Theses, Dissertations, and other Required Graduate Degree Essays

Receive updates for this collection

This collection contains digitized SFU theses except for those theses submitted within the last 12 months. If you cannot find the thesis you are looking for please search Recently Submitted Theses as it may be a recently submitted thesis and thus not yet available in Summit.

Attention-based skin lesion recognition

Author: 
Date created: 
2020-04-07
Abstract: 

Skin cancer is one of the most common types of cancers in the world and is a big concern for people’s health. In recent years, automatic algorithms to recognize skin cancers from dermoscopy images have gained lots of popularity, especially deep-learning-based methods. In this thesis, we propose an attention-based deep learning model for skin cancer recogni- tion. The attention modules, which are learned together with other network parameters, estimate attention maps that highlight image regions of interest that are relevant to lesion classification. These attention maps provide a more interpretable output as opposed to only outputting a class label. Additionally, we propose to utilize prior information by regulariz- ing attention maps with regions of interest (ROIs) (e.g., lesion segmentation or dermoscopic features). To our knowledge, we are the first to introduce an end-to-end trainable attention module with regularization for skin cancer recognition. We provide both quantitative and qualitative results on public datasets to demonstrate the effectiveness of our method. Experiments show that: (1) the attention module is capable of ruling out irrelevant areas in the image; (2) when the proposed attention regularization terms are applied, both the classification performance and the attention maps can be further refined; (3) the attention regularization is quite robust and flexible in that it can take advantage of sparse or even imperfect ROI maps. The code of this work is released at https://github.com/SaoYan/IPMI2019-AttnMel.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Ghassan Hamarneh
Department: 
Applied Sciences: School of Computing Science
Thesis type: 
(Thesis) M.Sc.

Restoring the comfort of home: Addressing the challenge of placing hard-to-house populations in seniors’ social housing in British Columbia

Author: 
Date created: 
2020-04-08
Abstract: 

Since the early 2000s, tenants of seniors’ social housing in BC have increasingly shared their buildings with younger persons who have severe mental illnesses and/or addiction issues. While this demographic shift does not neatly correspond with a specific policy change, academics, media sources, and the experts and stakeholders interviewed for this report all have suggested that it results from the prioritization of the hard-to-house by the provincial government. For many seniors, this new environment has produced a host of negative outcomes: increased levels of fear; greater social isolation; more disruptive and unpredictable living conditions; and exposure to criminal activity, threats, violence, and other disturbing or dangerous behaviors. This paper examines the emergence of this policy problem and explores possible policy solutions. It does this through a literature review, six case studies from American jurisdictions, and thirteen interviews with experts and stakeholders. Ultimately, the paper recommends two interventions: funding and creating training materials for resident service coordinators, and an environmental scan of the approaches currently being made by the more than 550 non-profit housing organizations which provide nearly 90% of British Columbia’s social housing units.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Joshua Gordon
Department: 
Arts & Social Sciences: School of Public Policy
Thesis type: 
(Project) M.P.P.

Fractured Foundations: Distrust and democratic decline in Canada

Author: 
Date created: 
2020-03-09
Abstract: 

This paper investigates the declining levels of trust in government and its impact on Canada’s democracy. Trust is foundational for the rule of law, economic growth, government stability and the development of political capacity in citizens. The extent of the trust deficit in Canada is determined by analyzing data recently collected by the Morris J. Wosk Centre for Dialogue’s national survey on democratic culture. The primary causes and consequences of distrust are identified using the survey data and interviews with academic experts. The research results suggest increasing citizens’ opportunities to meaningfully participate in government is the strongest approach to improving trust in government. Citizens’ reference panels, participatory budgeting and reforming to a proportional representation system are the specific options evaluated using standardized criteria and measures. The policy analysis demonstrates that implementing national participatory budgeting and citizens’ reference panels would both be effective steps towards rebuilding trust and increasing citizens’ capacities.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Doug McArthur
Department: 
Arts & Social Sciences: School of Public Policy
Thesis type: 
(Project) M.P.P.

Functional neural networks for scalar prediction

Author: 
Date created: 
2020-04-07
Abstract: 

We introduce a methodology for integrating functional data into densely connected feed-forward neural networks. The model is defined for scalar responses with at least one functional covariate and some number of scalar covariates. A by-product of the method is a set of functional parameters that are dynamic to the learning process which leads to interpretability. The model is shown to perform well in a number of contexts including prediction of new data and recovery of the true underlying coefficient function; these results were confirmed through cross-validations and simulation studies. A collection of useful functions are built on top of the Keras/Tensorflow architecture allowing for general use of the approach.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Jiguo Cao
Department: 
Science: Department of Statistics and Actuarial Science
Thesis type: 
(Project) M.Sc.

Listening for listening in art and inquiry

Author: 
Date created: 
2020-04-09
Abstract: 

Reflecting on my ways of being in the world as an artist, researcher, student, and mentor, I am drawn by the recurring theme of listening—not only in the sense of auditory perception, but also a wider attunement to the world, involving all my intermingling senses. Arts-based research practices of living inquiry, performative inquiry, embodied inquiry invite me to explore the multiple ways I enact listening in different contexts of my life, such as when reading, dancing, writing, transcribing and facilitating art engagement. Through poetry, theory, life-writing and meditations on my embodied experiences, I observe how different metaphors, intentions or practices can guide and enable different kinds of listening experiences. In particular, I propose that listening transcends an act of reception, constituting a creative and dialogical encounter. Listening calls us to release expectations, preconceptions and control in order to enliven our desire for curiosity, discover new possibilities, and bring forth our own unique voice.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Lynn Fels
Department: 
Education: Faculty of Education
Thesis type: 
(Thesis) M.A.

Search (and rescue) for the ultimate selfie: How the use of social media and smartphone technology have affected human behaviour in outdoor recreation scenarios

Author: 
Date created: 
2020-04-07
Abstract: 

The practice of outdoor recreation was historically a form of therapy and escape from the rigors of modern industrial daily work-life, and it remains a favored pastime today, with 70% of Canadians and 91% of British Columbia residents participating in “outdoor recreation or wilderness activities”. In recent years, there is a belief that the surge in popularity of hiking is due to beautiful destinations becoming more visible on social media. Further, the proximity of urban centres like Vancouver to such destinations reassures users that the safety benefits of urban technologies including smartphones, will remain accessible and reliable throughout their outdoor exploration and that help is available in the event of an emergency. This belief has led to many instances of Search and Rescue teams being activated, which would previously have been avoided by outdoor recreation participants making different choices based on their skill and experience. The culture of outdoor recreation has therefore been increasingly affected by smartphone technology in terms of users’ risk perception while recreating outdoors.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Peter Anderson
Department: 
Communication, Art & Technology: School of Communication
Thesis type: 
(Thesis) M.A.

Tensor completion methods for collaborative intelligence

Author: 
Date created: 
2020-03-20
Abstract: 

In the race to bring Artificial Intelligence (AI) to the edge, collaborative intelligence has emerged as a promising way to lighten the computation load on edge devices that run applications based on Deep Neural Networks (DNNs). Typically, a deep model is split at a given layer into edge and cloud sub-models. The deep feature tensor produced by the edge sub-model is transmitted to the cloud, where the remaining computationally intensive workload is performed by the cloud sub-model. The communication channel between the edge and cloud is imperfect, which will result in missing data in the deep feature tensor received at the cloud side, an issue that has mostly been ignored by existing literature on the topic. In this thesis I study four methods for recovering missing data in the deep feature tensor. Three of the studied methods are existing, generic tensor completion methods, and are adapted here to recover deep feature tensor data, while the fourth method is newly developed specifically for deep feature tensor completion. Simulation studies show that the new method is 3 − 18 times faster than the other three methods, which is an important consideration in collaborative intelligence. For VGG16’s sparse tensors, all methods produce statistically equivalent classification results across all loss levels tested. For ResNet34’s non-sparse tensors, the new method offers statistically better classification accuracy (by 0.25% − 6.30%) compared to other methods for matched execution speeds, and second-best accuracy among the four methods when they are allowed to run until convergence.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Ivan Bajic
Department: 
Applied Sciences: School of Engineering Science
Thesis type: 
(Thesis) M.A.Sc.

Paving the Pathway – Expanding youth initiatives to improve high school graduation rates of at-risk students in Vancouver

Author: 
Date created: 
2020-03-16
Abstract: 

High school graduation is an important milestone for future employment and opportunities, income potential and overall well-being. However, too many students in low-income neighbourhoods fail to complete high school. In Canada, youth from low-income families are three times more likely to drop out of high school than youth from middle-income families. Negative peer effects in school and weak neighbourhood dynamics impact socio-emotional development and educational outcomes of youth. Investments in education can help overcome barriers associated with poverty. This capstone assesses how comprehensive youth initiatives can be expanded to improve high school graduation rates of at-risk students in Vancouver, British Columbia. Five criteria were used to assess policy options and their ability to increase high school graduation outcomes. Through literature reviews, interviews and case studies, a recommendation is made for the Ministry of Education to establish an After-School Grant Program and to request written proposals from interested schools and organizations.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
John Richards
Department: 
Arts & Social Sciences: School of Public Policy
Thesis type: 
(Project) M.P.P.

Advanced subspace methods for low/mid-level vision

Author: 
Date created: 
2020-02-19
Abstract: 

Low- and mid-level vision tasks are fundamental to computer vision. They are important not only in themselves but also for higher-level tasks as cornerstones. Low-level tasks are basically about extracting primitive information, such as edges, textures, and correspondences from images. And mid-level tasks, from the Gestalt psychologists' perspective, are grouping mechanisms on low-level visual information. In particular, inferring the geometric information from images and segmenting an image into object-level regions are two major aspects of mid-level tasks. In this thesis, we make advances in solving real-world low- and mid-level problems using subspace based representations. For monocular visual SLAM, we solve the visual odometry in a rank-1 factorization and solve the pose-graph optimization by linear programming in multi-stage, which are more robust to initialization errors in the local 3D maps and the global pose-graph respectively. For dense 3D reconstruction, which is also a mid-level task, we represent a depth map as a linear combination of several basis depths from an underlying subspace, and learn a convolutional neural network to generate such a basis. To estimate the depth maps as well as the camera poses, we propose a differentiable bundle adjustment layer that optimizes for the depth map and camera poses by minimizing a feature-metric error. The feature-metric error is defined over a feature pyramid, which is learned jointly with the basis generator end-to-end. For broader low-level vision tasks, we also adopt a basis representation, but for a different purpose. Conventionally, a low-level task is formulated as a continuous energy minimization problem, where the objective function contains a data fidelity term and a smoothness regularization term. We replace the regularization term with a learnable subspace constraint and define the objective function only with the data term. This methodology unifies the network structures and the parameters for many low-level vision tasks and even generalizes to unseen tasks, as long as the corresponding data terms can be formulated. In summary, we explore the subspace based methods from manually derived low-rank formulation to learning based subspace minimization, which are conceptually novel compared to the existing methods. To demonstrate the effectiveness of the proposed methods, we conduct extensive experiments for all the involved tasks on public benchmarks as well as our own data. The results show that our methods have achieved comparable or better performance than state-of-the-art methods with better computational efficiency.

Document type: 
Thesis
File(s): 
Senior supervisor: 
Ping Tan
Department: 
Applied Sciences: School of Computing Science
Thesis type: 
(Thesis) Ph.D.

Material Matters: Using Regulation to Improve the Canadian Mining Industry’s Human Rights Record

Author: 
Date created: 
2020-03-04
Abstract: 

Canada has the most incorporated mining entities in the world. The Canadian mining industry has an international reputation as a mining power, but also one for human rights violations. The literature indicates that insufficient domestic accountability mechanisms, such as non-financial reporting, are one of the reasons why human rights violations persist in this industry. This study addresses the regulatory gap within Canadian securities regulations and identifies policy options aimed at improving the lack of accountability within the Canadian mining industry. Three policy options are evaluated, including: incorporating the term “salient human rights impacts” into the existing regulations; adjusting the definition of materiality to include human rights violations; and direct reporting to the federal government. Based off this analysis, a two-pronged approach including mandatory reporting on salient human rights impacts, alongside federal submissions and audits, is recommended as a possible solution to human rights violations occurring within the Canadian mining industry.

Document type: 
Graduating extended essay / Research project
File(s): 
Senior supervisor: 
Maureen Maloney
Department: 
Arts & Social Sciences: School of Public Policy
Thesis type: 
(Project) M.P.P.