Aviva2019
Notes for Aviva workshop. http://bit.ly/aviva19
Day 1 (10th Dec)
1pm Lunch
2pm Introduction: Deep Learning and AI: What it is, what it can, and cannot do. (JVS) slides
3pm Overview: From linear networks to convolutional networks. (SJE) slides
4pm Coffee
4:30pm Images (continued SJE) / Introduction to dimensionality reduction (LDC) / review of mathematical background
5:15pm end
5:30pm check in Selwyn
7pm Dinner at Caius college (6:30pm for drinks)
Day 2 (11th Dec)
10am Backprop: how it works (and how it fails). (JVS) slides
11am Dimensionality Reduction (LDC). slides
12 Lunch
1pm Using Python for networks:
- No hidden layer Practical/xor.py.
- One hidden layer bp/main.py.
- MNIST in Python / MNIST in Keras
See also https://playground.tensorflow.org
3pm-3:30pm Wrap up / discussion of next steps.
References and further reading
The core text we will refer to is AI Engines, of which you will receive a copy.
Most core papers that we mention should be available through our Paperpile collection. Click on the "View PDF" link below each reference to get direct access to the paper. These PDFs are provided only for educational use.
A book full of practical implementations in Python (and R) is by Chollet. Finally, slightly old now but still regarded as the "bible" and freely available online is Goodfellow et al.