“Group project time!”

This is a phrase that always incites fear in me. And I’m not going to lie, I was nervous when I realized that my Capstone project was going to be a group project. I was worried that there would be group members not pulling their weight and that I would end up having to do the work of multiple people (which has happened many times before). For a small project it’s manageable because it just means a couple extra hours of work. But for a big project… a project that is already going to take about 100 hours (just for MY portion), I was concerned. So I’m glad to report that so far, I have been pleasantly surprised. I’m not sure if it’s because people take group work more seriously when the entire course is just the one group project, or if I just lucked out having a great group. Everyone is responsive and is pulling their weight. I chose a topic that has not been taught previously in any of my courses because I have a strong interest in it and want to learn as much as possible from this project. Everyone else seems to be in the same boat as me so we ended up being a group of people all wanting to do as much as possible.

All that being said, the process has not been easy. While I have a great team working with me, we all have a ton to learn. I understood the very basics of neural networks before starting this project but had never looked into all the intricate details. And wow, there is a lot of math involved.

Luckily, we can use Tensorflow and Keras to hide a lot of that math, but to actually understand what’s going on (which is my goal), I want to dive deeper. One of the best resources I’ve found so far is this video where a neural network is built from scratch without using the usual libraries. It’s math-heavy but I would recommend it to anyone wanting to learn EXACTLY what is happening.

I’m looking forward to the coming weeks and I’m excited to see the final product. There’s lots of work ahead but this is going to be something that I’m proud to have made, so bring it on!