My experience at DIYA was an extremely valuable learning opportunity and exposure to an area of computer science I hadn’t explored before.
I began the internship with a moderate degree of experience in computer science, having worked in java, html, javascript, and python before, but I had no understanding of Data Science or Machine Learning and only knew the terms as buzzwords. At DIYA, I was able to get an interactive introduction to Machine Learning that allowed us to apply concepts to real programs as we learned about them.
The process of going through the stages of research from exploratory analysis to exploring different ways to train the model to analyzing results was an extremely useful learning experience that allowed us to learn through doing. When we actually began to use and tweak our models. I understood much more about how they worked. Filling out the project proposal plans and templates were particularly useful in guiding me into understanding what to manipulate and what to look for. Through this guided experience, not only did I learn basic fundamentals of machine learning models and how they worked, but also got to see them in direct application through my project. The guidance of the mentors and TA throughout the process and ability to interact and compare results with peers was transformative in helping me truly understand the underlying concepts of what we were doing, and why we saw different results. Exchanging observations with others and receiving mentor feedback was very useful because it allowed me to see the influence of variables and perspectives that I hadn’t even begun to consider.
Having a diverse group of peers added another dimension of learning as we each tried to answer different questions with the same data, and were able to learn from both the similarities and differences in our approaches. Since we all used the same decision tree and random forest machine learning models, we could compare our results to share what trends to look for in our models and understand if we were heading in the right direction, and were able to discover unexpected issues like a data imbalance which we all could see independently. At the same time, we could learn from each others’ successes – seeing what elements of each others’ models and approaches we might want to incorporate for next time.
As a whole, DIYA was a great introduction to machine learning for me and fostered an new interest in an aspect of computer science that I once thought was out of my reach. Seeing how elements of mathematics, statistics, and scientific analysis all came together during my experience at DIYA makes me eager to explore more Data Science in the future.
