Kaggle Higgs Boson Machine Learning Challenge

Undergraduate course project, STAT 406, 2021

  • Higgs Boson Machine Learning Challenge is a well-known Kaggle competition, the objective is to find the most accurate method to classify this dataset into the classes “tau tau decay of a Higgs boson” versus “background”.
  • Our team is currently evaluating classification models such as PCA, random forest, logistic regression and are going to utilize ensemble methods with parallel computing.
  • The link to the project will be provided soon.

VAE Lecture Project

Undergraduate course project, CPSC 440, 2021

  • Created a lecture on Variational Autoencoders which includes Variational Inference and KL-divergence derivation as well as the transition from Autoencoders and KL- divergence to VAE, we showcased and implemented Keras versions of Autoencoder, VAE as well as their deep variations and tested our implementation on MNIST, we also created assignment questions regarding KL-divergence derivation and its implementation.
  • GitHub link to project

Kaggle Autonomous Driving Prediction Challenge

Undergraduate course project, CPSC 340, 2020

  • To create an ego-centric predictive model of vehicle motion conditioned on past positions of both the ego vehicle and the motion of other agents moving around the same intersection near the same place in space and time, our group utilized two MLP models to train the x and y coordinates of the ego vehicle, we used the vehicles current positions as well as the coordinates of the neighbouring objects to predict the future positions of the ego vehicle.
  • GitHub link will be added soon.

COVID-19 Canada Daily Deaths Forecast

Undergraduate course project, CPSC 340, 2020

  • Predict the number of deaths in Canada caused by COVID-19 per day over a certain number of days using a Linear Auto-regressive model and boxcar smoothing. Predicted cases from 10/6/2020 -> 10/16/2020 for model selection and forecasted 10/26/2020 -> 10/30/2020.
  • GitHub link will be added soon.

InsightUBC Web Application

Undergraduate course project, CPSC 310, 2020

  • A TypeScript full stack web development project with Node.js as server-side development, the project as a RESTful web server aims to enable effective querying of the metadata from around campus, including course info and past averages.
  • GitHub link to project