Since the establishment of West Nile virus (WNV) in North America, concern has arisen that this virus may also pose a biosecurity threat to Australia. West Nile virus is maintained in an enzootic transmission cycle between birds and ornithophilic Culex spp. mosquitoes. Since its introduction into the United States in 1999, it has been responsible for over 11,000 human cases of neurological disease and over 1,000 human deaths. In addition, WNV has caused extensive equine and avian mortality. Should an exotic, virulent strain of WNV be introduced into Australia, it may have significant impact on human and animal health. This thesis examines entomological factors which may influence the potential for an exotic, pathogenic strain of WNV to become established in eastern Australia and be maintained in urban transmission cycles. Further, it aims to elucidate the most likely candidate vectors for the maintenance of WNV in natural transmission cycles. Components include: a) laboratory-based vector competence experiments to assess the ability of Australian mosquito species to become infected with and transmit a North American strain of WNV; b) identification of the vertebrate origin of mosquito bloodmeals using serological and molecular assays; c) an evaluation of the efficacy of bird-baited traps for collecting ornithophilic mosquitoes at ground level and in the tree canopy; and d) examination of the prevalence of endemic arboviruses, including Kunjin virus (KUNV; an Australian subtype of WNV) in mosquitoes collected from urban centres of eastern Australia. Vector competence experiments were conducted with field-collected Australian mosquito species. Nineteen species were were exposed to a 1999 New York strain of WNV using an artificial membrane feeding apparatus and transmission was attempted using an in vitro modified capillary tube method. A number of Culex species, including Culex quinquefasciatus and Culex gelidus, demonstrated high vector competence for WNV. Culex annulirostris, the primary KUNV vector, was the most competent vector, displaying transmission rates of 84%. Other common species, including Aedes vigilax, Aedes procax and Verrallina spp. were less susceptible to infection. Approximately 1,200 blood engorged mosquitoes were analysed using an enzyme-linked immunosorbent assay to determine host species. Overall, 90% of bloodmeals from Cx. quinquefasciatus were avian in origin, while 22% of bloodmeals from Cx. annulirostris were from birds. Molecular assays were further used to determine which bird species were commonly targeted by a selection of ornithophilic mosquito species. Mitochondrial DNA sequence data indicated that most (93%) avian bloodmeals from Cx. quinquefasciatus were obtained from passerine birds (order Passeriformes). Avian species commonly targeted by Cx. quinquefasciatus included Sphecotheres vieilloti (Australasian figbird; 39% of total identified), Sturnus tristis (common myna; 19%) and Philemon buceroides (helmeted friarbird; 17%). Alternatively, the majority (87%) of avian bloodmeals identified from Cx. annulirostris were from non-passerine birds, with most bloodmeals obtained from ducks (72%; family Anatidae). Seven field-based Latin Square trials conducted in Brisbane and Cairns showed that conventional CO2-baited CDC traps yielded higher numbers of Culex spp. than novel bird-baited traps, both at ground level and in the tree canopy. In Brisbane, CO2-baited CDC traps placed at ground level (approximately 1 m from ground) collected significantly more Cx. annulirostris than those at canopy height (5-10 m) during two trials, and more Ae. vigilax and Aedes procax during one trial each. Conversely, in Cairns, CO2-baited CDC traps placed in the tree canopy collected significantly more Culex squamosus during two trials, and significantly more Aedeomyia catasticta and Culex cubiculi during one trial each. Consistently low yields of ornithophilic species including Culex australicus and Cx. quinquefasciatus in all traps tested suggests the need to employ alternative trapping measures to target these species in urban areas. Over 1,800 mosquito pools (comprising over 65,000 individual mosquitoes collected from urban environments) were tested for evidence of arboviral infection. None of the pools showed evidence of flavivirus infection, but two alphaviruses were isolated from mosquito pools from Cairns. One isolate of Barmah Forest virus was obtained from a pool of 100 Ae. vigilax, and a pool of 89 Verrallina carmenti yielded an isolate of Ross River virus. In addition, an isolate most similar to a newly described bunyavirus, Stretch Lagoon orbivirus, was obtained from a single Aedes alternans collected from Sydney. The outcomes of this thesis demonstrate that a number of Australian mosquito species can become infected with and transmit a virulent strain of WNV, should it be introduced into Australia. Analysis of the host feeding patterns of these potential vectors in urban environments demonstrates that a number of candidate vectors, including Cx. quinquefasciatus and Cx. annulirostris, readily feed on common bird species. Further, the absence of detectable flavivirus infection in urban mosquito populations shown here suggests that endemic urban flaviviruses would have minimal influence on potential WNV transmission. Notwithstanding other ecological factors that may influence the likelihood of enzootic or epizootic transmission including susceptibility of vertebrate hosts and mechanisms of introduction, the potential establishment of an exotic, virulent strain of WNV in urban areas of Australia is certainly possible given the evidence provided in this thesis.

The problem of embedding a manifold in Euclidean space is considered. Manifolds are introduced in Chapter I along with other basic definitions and examples. Chapter II contains a proof of the Regular Value Theorem along with the "Easy" Whitney Embedding Theorem. In Chapter III, vector bundles are introduced and some of their properties are discussed. Chapter IV introduces the Stiefel-Whitney classes and the four properties that characterize them. Finally, in Chapter V, the Stiefel-Whitney classes are used to produce a lower bound on the dimension of Euclidean space that is needed to embed real projective space.

As the quantity of text documents created on the web grows the ability of experts to manually classify them has decreased. Because people need to find and organize this information, interest has grown in developing automatic means of categorizing these documents. In this effort, ontologies have been developed that capture domain specific knowledge in the form of a hierarchy of concepts. Support Vector Machines are machine learning methods that are widely used for automated document categorization. Recent studies suggest that the classification accuracy of a Support Vector Machine may be improved by using concepts defined by a domain ontology instead of using the words that appear in the document. However, such studies have not taken into account the hierarchy inherent in the relationship between concepts. The goal of this dissertation was to investigate whether the hierarchical relationships among concepts in ontologies can be exploited to improve the classification accuracy of web documents by a Support Vector Machine. Concept vectors that capture the hierarchy of domain ontologies were created and used to train a Support Vector Machine. Tests conducted using the benchmark Reuters-21578 data set indicate that the Support Vector Machines achieve higher classification accuracy when they make use of the hierarchical relationships among concepts in ontologies.

The objective of this work is to study state-of-the-art deep neural networks based speaker verification systems called x-vectors on various conditions, such as wideband and narrowband data and to develop the system, which is robust to unseen language, specific noise or speech codec. This system takes variable length audio recording and maps it into fixed length embedding which is afterward used to represent the speaker. We compared our systems to BUT's submission to Speakers in the Wild Speaker Recognition Challenge (SITW) from 2016, which used previously popular statistical models - i-vectors. We observed, that when comparing single best systems, with recently published x-vectors we were able to obtain more than 4.38 times lower Equal Error Rate on SITW core-core condition compared to SITW submission from BUT. Moreover, we find that diarization substantially reduces error rate when there are multiple speakers for SITW core-multi condition but we could not see the same trend on NIST SRE 2018 VAST data.

As Artificial Intelligence systems are used by human users at an increasing frequency, the need for such systems to understand and predict human behavior likewise increases. In my work, I have considered how to predict human behavior in repeated games. These repeated games can be applied as a foundation to many situations where a person may interact with an AI, In an attempt to create such a foundation, I have built a system using Attitude Vectors used in automata to predict actions based on prior actions and communications. These Attitude Vector Automata (AVA) can transform information from actions in one game with a given payoff matrix into actions in another game. Results show that prediction accuracy was ultimately below other, similar work, in general in several repeated games. There are however some aspects, such as scenarios involving lying, in which my predictor showed potential to outperform these other systems. Ultimately, there is potential in using ideas presented as AVA to build a potentially more robust system for future efforts in human behavior prediction.

This project describes the fundamental concepts and basic theory of polarization mode dispersion (PMD) in optical fibers. There are described basic the relation between Jones vectors and Stokes vectors, rotation matrices, the definition and representation of PMD vectors, the laws of infinitesimal rotation. After the introduction the first conception with bases of polarization mode dispersion (PMD) in optical fibers, they have become an important body of knowledge basic for the design of high-capacity optical communication systems. PMD effects are linear electromagnetic propagation phenomena occurring in so-called “single-mode“fibers. Despite their name, these fibers support two modes of propagation distinguished by their polarization. Because of optical birefringence in the fiber, the two modes travel with different group velocities, and the random change of this birefringence along the fiber length results in random coupling between the modes. With current practical transmission technology the resulting PMD phenomena lead to pulse distortion and system impairments that limit the transmission capacity of the fiber. I describe different ways of measuring PMD in optical fibre, PMD compensation techniques and analyse PMD results in optical fibres.

Authorship identification is the field of classifying a given text by its author based on the assumption that authors exhibit unique writing styles. This thesis investigates the semantic shortcomings of the vector space model by constructing a semantic kernel created from WordNet which is evaluated on the problem of authorship attribution. A multiclass SVM classifier is constructed using the one-versus-all strategy and evaluated in terms of precision, recall, accuracy and F1 scores. Results show that the use of the semantic scores from WordNet degrades the performance compared to using a linear kernel. Experiments are run to identify the best feature engineering configurations, showing that removing stopwords has a positive effect on the financial dataset Reuters while the Kaggle dataset consisting of short extracts of horror stories benefit from keeping the stopwords.

No description available.

Precipitation is a dominating quantity in microwave radiometry. The large emission and scattering signals of rain and ice, respectively, introduce large contributions to the measured brightness temperature. While this allows for accurate sensing of precipitation, it also results in degraded performance when retrieving other geophysical parameters, such as near-surface ocean winds. In particular, the retrieval of wind direction requires precise knowledge of polarization, and nonspherical particles can result in a change in the polarization of incident radiation. The aim of this dissertation is to investigate the polarizing effects of precipitation in the atmosphere, including the existence of a precipitation signal in the third Stokes parameter, and compare these effects with the current sensitivities of passive wind vector retrieval algorithms. Realistic simulated precipitation profiles give hydrometeor water contents which are input into a vector radiative transfer model. Brightness temperatures are produced within the model using a reverse Monte Carlo method. Results are produced at three frequencies of interest to microwave polarimetry, 10.7 GHz, 18.7 GHz, and 37.0 GHz, for the first 3 components of the Stokes vector.

This thesis combines Keysight’s SystemVue software with a Vector Signal Analyzer (VSA) and Vector Signal Generator (VSG) to test receiver equalization schemes over physical channels. The testing setup, “Equalization Verification,” is intended to be able to evaluate any equalization scheme over any physical channel, and a decision-directed feed-forward LMS equalizer is used as an example. The decision-directed feed-forward LMS equalizer is shown to decrease the BER from 10-2 to 10-3 (average of all trials) over a CAT7 and CAT6A cable, both simulated and physical, for 1GHz and 2GHz carrier, and 80MHz data rate. A wireless channel, 2.4GHz Dipole Antenna, is also tested to show that the addition of the equalization scheme decreases BER from 10-5 to less than 10-5. Then the simulation and equalization parameters (LMS step size, PRBS, etc.) are changed to further verify the equalization scheme. The simulated channel BER results do not always match the physical channel BER results, but the equalization scheme does decrease BER for both wired and wireless channels. Then transistor-based equalization model is created using both HDL SystemVue components and blocks easily implemented by transistors. The model is then verified using HDL, Spice, and SystemVue simulation. Overall this thesis accomplishes its goal of creating a testing setup, Equalization Verification, to show that adding a given simulated equalization scheme in SystemVue can improve the quality of the link, by decreasing BER by at least an order of magnitude, over a specific physical channel.