• Canada
• Research dataclose
• Other research productsclose
• Open Access close
• MSpace at the University of Manitoba close
• Hindawi Publishing Corporation close

Dataset of driving behaviour within the perimeter of Winnipeg collected from May 2009 to May 2010.

The current telephone network protocol is best suited for audio transmissions, and is not particularly video- or data-friendly. However, data traffic growth is four times that of voice traffic. This trend indicates the need for a network protocol that incorporates multi-media traffic. One solution is a distributed intelligence network running over an ATM network. The intelligence is distributed in such a way as to allow the switches to provide bandwidth and some intelligence, while allowing end terminals to provide for additional functionality. This is similar to the Internet model which also employs distributed intelligence across its network. The Specification and Description Language (SDL) is used to develop and simulate the protocol. Possible applications for this protocol include use as a firewall for connection to the telephone network and as an advanced server providing supplementary services such as call forwarding based on the callerID.

The recruitment of participants to online research can be difficult when they must meet restrictive requirements, a situation we faced in recruiting the parents of 2-month-olds. Here we describe a new method, blog commenting, and compare it to the more common online technique of posting recruitment information on parent-oriented online forums. In the blog method, we searched blogs for infant-specific terms and phrases; we then read entries from those retrieved blogs and identified ones written by a parent of an appropriately aged infant. We then posted to the blog a comment in which we invited the parent to participate and to visit our research web site. Rates of study completion and most participant characteristics did not differ for blog- and forum-recruited participants. We discuss the particular strengths and weaknesses of blog recruiting and conclude that it is well suited for topics that people care to write about.

Ce mémoire est le résultat d’une exploration des thèmes de la quotidienneté et de la banalité qui ont mené à un projet de création. La première partie de ce travail est composée d’un court essai qui s’intéresse à trois axes en particulier : les entrées et les sorties de la narration, les lectures de tous les genres, ainsi qu’un commentaire sur quelques-uns des textes créés dans le cadre de cette maîtrise. La deuxième partie est consacrée à la création et est constituée d’un recueil de quatorze nouvelles intitulé Nouvelles orphelines.

Pseudomonas chlororaphis strain PA23 is able to inhibit fungal phytopathogens, primarily through production of excreted secondary metabolites, deeming this organism a promising biocontrol candidate. Antibiotics, including phenazine (PHZ) and pyrrolnitrin (PRN), in addition to degradative enzymes and volatile organic compounds, make up the arsenal that combats pathogenic fungi like Sclerotinia sclerotiorum. Production of these exometabolites is tightly controlled via a hierarchy, which includes the Phz quorum sensing (QS) system. Previous RNA-seq analysis focusing on two QS-deficient strains, namely PA23phzR and PA2--AHL, revealed that 13% of the PA23 genome is QS-regulated. A number of unique differentially expressed genes in the QS-deficient strain compared to the phzR mutant suggested that additional regulators are involved in PA23 QS. A second system, called Csa (cell surface alterations) has been characterized in the closely related P. chlororaphis 30-84. While csaI and csaR are present in the PA23 genome, the role of this system in PA23 biocontrol has not yet been investigated. The purpose of this research was to characterize the PA23 Csa QS system and elucidate its role in regulating secondary metabolite production. Single mutants of the Csa and Phz QS systems, PA23csaR, PA23csaI and PA23phzI, in addition to double mutants, PA23phzRcsaR and PA23phzIcsaI, were generated and subjected to phenotypic analysis. Results indicated that the Csa system negatively regulates PA23 AF activity. PA23csaR and PA23csaI, as well as the phzIcsaI double mutant, all exhibited increased S. sclerotiorum inhibition in plate assays. While phenotypic analysis revealed very minor differences in chitinase, protease, and HCN production as well as biofilm formation in one or both csa mutants, these differences ii alone are insufficient to account for the increased antifungal (AF) activity. Expression of prnA, which encodes the primary antibiotic involved in PA23 biocontrol, was unchanged in the csa mutant background suggesting that PRN is likely not affected. Therefore, the traits responsible for the enhanced AF activity have yet to be identified. Additionally, the relationship between the Csa QS system and other components of the regulatory cascade require further investigation to elucidate to what degree, directly or indirectly, the Csa QS regulates PA23 biocontrol.

Understanding which parts of an image are most salient is an ongoing research problem in the field of computer vision. In this thesis, we consider the problem of saliency prediction insofar as this can benefit from learning image inpainting error. We address this problem by presenting a novel approach to predict saliency maps. The first model learns which parts of the image are most difficult to predict by applying an inpainting algorithm on a regular grid and measuring the error in the resulting prediction. A convolutional neural network is trained to predict the degree of error subject to this inpainting process. A second network uses transfer learning from the first network to predict saliency maps by taking advantage of image inpainting error. We demonstrate that saliency prediction can benefit from first learning image inpainting error. We evaluate and compare our results both by considering image inpainting error, and also through an ablation study we consider a comparison to direct training on the saliency data without transfer learning. We then evaluate our results with the previous state of the art models. We test our networks on two well-known saliency datasets including CAT2000, and SALICON.