Speech communication is gaining in popularity in many different contexts as technology evolves. With the introduction of mobile electronic devices such as cell phones and laptops, and fixed electronic devices such as video and teleconferencing systems, more people are communicating which leads to an increasing demand for new services and better speech quality. Methods to enhance speech recorded by microphones often operate blindly without prior knowledge of the signals. With the addition of multiple microphones to allow for spatial filtering, many blind speech enhancement methods have to operate blindly also in the spatial domain. When attempting to improve the quality of spoken communication it is often necessary to be able to reliably determine the location of the speakers. A dedicated source localization method on top of the speech enhancement methods can assist the speech enhancement method by providing the spatial information about the sources. This thesis addresses the problem of speech-source localization, with a focus on the problem of localization in the presence of multiple concurrent speech sources. The primary work consists of methods to estimate the direction of arrival of multiple concurrent speech sources from an array of sensors and a method to correct the ambiguities when estimating the spatial locations of multiple speech sources from multiple arrays of sensors. The thesis also improves the well-known SRP-based methods with higher-order statistics, and presents an analysis of how the SRP-PHAT performs when the sensor array geometry is not fully calibrated. The thesis is concluded by two envelope-domain-based methods for tonal pattern detection and tonal disturbance detection and cancelation which can be useful to further increase the usability of the proposed localization methods. The main contribution of the thesis is a complete methodology to spatially locate multiple speech sources in enclosed environments. New methods and improvements to the combined solution are presented for the direction-of-arrival estimation, the location estimation and the location ambiguity correction, as well as a sensor array calibration sensitivity analysis.

Meniere's disease is a common inner ear disorder that affects balance and hearing. Electrovestibulography (EVestG) is a relatively new vestibular driven test that measures spontaneous and driven field potential activity recorded in the external ear canal in response to various vestibular stimuli. The main objectives of this thesis were to record and analyze EVestG signals in order to 1) testify whether the EVestG technology is capable of classifying individuals with Meniere’s from healthy ones, and if it is, then 2) identify the EVestG tilt stimulus providing the most informative response in relation to identifying Meniere’s symptoms; thus, optimizing the EVestG experimental protocol as a Meniere’s disease diagnostic aid. EVestG signals of two groups of Meniere’s and control individuals during seven different EVestG tilt stimuli were recorded and analyzed by linear and nonlinear signal processing techniques. Data of 14 with Meniere’s disease and 16 healthy individuals were used as the training set, while additional data of 21 individuals with vertiginous disorders (and suspected of Meniere’s disease) and 10 controls were used as the test set. An ad-hoc voting classifier built upon single-feature linear classifiers was designed, and used for classification of the two groups of both training and test datasets. The results showed an overall accuracy of 87% and 84% for training and test datasets, respectively. Among the seven different tilts that each evokes a specific part of the inner ear organ, the side tilt which stimulates most of the labyrinth and particularly the utricle, was found to generate the best characteristic features for identifying Meniere’s disease from controls. Thus, one may simplify the EVestG protocol to only the side tilt stimulus for a quick screening of Meniere’s disease. The proposed method encourages the use of EVestG technology as a non-invasive and potentially reliable diagnostic/screening tool to aid clinical diagnosis of Meniere’s diseases. / October 2016

A key component of autonomous outdoor navigation in unstructured environments is the classification of terrain. Recent development in the area of machine learning show promising results in the task of scene segmentation but are limited by the labels used during their supervised training. In this work, we present and evaluate a flexible strategy for terrain classification based on three components: A deep convolutional neural network trained on colour, depth and infrared data which provides feature vectors for image segmentation, a set of exchangeable segmentation engines that operate in this feature space and a novel, air pressure based actuator responsible for distinguishing rigid obstacles from those that only appear as such. Through the use of unsupervised learning we eliminate the need for labeled training data and allow our system to adapt to previously unseen terrain classes. We evaluate the performance of this classification scheme on a mobile robot platform in an environment containing vegetation and trees with a Kinect v2 sensor as low-cost depth camera. Our experiments show that the features generated by our neural network are currently not competitive with state of the art implementations and that our system is not yet ready for real world applications.

In this discussion paper, we argue that the literature on tree algorithms is very fragmented. We identify possible causes and discuss good and bad sides of this situation. Among the latter is the lack of free open-source implementations for many algorithms. We argue that if the community adopts a standard of creating and sharing free open-source implementations for their developed algorithms and creates easy access to these programs the bad sides of the fragmentation will be actively combated and will benefit the whole scientific community. (authors' abstract)

In this thesis we make several contributions to the theory of moduli spaces of smooth manifolds, especially in dimension two. In Chapter 2 (joint with Soren Galatius) we give a new geometric proof of a generalisation of the Madsen-Weiss theorem, which does not rely on the tangential structure under investigation having homological stability. This allows us to compute the stable homology of moduli spaces of surfaces equipped with many different tangential structures. In Chapter 3 we give a general approach to homological stability problems, especially focused on stability for moduli spaces of surfaces with tangential structure. We give a sufficient condition for a structure to exhibit homological stability, and thus obtain stability ranges for many tangential structures of current interest (orientations, maps to a simply-connected background space, etc.), which match or improve the previously known ranges in all cases. In Chapter 4 we define and study the cobordism category of submanifolds of a fixed background manifold, and extend the work of Galatius-Madsen-Tillmann-Weiss to identify the homotopy type of these categories. We describe several applications of this theory. In Chapter 5 we compute the stable (co)homology of the non-orientable mapping class group, and find a family of geometrically-defined torsion cohomology classes. This is in contrast to the oriented mapping class group, where few are known. In Chapter 6 (joint with Johannes Ebert) we study the divisibility of certain characteristic classes of bundles of unoriented surfaces introduced by Wahl, analogues of the Miller-Morita-Mumford classes for unoriented surfaces. We show them to be indivisible in the free quotient of cohomology.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

The Swede, Carl Linnaeus (1707-1778), introduced an artificial " Sexual System " of plant classification in 1735, and a binomial system of nomenclature in 1753. They made plant identification much easier. The Linnaean period in Britain lasted from 1760 until [1810-]1830. It is demonstrated that it was during this period that it was first recognised that an unusually high number of rare plants grow in Upper Teesdale. Most of the rare plants of the then very remote Upper Teesdale were discovered shortly after 1783 by William Oliver (1760-1816), alone. He was a surgeon and part of a medical dynasty. How he became a botanist, with his medical background, is examined in detail. He trained at Edinburgh but did not do botany. However, he knew John Hope, the Professor of Botany. Hope was one of only two people teaching the Linnaean system in Britain at this time. The appearance of Linnaean floras of Britain in English from the 1770's onwards made field botany accessiblet o anyone. Previously complex natural systems of plant classification and the use of Latin had restricted access. How Oliver's discoveries were made known is examined in detail. It involved Rev. John Harriman (1760-183 1) who was influenced by the Linnean Society of London, formed in 1788, and the Linnaean English Botany which began in 1790. H-e wanted to become a Fellow of the Linnean Society. James Edward Smith was President of the Linnean Society and an author, with James Sowerby, ofEnglish Botany. IV alic, ,j Lrf Edward Robson (1763-1813), a Quaker botanist and already an Associate of the Linnean Society, and his compilation: Plantae rariores agro Dunelmensi indigenae of 1798, and John Binks (1766-1817), an artisan botanist. Medicine made botanists of both Harriman and Binks, as well as Oliver. Linnaeus influenced the teaching of materia medica (the plant simples).

Master of Science / Department of Computing and Information Sciences / William H. Hsu / Logistic regression is a model for function estimation that measures the relationship between independent variables and a categorical dependent variable, and by approximating a conditional probabilistic density function using a logistic function, also known as a sigmoidal function. Multinomial logistic regression is used to predict categorical variables where there can be more than two categories or classes. The most common type of algorithm for optimizing the cost function for this model is gradient descent. In this project, I implemented logistic regression using conjugate gradient descent (CGD). I used the 20 Newsgroups data set collected by Ken Lang. I compared the results with those for existing implementations of gradient descent. The conjugate gradient optimization methodology outperforms existing implementations.

The Civano community of Tucson, Arizona, is built for sustainability. Trees and plants are precious resources in the community and balancing human needs and natural resources. The design of rainwater harvesting systems and the usage of reclaimed water inside the community effectively irrigate plants and save drinking water. This project estimates canopy changes over time and explores the effect of water resources on plant growth for developed areas and natural areas, respectively. This project generates land cover classifications for 2007, 2010, and 2015 using supervised classification method and measures canopy cover change over time. Based on City of Tucson Water “harvesting rainwater guide to water-efficient landscaping”, this project discusses if water supply meets plant water demand in the developed areas of the community. Additionally, the normalized difference vegetation index (NDVI) data for developed area and natural area over ten years are compared and provide a correlation analysis with water sources. The results show that canopy cover across the entire community decreased from 2007 to 2010, then increased from 2010 to 2015. Water supply in the developed areas is sufficient for plant water demand. In natural areas plant growth changes dramatically as a result of precipitation fluctuation. In addition, it’s proved that 2011 National Land Cover Database (NLCD) tree canopy underestimates canopy cover in the Civano community. The final products not only provide the fundamental canopy cover data for other studies, also serve as a reference of water efficient landscaping within a community.