arxiv code search
2022Searching through arxiv papers (with ML) to see if they include the code and data to reproduce the work. Active research.
Always building. Always learning. Here are some of the projects I have been working on:
Searching through arxiv papers (with ML) to see if they include the code and data to reproduce the work. Active research.
External knowledge can enhance machine learning. We use knowledge from reliability engineering and integrate it into a machine learner through the use of a Weibull-based loss function. Demonstrated on bearing remaining-useful-life prediction.
Anomaly detection on the UC Berkeley milling data set using a disentangled-variational-autoencoder (β-VAE). Published in the International Journal of Hydromechatronics.
Machinery data, made easy. Open-source Python package to easily download and prepare common PHM (prognostics and health management) datasets. Use PyPHM before feature engineering or model training.
Using generative adversarial networks (GANs) for modeling of time series. Applied to applications in manufacturing systems. Active research.
Developing scalable ETL/ML pipeline for rapid testing of feature engineering and machine learning techniques. For use on industrial time series data. Leverages HPC or cloud infrastructure.
Personal project exploring the CDC birth data files from 1968 to 2020. Developed for HPC and local compute.
A collection of beautiful plots and other data visualization explorations.
New to Compute Canada and high performance computing? Here are some tutorials to get you started.
Demonstrating feature engineering and classical machine learning for use on tool wear monitoring. Applied to industrial partner's manufacturing environment. Discussed in thesis, "Feature Engineering and End-to-End Deep Learning in Tool Wear Monitoring".
Talk on real-world anomaly detection methods within health care, astronomy, finance, and manufacturing. Presented at Pycon Canada 2019.
Presentation on AI within the public health domain. Talk given to medical residents and MOH's at KFL&A Public Health.
Final project for Deep Learning course (CISC-867). Used a convolutional autoencoder to detect faults in medium voltage power lines. Received an A+ in the course!