It’s times like this that I wish I saved more notes from freshman year. This semester, as part of my PDP in Operations Research Engineering (a combined program where you can work toward your master’s during undergrad), I’m taking ISE-633: Large-Scale Optimization and Machine Learning. The class deals with the math behind modern machine learning applications, many of which can be modeled as optimization problems, a central concept in ISE. For some context, here’s a simple example: when you perform linear regression, you are simply attempting to draw a line through a given set of datapoints that minimizes (optimizes) the total distance, or error, between the line that you draw and all of the given data. Essentially, you choose an intercept and a slope that will give you the best possible line for predicting values of future observations. This kind of optimization problem is relatively simple and can actually be broken down into a few simple matrix operations from linear algebra. However, for more complex machine learning models (e.g., neural networks) and extremely large datasets, it becomes much more difficult to compute a solution that truly minimizes the error between your model’s predictions and the data that you have observed.
That’s exactly what this class deals with: the advanced concepts from linear algebra and calculus that allow us to design algorithms that efficiently find parameters to minimize prediction errors. Now, I am obviously very fascinated by these ideas, but of course I am also a bit daunted by the math involved. One thing I love about the ISE undergraduate major is that it’s very broad and has classes that teach both hard skills and soft skills. But I’m also having a bit of a reality check these days in that I’ve spent the last two semesters largely taking more of the “soft skills” courses, meaning that my math chops are a bit rusty—not gone, but definitely in need of a tune-up. So as I dive back into ideas from linear algebra and calculus that I haven’t dealt with in quite some time, wish me luck. They’re both subjects that I love, but it’ll certainly take some work. Regardless, I’m really excited! And since freshman year, I’ve luckily become much more diligent about keeping records of all my class notes. That’ll come in handy.
