Click the thumbnail to watch our results summary!

The goal of this project is to not only create four working models, but to also learn the inner workings of a neural network, how they operate, and what can be manipulated to change a network’s effectiveness.

This project involved creating four neural networks using Python, PyTorch, and PyCharm. These four networks were created by us, as well as trained and tested by us. Packages used include scikit-learn and scikit-image, matplotlib, pandas, seaborn, and PyTorch. Also included use of anaconda for managing virtual environments.

Our pair had very limited prior experience with Python and PyTorch as well as machine learning. After an initial learning period, we were able to successfully create four models, alongside numerous testing and helper functions. The results showed what we hoped for, with all models learning properly and achieving reasonable accuracy results. With more time, it could be possible to further increase the accuracy of our models, however this was outside the time scope of this project. Overall it was very enjoyable and fascinating to get to work on machine learning. It has become such a buzzword, it can be difficult to understand what it truly is. This project gave me new insight into the limitations of current machine learning, and its applications in the future on more complex problems.