By Adam Czarnecki
What did your Stretch Experience involve?
I worked with Alpha Phenomics, a local company developing livestock monitoring technologies, for my Stretch Experience. The project that they were working on was a system to track individuals in a certain population and accurately predict if they have a virus based on various biometrics. The two main areas where they are hoping to implement this system are tracking employees in a workplace to help monitor COVID-19, and monitoring the health of dairy cows at a farm. I worked with the Chief Technical Officer and one of the head engineers to develop the part of the system responsible for recognizing individuals in the population being considered.
First, I spent some time researching how artificial intelligence (AI) works and how to develop neural networks. After implementing some neural networks in basic examples, I took on the challenge of implementing an AI to facial recognition. I was given a set of images of people for the AI to sort. I designed the AI to go through the list of images and group them together so that each group has images of the same person.
After discussing and implementing improvements to the accuracy of the program, I was asked to quantitatively compare my various sorting algorithms’ performance. I didn’t know how to do this, so I reached out to a friend who works in data analytics. He taught me that I can do random sampling throughout the databases that my algorithms created, and look manually to see how they compare. While this is a qualitative observation, it gave me a good understanding of how the AI was working and what factors influence its performance.
However, while making these comparisons, I realized these sorting algorithms were addressing a slightly different problem than what we actually wanted to solve. We wanted to categorize faces, not group faces together. I discussed this tangent that we went off on with the team and suggested we turn towards the right direction for our project, and how we can use what I observed in my algorithms. Afterwards, I used the AI to detect people’s faces from a live video feed and compare them with reference images in a database of known people that the program has stored.
Prior to this project, I only had one encounter working with AI and developing neural networks, and it wasn’t a good experience. It left me with the impression that it is something unattainable for me. I was nervous going into this project because I wasn’t sure if I would be able to make any progress, or if I would be able to come up with any useful contributions to the team. This project definitely felt out of my comfort zone. But with some help from the Alpha Phenomics team and a machine learning expert from AlphaML, and also some of my own online research, I was able to figure out the ropes in this field that was almost new to me. I quickly figured out what we needed to do to reach the goal of the project, and started coming up with ideas on how to execute those tasks.
My part of this project gave Alpha Phenomics a program that will be integral in the system for tracking a virus in a given population. When applied in work places, this system will help tremendously in contact tracing and containing outbreaks of a virus. This will be vital in our society now during the pandemic. The system will also be applied to cows at dairy farms, and potentially other livestock. This will aid farmers in tracking viruses in their livestock and keeping them healthy.
The leadership skills that I most developed during this project were my communication skills. As per my learning outcomes, I learned how to collaborate effectively with others and how to discuss ideas effectively. These will be very useful in almost any future collaborations with others. More specifically, I will be joining a research group at Quantum Silicon where I will be working on various new technologies in quantum computing. Being able to communicate effectively with others will be vital to my success in this work.
I also learned technical skills in artificial intelligence during this summer. I needed to apply a lot of background math knowledge that I had to learn these skills. Through this experience, I learned how it is possible to apply various theoretical concepts that I am learning during my undergraduate studies in the industry. While AI is only one application of these theoretical concepts, I found it very enjoyable to apply this background knowledge for practical purposes. This experience makes me more willing to consider a future career in the industry, but I think I still need more diversified experiences in other industries to make a better decision for my future career. These technical skills will also be very useful in my work at Quantum Silicon as they apply a lot of AI to detect defects in their products.
Adam Czarnecki is an undergraduate student in the Faculty of Science.