Rangle x The Knowledge Society: How I Started My Journey Into Machine Learning
The Knowledge Society (TKS) is an incubator for students to explore innovative topics with like-minded people. TKS was designed to prepare young people ages 13-17 to become leaders in their industries by encouraging them to take risks and think big. At Rangle we consider ourselves to be like-minded in that we are also constantly thinking about what’s new and what’s next. That’s why we’ve collaborated with some of these bright minds to get their take on some of the biggest things happening in tech and development right now.
With that being said, the following blog post is a feature written by TKS student Seyone Chithrananda on how and why he became so interested in Artificial Intelligence and Machine Learning.
For most high schoolers, our world revolves around school, sports, friends, and extracurriculars. But thanks to the internet and faster chains of communication, knowledge is being distilled at exponential rates and has lowered the bar of entry into many fields. With easy access to all of this information, I began to learn more about the field of genomics and machine learning about a year ago and realized there are so many opportunities for me to learn beyond school.
Biologist Eric Lander, who helped lead the effort to sequence the full human genome back in 2003, famously said, “Genome: Bought the book; hard to read.” My interest in genomics led me to write an essay on modern therapeutics in the field of ocular genetics a few articles surrounding CRISPR-CAS9, a tool to help us edit our genome and its functionality, as well as current problems with the tool such as unwanted deletions. I first started learning about genetics and biotechnology while working on the essay on ocular genetics, and I was fascinated by the genome and the fact that diseases like spinal muscular atrophy were caused by one simple mutation in the SMN1 gene. Typically, students like me rarely have the opportunity to learn about machine learning, genomics and the other technologies I’ve become interested in. I was searching for a way to learn more beyond what I had access to online.
Thankfully, a few months ago I joined the Knowledge Society (TKS), a youth accelerator that aims to expose youth to emerging technologies such as AI, Quantum Computing, Genomics, and Blockchain. It was shortly after joining TKS that I realized I specifically wanted to focus on machine learning after reading into its applications in genome biology for things such as drug discovery and recognizing gene expression. During this period of time, I not only read a variety of papers on different model implementations, I also began to learn from role models in the field. Some of my influences include Brendan Frey who leads Deep Genomics, a company working on creating genetic medicine using deep learning, as well as Geoffrey Hinton, who many call the Godfather of AI and is one of the pioneers of neural networks and backpropagation.
Living in Toronto, I’m privileged that my city is rapidly becoming a hotbed for AI. Because of this, I have even been able to meet and talk with a handful of remarkable researchers over the last couple months at conferences such as Re-Work and the Toronto Machine Learning Summit. Recently, I decided to focus on a difficult to solve problem - how to develop blood based tests for cancer detection. In approaching this issue I focused on CNNs and developing models for different problems surrounding genomics. We don’t have many techniques for using blood-based tests for cancer detection right now, but with deep learning we could potentially use liquid biopsies to classify cancer by stage. What interests me most about machine learning in genomics is the sheer amount of data - each genome is roughly 200 GB! That being said, my goal going further is to develop models in the ML space surrounding genomics and have the opportunity to work in a lab and do research in this exciting field.
What get's me the most excited about the future of machine learning and computational biology is the possibility of finding a solution to interpreting DNA. For example, through algorithms and models, we can now understand the living, breathing genome in motion. As more data becomes available as a result of companies like 23andMe and Illumina, we have the opportunity to accelerate drug discovery through molecule-target bonding or identification of novel biomarkers. We can even potentially classify and categorize cancer using cell-free DNA. The fact that the two fields of genomics and machine learning are rapidly growing at the same time really excites me about the future applications that will be developed!
Although there are limitless opportunities in the ML space, not many of my peers know about these opportunities or where to even begin. If we can get more youth exposed by bringing in mentors from the industry to high schools and other educational institutions, we can help spur growth in the number of aspiring developers. My hope for the future is that more people my age can realize the possibilities in the field and how AI is going to positively disrupt so many industries.
You can check out more of Seyone Chithrananda's Machine Learning insights on his personal Medium, here.