Convolutional neural networks are a class of machine learning algorithms that have proven rather powerful in the analysis of images. Relying on principles inspired by our own visual system, they capitalize on a combination of filters that learn the spatial correlation structure of the training data, and a hierarchical organization that allows a gradual transformation of the information from the input into a representation that lends itself more readily to interpretation.
In this lesson, we will learn first about the basic principles underlying learning by neural networks, and then focus in on convolutional neural networks with a particular application in image classification.
The lesson template used for creating this lesson is based on the lesson template used in Data Carpentry and Software Carpentry workshops.
| 09:00 | Artificial neural networks |
Why use artifical neural networks?
What is the mechanism of learning in neural networks? |
| 09:15 | Keras |
How do we classify images with a neural network?
How can we build neural networks with Keras? How are models fit and evaluated in Keras? |
| 09:25 | Convolutional networks | Break |
| 09:40 | Deep CNNs |
How can we make CNNs more efficient?
What can we do to intepret CNN results? |
| 09:55 | Wrap up | What are things we should add to |
| 10:10 | Finish |