Enough Keras for students to put on their resume, use to supplement capstone projects, and provide starter scaffolding/kindling further study.
How ANNs function: Keras workflow, layers, representations
- Concepts, Vocabulary, and Anatomy
- Interpretability, caution, and ethics
- Just enough detail on how things work, not all the math or a deep dive on the history or cutting edge
- I do: Build an even/odd number classifier. Identify the components, explain how each piece works.
- We do: Build out FizzBuzz in Keras
MNIST Digits and Fashion datasets with a CNN
- I do: MNIST handwritten digit classifier
- Connect to the big picture
- Explore and visualize
- You do: Fashion image classifier
- https://keras.io/examples/vision/image_classification_from_scratch/
- CIFAR10 small images classifier
- Preprocessing resources. How to turn images/text/numbers into tensors/numpy arrays
- Further datasets for practice:
- Classify movie review sentiment
- House price predictor using regression
- Classify newswire topics (multiclass)
Cats and Dogs Dataset https://www.microsoft.com/en-us/download/details.aspx?id=54765
- Madeleine's slide presentation
- Dedicated reading time of handout.pdf
- Guided demo of: