Intermediate Machine Learning (Python or R)

This ML workshop is build upon the introductory workshop and will cover Deep Learning (specifically ANN and CNN).

The workshop is prepared by:

• Daniel Marrable
• Rebecca Lange

from the Curtin Institute for Computation at Curtin University in Perth, Australia.

The notebook materials have been referenced and adapted from the Deep Learning with Python / R books.

Notebooks

R:

• this workshop
• additional notebooks from the Deep Learning with R book

Python

• this workshop
• additional notebooks from the Deep Learning with Python book

Cheatsheets

Python: https://s3.amazonaws.com/assets.datacamp.com/blog_assets/Keras_Cheat_Sheet_Python.pdf

R: https://www.rstudio.com/resources/cheatsheets/#keras

Data files

[cata and dogs] https://drive.google.com/open?id=1H5do852RlW7EdALFQOiP6hro7zB--AY9
[model files] https://drive.google.com/open?id=1cWi0hhl6c9CnQmw8nwYPwtiCfHLG2EA6

What to expect and what will be covered

Prerequisites

A working knowledge of Python or R and Jupyter notebooks is essential for this workshop. i.e. knowledge of basic data structures, operations and how to write scripts. The Python notebooks used throughout the workshop have been developed using Python 3.6.3 and the R notebooks have been developed using R version 3.5.1.

The workshop will use Keras, a high-level API for implementing neural networks. Keras is available for both R (more infor here) and Python (more info here). Please make sure you have it installed before the workshop. If help is required with installation do come along to the Curtin University Hacky Hour, every Wednesday 3-4pm at the Library Makerspace.

Other required packages for R:

• ggplot2
• gridExtra
• TTR

Other required packages for Python:

• numpy
• matplotlib

Please make sure you have everything installed and ready to go for the workshop!

What will be covered:

• Deep Learning Workflow:
  • data prep,
  • building a nework,
  • training the model, and
  • evaluation of the model
• Artificial Neural Networks
• Convolutional Neural Networks
• Transfer learning
• Hands on examples and exercises throughout

What will not be covered:

• Rigorous mathematical working and proofs.
  There simply isn't enough time in a workshop format.
  We will provide links and material that you can and should read.

This work is made available under the Creative Commons Attribution 4.0 International License.