Add neural network tutorial for defect prediction

[EvanMendenhall]

Hello! I added a small tutorial folder that builds on the existing defect prediction material.

The goal is to give other students in the workshop another way to run:

1. Linear regression
2. Add a new learning algorithm of: Gradient Boosted Trees, to see how it compares.
3. Offers a short 20 Epoch local pre-training tutorial on a 92M paramenter neural network that can be completed in a few minutes. This is a runnable example of training models on the defect prediction data which will add 3-4% accuracy depending on training runs.

What this adds:

• setup and run instructions
• data download/preprocessing code
• baseline model examples
• a neural network training example
• model evaluation output focused on accuracy and related metrics

I tested this inside a fresh copy of the repo with this folder added, and the instructions worked for me. I enjoyed your class, and think this extra 10 minute walk through would be an amazing addition.

Students can leave saying: "I built my own neural network at the Coveros AI for Testers training workshop!" I personally had fun, so I wanted to give this to your course. Hope you enjoy it!

This is meant to support the existing defect prediction notebook, not replace it. I’m very open to changing the folder name, structure, or approach if there’s a better way you’d prefer this to fit into the course materials. It should be fully self contained as a 1, 2, 3, step extra exercise you can use.