in their daily lives, so a face can be regarded as a unique ID of a person.
Notional aims to generate a unique pattern by collecting facial data of a person, such as face length and forehead width, to show that each person is unique. The final pattern can even be used as a form of identity in the context of the metaverse
How to implement online warnings to prevent the use of child sexual abuse material
No full textOnline CSAM offending is a challenge for law enforcement, policymakers and child welfare organisations alike
Predicting responses to a heat acclimation protocol in trained triathletes
Get PDFChapter One reviews the differences between performing in temperate and hot and humid environments. When compared to performance in temperate environments, heat-stressful environments acutely alter physiology heart rate, sweat rate, both core and skin temperature, plasma volume and local blood flow. Heat acclimation (HA) protocol introduced prior to competition in varying ways specific to the competition type has been shown to have large impacts on eventual performance, often summarised through heat response tests (HRT) or adaptations of thermoregulatory physiology. Individual responses to HA have been proven to vary between individuals with little evidence presented to explain why such variation occurs. In Chapter Two, 10 endurance-trained competitive triathletes (aged 33.6 ± 10.6 years, 8 males and 2 females; VO2 max 56.9 ± 11.1) completed a 14-day, cycling-based HA protocol (36 °C, 65% relative humidity). Participants completed two HRTs (Day 1 & Day 13), composed of 20 minutes at a steady state intensity immediately followed by a 30-kilometer time trial. Between these HRTs were seven HA sessions. Performance in the 30-kilometer time trial significantly increased as a result of the HA protocol (p = 0.037). Two regression models for predicting performance outcomes were generated using stepwise regression analysis. A ‘best overall fit’ model using baseline VO2, sweat loss, and sweat sodium composition explains 53% of the variability of performance improvement (R2 = 0.53), and a ‘best practical fit’ model using baseline VO2 and sweat loss explains 49% of the variability of performance improvement (R2 = 0.49). Both models are statistically significant (p = 0.017; p = 0.023). The ‘best practical fit’ model interprets that those with low baseline VO2 but greater sweat loss in the first HRT demonstrated the best performance improvement, whereas those with high baseline VO2 and low sweat loss do not see much performance improvement. Chapter Three investigates the validity of the Kenzen™ wearable core temperature sensor. Ten participants engaged in two HRTs each (20 minutes at a steady state intensity immediately followed by a 30-kilometer time trial) whilst wearing a Kenzen™ device and a rectally inserted thermometer. Bland-Altman plots were used in conjunction with a direct comparison of the differences of the measures to determine validity. The Kenzen™ device is accurate between the range of 37-38 °C, but once core temperature measured rectally reaches 38.5 °C, the validity of the Kenzen™ device comes into question (when rectal temperature ≥ 39 °C, mean difference = 0.79 °C). The difference in measurements suggests that the Kenzen™ is not valid once moderate hyperthermia is reached (38.5 °C). Chapter Four presents a summary of the previous chapters. There is evidence to suggest that predicting variation in HA success is possible, potentially even more so with some refinement of the model through additional thermoregulatory measurement or protocol modification. The validity of wearable core temperature technology does not currently appear suitable in high-performance athletic environments where athletes are expected to reach core temperatures defined as hyperthermic, as most wearable sensors do not seem valid compared to gold standard measures. It is likely that the technology will advance as the popularity of such wearables increases
E pā tō reo e Te Matatini, Herewini Parata
No full textWaita ā-ringa (action song) Lyrics composed by Te Kahautu Maxwell, Musical score composed by Te Whānau Mitai, Eric Kurei. Performed by Ōpōtiki Mai Tawhiti (Kapa Haka group) (3mins)
Waiata ā Tira: Aroha
Whakaeke: Pinepine Te Kura, Hau Te Kura
Mōteatea: Te Ao o Te Pārerā
Waiata ā Ringa: E pā to reo, Te Matatini, Herewini Parata
Poi: E ia e ia
Haka: Te Kōtiritiri, Te Kōtaratara
Whakawātea: Ōranga Tōiti, tōnui kōrer
Extracting information from time series data from rechargable batteries
No full textIn practical applications such as electric vehicles, mobile phones, etc, rechargable batteries are continually charged and discharged, resulting in time-varying voltages and currents. The relationship between voltage and current depends on the properties of the battery and analysis of the time-series can yield useful information about battery performance. However, analysis is complicated because the charge-discharge cycles are not periodic or regular. For example, in an electric vehicle the battery might undergo intensive periods of discharge and charge, corresponding to acceleration and regenerative braking respectively, while the vehicle is in use, followed by long periods of inactivity. There is structure on the timescales of seconds through to days. We have recorded voltage, current and temperature of several batteries in a laboratory as they have been continually cycled over many months using both regular and irregular cycles, the latter mimicking cycles experienced in practical applications. We have used established methods of time-series analysis such as the Hilbert Transform to investigate the relationships between voltage and current, and have shown that results agree with more sophisticated but less practical methods of battery characterisation such as electrical impedance spectroscopy. The relationships imply that several different processes occur, depending on time-scale
The hydrodynamic evolution of the Maketū Estuary after the re-diversion of the Kaituna River
No full textThe Maketū Estuary and lower Kaituna River located in the Bay of Plenty, Aotearoa, New Zealand have undergone anthropic disturbances since late 1800s, including land reclamation, river diversion and re-diversion, and modifications to the river channels and estuary tidal inlet. This has resulted in estuarine degradation, as is a common issue worldwide. Globally different approaches to estuary restoration have aimed at mitigating the adverse impacts of estuarine degradation. However, there are still gaps in restoration frameworks and how the success of the scheme is evaluated that mean there is no ready-made solution to the issues at Maketū.
This thesis examines the important role of the hydrodynamic processes (including water quality) that govern the overall physical response of Maketū Estuary to two stages of partial re-diversion of the Kaituna River into the estuary. The assessment of this physical response highlighted the necessity of a proper framework in which restoration attempts, such as at Maketū are evaluated.
Field time series data were collected before and after two stages of river re-diversion, Stage 1 in 2020 and Stage 2 in 2021, each over a span of a month, to record the immediate response of the estuary to the added freshwater flows. Freshwater flow into the estuary was constrained to approximately 4 hours during flood tides when the water level In the Kaituna River was at least 40 mm higher than the upper Maketū Estuary. The control gates also operate as a flood control scheme, so they remain shut if there is flood risk causing all the flood volume to discharge directly to the sea through Te Tumu Cut river mouth.
The field data were obtained using Aquadopp ADPs, RBR Maestros, RBR Concertos, and Solinst water level loggers at seven different sites throughout the estuary and the Kaituna River. A hydrographic survey of the estuary using RTK-GNSS was undertaken and combined with LiDAR to produce a bathymetric map of the estuary and Lower Kaituna River. Comparison of this map to previous compilations showed that the tidal inlet and lower estuary have not undergone major changes since the original river diversion in 1958, apart from an expansion in the area occupied by the flood tidal delta. However, the upper estuary has undergone both artificial and natural changes, including construction of artificial channels, causeways and reclamation, which have influenced the flow regime, circulation patterns and the flushing ability of the upper estuary.
Based on field observations, the freshwater inflow to the estuary during Stage 1 was higher than Stage 2 even though Stage 1 was a partial re-diversion restricted to 400,000 m3 per flood tide, compared to the maximum 600,000 m3 for Stage 2. There was great spatial variability in the response of the estuary to the added freshwater flow despite the estuary’s small size and lack of inter-tidal vegetation. To better isolate the impact of freshwater flow and eliminate variables such as spring/neap tidal events, rainfall and wind, all of the data were averaged over a semi-diurnal tidal cycle of 12.4 hours. There was an increase in mean water level after both stages. Flow in the mid and upper estuary shifted towards ebb-dominance after Stage 1, but no further shift to ebb-dominance was observed after Stage 2. Mean bottom salinity increased in the upper estuary while it decreased everywhere else after Stage 1. Mean bottom salinity reduced uniformly throughout the estuary after Stage 2, most likely due to higher rainfall rate during the monitoring period, and not the added freshwater flow.
To better understand the mixing patterns within the estuary, Estuarine Richardson number (R_iE) was calculated at 3 main sites before and after both stages of re-diversion. R_iE increased after Stage 1, but not sufficiently to cause a shift in classification; therefore, the estuary remained partially-mixed. Current dynamics within the estuary varied depending on proximity to the tidal inlet, with sites close to the inlet showing a strong tidal signal, and the ones near the control gates showing a decreased tidal influence and an increased response to freshwater inflows during flood tide when the control gates were open. Surface currents were influenced by strong winds during storm events and dictated the mixing (stratification) in the mid and upper shallow estuary. However, it was difficult to assess the significance of these changes in relation to the community aspirations for the re-diversion and estuarine restoration as there were insufficient quantitative or qualitative targets defined.
To complement the field observations and better evaluate the impact of alternative restoration options, a 2D numerical model of the estuary was developed using Delft3D. After successfully being calibrated and validated, Delft3D-FLOW and PART were used to simulate different scenarios to assess the flow regime and flushing ability of the estuary. The scenarios included Stage 1 re-diversion, river inflow flow without control gates; and a pulsed flood flow using control gates that imitated natural flood events. Particle tracking was used determine the residence time across the estuary, and assess within which zones decay rate changed with tidal fluctuations, and which zones were sheltered from the tidal fluctuations. Model results showed that the pulsed flood scenario created the optimal conditions for sediment transport, increased flushing ability, and reduced salinity throughout the estuary. The model results also highlighted differences in how various zones in the estuary responded to the added freshwater flow, further emphasising the need to monitor them separately with an improved monitoring framework.
The results of this study have led to a better understanding of the impact of added freshwater flow on estuarine hydrodynamics and water quality: specifically how the river input moves the boundaries of fluvially or tidally dominated, and mixing zones. The most vulnerable zones in terms of the flushing ability under different scenarios of freshwater restoration were also identified. This study also highlighted the importance of setting tangible targets in the restoration scheme in alignment with the environmental uncertainties such as the average river flow and limitations due to flooding issues in the upper catchment
Using tertiary building performance to define post-disaster functionality timeframes for community recovery and resilience
No full textBuilding damage from earthquakes can have significant consequences for communities, leading to direct and indirect losses. These losses include casualties and repair costs, reduced productivity and well-being. Current building codes prioritize life safety by aiming to prevent deaths, but lack provisions for repairability and functionality, resulting in potentially long recovery times. There has been growing interest in enhancing post-earthquake building functionality, which refers to a building's ability to serve its intended purpose. Beyond life safety, additional functionality states have been defined, including re-occupancy (safe shelter), functional recovery (basic service provision), and full functionality (restoration to pre-earthquake conditions). Several frameworks and design methodologies have been developed in the United States to understand the requirements for achieving these higher functionality states. However, these frameworks predominantly focus on structural and non-structural component damage, with limited consideration for external impacts (e.g., neighbouring buildings) and social factors (e.g., pre-event planning and occupancy policies). Furthermore, their applicability outside the United States and on a broader scale remains uncertain. In response to the disruptions caused by the Canterbury earthquake sequence in 2010/2011, New Zealand government organizations have expressed a desire to move beyond existing life safety codes and increase the country's resilience. The requirements for achieving higher functionality states within a New Zealand context remains an active research area. This poster proposes developing a framework for post-earthquake building functionality tailored explicitly to New Zealand. Tertiary education institutions in New Zealand are selected as a case study due to their functional similarities to other community buildings such as libraries, offices, cafes, and auditoriums. The research objectives include learning from institutions affected by the Canterbury earthquake sequence, understanding the needs of institutions nationwide, and investigating the framework's applicability to other types of buildings. The poster presents these objectives and details
on required resources, timelines, risks, and ethical considerations
Reframing the art of music video: “mismatched eyeballs” and reanimated collaborators
No full textJust as music videos have been studied through many frames, so too, has David Bowie. When undertaking research for my book David Bowie and the Art of Music Video, a reframing process was necessary. This entailed starting from the premise that music videos are functionally diverse, complex audiovisual configurations that go well beyond serving as a promotional tool for a musician or
as a stepping stone to auteurism for a director. This act of ‘reframing’ also involved examining Bowie’s creative process through a fresh lens. Through a combination of contextual research, interview materials, audience research, and multimodal analysis of Bowie’s videos, I demonstrate that Bowie played an important part in helping develop music video as a collaborative artform with diverse functions, affects, and contributions to social, cultural, and political understandings. By reframing the art of music video, this book firmly places the spotlight onto the relations between collaborative process and audiovisual assemblage, demonstrating that music
video is a flexible form for transmedia storytelling, intertextuality, and the remediation of numerous artforms – all of which supports my argument that music video is a form worthy of deep analysis. By reflecting on my response to unique challenges (such as limited
opportunities to interview music video directors), this presentation will discuss my research process as an example of reframing the art (and research) of music video