gnaixuj / til-ai-camp

TIL AI Camp - Upskilling Content

Description

This repository contains 3 folders that will be used during the upskilling workshop for TIL AI Camp 2019.

Participants can refer to the folder1-mnist and 2-cifar10 to familiarise themselves with the basics of machine learning where we make use of convolutional neural network (CNN) to identify digits from the MNIST handwritten digits dataset as well as to perform object classification using the CIFAR10 dataset. In both notebooks, participants are expected to fill in the missing fields before running the cells.

The solution folder contains the solutions to both notebooks, with the missing fields populated.

Folder structure

1-mnist
|- 1_MNIST_CNN.ipynb

2-cifar10
|- 2_Image_Classification_CIFAR10.ipynb

solution
|- 1_MNIST_CNN_solution.ipynb
|- 2_Image_Classification_CIFAR10_solution.ipynb

Links

• AWS Educate - http://awseducate.com
• AWS Educate Login Page - https://www.awseducate.com/signin/SiteLogin
• Kahoot - https://kahoot.it/
• TensorFlow Playground -https://playground.tensorflow.org/
• 3D Visualisation - Solving MNIST using CNN - http://scs.ryerson.ca/~aharley/vis/conv/
• Brainhack - Compilation of resources for getting started into the world of AI - https://github.com/brainhack-dsta/Brainhack

More Links

There are many things you can try to improve on this baseline. Here's a short and non-exhaustive list of tricks that deep learning practitioners are normally up to. Not all are free, but it's all great material:

• https://keras.io/
• https://blog.keras.io/building-powerful-image-classification-models-using-very-little-data.html
• https://stanford.edu/~shervine/teaching/cs-230/cheatsheet-deep-learning-tips-and-tricks
• https://www.deeplearningbook.org/contents/regularization.html
• https://www.amazon.com/Deep-Learning-Python-Francois-Chollet/dp/1617294438