Three Professional Master’s Program groups win Innovation Prize for project presentations
After presenting their projects to advisors, three teams in the Professional Master’s Program have received Innovation Prizes for their work. Two groups from the Big Data concentration were selected, as well as one group from Visual Computing.
“The Innovation prize winners were selected by the PMP Advisory Council Members based on a combination of three aspects: innovation, potential impact, and quality of implementation. The selection process was highly competitive,” says Jiannan Wang, director of the Professional Master’s Program.
“Congratulations to the three winning teams. The quality of their work is phenomenal. I am so proud of their achievements.”
Below is information on each of the selected projects:
Deviation Finder - An Elevator Anomaly Detection System
Github link: https://github.com/Ria-Thomas/DeviationFinder
This project idea was originally proposed by Technical Safety BC. After seeing the work of the students, they were impressed by the project, sharing their research on LinkedIn.
Abstract: Elevators have become an integral part of any building structure due to its ease of transporting people in minimal time. Elevators do not just help us get places a little faster – they make dense urban living possible, allowing the efficient construction of high-rise structures. Like every other machine, elevators also require periodic maintenance. Periodic health check-up of lifts is done by elevator technicians through routine inspections. Increasing development of new infrastructures has resulted in exponential installation of lifts demanding more inspectors in this area. Acquiring new workforce and training them can be expensive and time consuming. Hence, it is important to setup an automated process that can effectively find anomalies in elevators using Internet of Things and Artificial Intelligence.
The objective of this project was to design a system that can predict anomalies in elevators using accelerometer data. We have created a data science pipeline that incorporates methodologies such as data cleaning, data pre-processing, exploratory data analysis, data modelling and model deployment to meet the objective. We have experimented various unsupervised models such as K-Means, DBSCAN, Isolation Forest, LSTM, and ANN Auto-encoders to capture deviations in normal patterns. LSTM Auto-encoder outperformed other models with an F1-score of 67%. A demonstration of model deployment on web using Kafka, AWS Dynamo DB, Flask and Plotly Dash is also achieved as part of this project.”
The WikiPlugin: A new lens for viewing the world’s knowledge
No link available
Abstract: In this project, we used the open datasets released by Wikimedia to leverage both the underlying graphical nature of Wikipedia, as well as the semantic information encoded within each article's text using modern NLP techniques. We used that representation for each article to train a model to predict whether or not it would be a difficult concept to understand. Then we carefully designed an ETL pipeline to update a back-end system to support the model-scoring on a monthly basis. We have a database and web application supporting the home page of a Chrome extension, allowing the user to highlight the important concepts of an article, and to see the expected time required to read the whole page. They can find similar articles to the one that they’re trying to learn about, or analogous concepts in other subjects that weren't connected through links. We also built a simplification priority queue for all the articles that don't currently have an existing simplified version, based on the expected amount of time the article would take to read. This could be used in conjunction with the article click demand to have a bounty system to incentivize the articles most in need to having a simplified version next.
Hairstyle Transfer — Semantic Editing GAN Latent Code
Github link: https://github.com/Azmarie/Hairstyle-Transfer
Abstract: Motivated by the success of StyleGAN, where stochastic variation is incorporated in generating realistic-looking images, we proposed to focus on the hairstyle attributes of a face. The right hairstyle can often only be discovered through trials and errors. Thus, being able to virtually “try on” a novel hairstyle through a computer vision system holds practical value in reality. In this project, we propose an end-to-end workflow for editing hair attributes on real faces. Hairstyle Transfer leverages fixed pre-trained GAN models, GAN encoders, and manipulations of the latent code for the semantic editing. Moreover, we further confirmed the linear separability assumption of hair-related semantic attributes.