An introduction to Convolutional Neural Networks

Jonathan Fernandes | Sunday 10:30 | Room I

Using a hands-on based approach with Jupyter notebooks, attendees will go from beginner level to covering intermediate-level content in 3 hours.

Section 1 In this first section, attendees will learn what convolutions are, and use convolutions in 1-D and 2-D. We will look at image filters such as Gaussian blurs and edge detection.

Section 2 Now that the participants have had a gentle introduction, we will introduce concepts such as pooling layers, convolutional neural networks (CNN) and weights in CNN. This segment will be peppered with exercises for participants to practice and confirm their learning.

Section 3 In this final section, we will look at some more advanced concepts and applications. We will take a look at numerous applications off CNNs including the StreetView House Number dataset and Facial expression recognition.