I investigated and worked with various Model Interpretation methods under Professor Kurfess such as Integrated Gradients, it would be cool to use some of them on our model to see what exactly the model is learning and what image features can be attributed to its decisions.