The 3rd Annual 3MT Competition

The 3rd Annual 3MT Competition was held as a part of the 8th Annual Graduate Students Research Colloquium. The event showcased 9 presentations.

The event featured a prestigious panel of judges which included Jean-Vianney Auclair, A/Assistant Deputy Minister of the Advanced Learning Division, Province of Manitoba; Eric Johnstone, Chair, the University of Board of Regents; Roger Currie, News Director, CJNU Radio. 

Winner:

Oluwayemisi Olugboji - Applied Computer Science & Society

Improving the Prognosis and Diagnosis of Chronic Kidney Disease Using Naive Bayes Algorithm

Big data represents a ground-breaking opportunity for healthcare industry to improve service quality. Clinical Decision Support will help the clinician answer clinical questions quickly when information is needed, especially if the learning method is built into the Electronic Health Record (EHR). This paper delineates how unstructured data can be mined from different point of care devices to improve the prognosis and diagnosis of Chronic Kidney Disease (CKD). This proposed mining system will scrape and analyze data from ubiquitous point of care devices and make predictions. This system will help predict potential events which may be symptoms of Kidney disease or a drug interaction that could lead to kidney problems. CKD is among the top ten leading causes of death in Canada according to Statistics Canada. This research will help clinicians in the areas of chronic disease management, prenatal patients monitoring, preventative medicine, and aggregation of medical information.

Oluwayemisi's win qualified him to travel to Kelowna, BC for the Western Regional Three Minute Thesis Competition. To read more about Oluwayemisi Olugboji's experience at the Western Regional, please visit:

Student Feature: Oluwayemisi Olugboji.

2016 Presentations

Syed Aqeel Awais — Applied Computer Science & Society

Descriptive Approach to Quantifying the Similarity of Graphs

Finding similarity between sets of disjoint objects is at once intuitive and practically applicable. Nearness and similarity between objects can be assessed by their spatial relationship or by feature vectors representing object descriptions. Additionally, sets of objects can be represented by graphs, which provide another avenue for assessing nearness, where each object in a set is associated with the vertex in a graph. Object descriptions associated with vertices and edges determine relationships between these objects and form patterns of interest. The graph edit distance is a well-known measure for accessing similarity of two graphs. In addition to this approach it is possible to assess the similarity of graphs using near set theory. Since, near set theory provides a theoretical framework for quantifying the similarity of disjoint sets. The focus of this work is applying tolerance near set theory to the problem of graph similarity for comparison with graph edit distance.

Joana Grace Beltrano— BioScience, Technology & Public Policy

Living in a Fast Paced World: Can Behaviour Affect the Diversity of Reproductive Adam Genes?

The results of molecular evolutionary studies continually confirm that genes coding for proteins involved in reproduction are rapidly evolving. Sperm and testes-expressed Adam genes are known to undergo bouts of positive selection in mammals. More modification is seen in the Adam genes directly involved with the intricate processes of fertilization, which has lead to speculation that diversification might be driven by species-specific sexual adaptations. This project will study proxies of sexual selection to gain further insight in the pervasive positive selection at sperm surface genes by analysing DNA sequences of species with varying mating systems such as those of Microtus spp.(Voles). DNA sequences will go through molecular evolutionary tests that uses population genetics tools and phylogenetic approaches. By exploring the selective pressures fundamental to the forces of evolution from a molecular stand point, we hope to provide an innovative edge that will have guiding implications for the design of population genetics studies.

Gabriela Jimenez— Development Practice: Indigenous Development

Indigenous Participation In Water Governance To Improve Access To Drinking Water

In Venezuela, the Indigenous population is about 3% of the total population of the country. Similar to Canada, Indigenous communities in Venezuela are an evident minority suffering from lack of drinking water, and whose voices need to be included in the planning process for the sake of their sustainable development. The purpose of this case study is to explore the potential of integrating Indigenous knowledges and participation of Warao communities from Delta Amacuro State, Venezuela, into the existing formal water governance structure as a sustainable alternative to drinking water strategies. The specific objectives pursued are: (i) Document Indigenous knowledges and practices related to drinking water treatment and management used by Warao communities, (ii) Identify the existing challenges in integrating Indigenous knowledge and participation of Warao communities in the local water management systems. (iii) Generate strategies/recommendations to integrate Indigenous knowledges and practices into existing formal water management institutions.

Majing Oloko — Indigenous Governance

Indigenous Food: A Viable Alternative to Food Security

Despite the central role that indigenous foods can potentially play in meeting people’s food security needs in Nigeria, it has continually been ignored by Government and policy makers. In order to encourage the recognition of indigenous food and knowledge, this study focused on indigenous foods utilized by the Irigwe people of North-Central Nigeria, their uses and how they contribute to meeting the people’s food security needs. Data was collected from 30 participants through Interviews and talking circles. Participants identified 23 indigenous species they utilize for nutritional and medicinal purposes. Insecurity and lack of access to farm input were some of the challenges faced by participants in their efforts to access their food. Participants emphasised the need for cultural revitalization, youth engagement and access to modern farming equipments, as ways to remedy challenges and strengthen the community’s food security. 

Ashmeet Singh — Applied Computer Science & Society

Music Classification Using Tolerance Near Sets

Automatic music genre recognition from acoustic features is a very popular music information retrieval (MIR) task where machine learning algorithms have found widespread use. MIR research is a multidisciplinary study that encompasses engineering, computer science and psychology.  Our research involves a tolerance-based machine learning approach to classify music into different genres on the MillionSong Dataset. Tolerance near sets provide the mathematical foundation for music perception based on audio signals. The notion of tolerance is related to the idea of closeness (proximity) between objects, such as image/audio segments that resemble each other with a tolerable level of difference.  A supervised form of learning algorithm is proposed. In the learning phase, a representative of each tolerance class is computed. In the testing phase, test objects are classified based on the lowest distance value from the tolerance class representative.  For a sample data set, our algorithm demonstrates comparable performance with other classical algorithms.

Quinn Webber — Bioscience, Technology, & Public Policy

S-I-R Model: Susceptible, Infected, Recovered Model

Certain individuals can disproportionately infect conspecifics during a pathogen outbreak. Consistent differences in behaviour, or personality, could influence this variation with the most exploratory individuals expected to spread pathogens. I quantified personality of little brown bats (Myotis lucifugus) and tested the hypothesis that exploratory individuals would be more likely to spread a proxy pathogen. I captured 10 groups of 16 bats and held each group in an outdoor flight-tent. I used hole-board tests to quantify exploration and then randomly selected one individual from each group for ‘infection’ with non-toxic, ultraviolet-fluorescent powder. Infected bats were released into the flight-tent for 24-hours after which I photographed individuals under ultraviolet-light and digitized photographs to quantify infection. As predicted, exploratory individuals caused highest intensities of infection for the rest of the group. My results highlight the importance of personality for pathogen dynamics in wildlife and have implications for conservation and public health.

Yexiao Wu — Applied Computer Science & Society

Image Analysis via Discrete Orthogonal Racah Moments

In this research, we have analyzed images with the Racah moment descriptors. First, we have studied the Racah moments computing and verified the unique image feature representation capabilities of Racah moments by performing the image reconstructions. Then, to investigate the corresponding contributions of the nth and mth orders of Racah moments, we have conducted the image reconstructions from partial orders of Racah moments. This leads to the discovery that the nth and mth orders of Racah moments can be called as the horizontal and vertical orders of Racah moments, which preserve the image information horizontally and vertically, respectively.

Jiajie Yu — Applied Computer Science & Society

Descriptive Topological Spaces for Performing Visual Search

The focus of this work is on simulating human visual search in automated systems. Human visual attention can be simulated using either bottom-up or top-down models. The bottom-up approach directs attention based on salient objects in the visual field determined by low level visual features. Further, it can be used to identify candidate salient objects, which form a basis for making high-level comparisons regarding the presence of objects of interest. Similarly, the top-down model assumes some a-prior information is present that is necessary to rule out salient regions of the visual field that are not similar to the object of interest. This research investigates the use of descriptive topological spaces for generating patterns from salient regions in order to make decisions on the whether a digital image contains the desired object. Further, the tolerance-based descriptive set intersection operator between patterns is the main mechanism used to determine similarity.

For more 3MT presentations, please visit here.