Chapter 5. From Novice to Master Predictor: Maximizing Convolutional Neural Network Accuracy

In Chapter 1, we looked at the importance of responsible AI development. One of the aspects we discussed was the importance of robustness of our models. Users can trust what we build only if they can be assured that the AI they encounter on a day-to-day basis is accurate and reliable. Obviously, the context of the application matters a lot. It would be okay for a food classifier to misclassify pasta as bread on occasion. But it would be dangerous for a self-driving car to misclassify a pedestrian as a street lane. The main goal of this chapter is thus a rather important one—to build more accurate models.

In this chapter, you will develop an intuition for recognizing opportunities to improve your model’s accuracy the next time you begin training one. We first look at the tools that will ensure that you won’t be going in blind. After that, for a good chunk of this chapter, we take a very experimental approach by setting up a baseline, isolating individual parameters to tweak, and observing their effect on model performance and training speed. A lot of the code we use in this chapter is all aggregated in a single Jupyter Notebook, along with an actionable checklist with interactive examples. It is meant to be highly reusable should you choose to incorporate it in your next training script.

We explore several questions that tend to come up during model training:

• I am unsure whether to use ...