/ProbStats-with-TFP

Prob&Stats with TensorFlow Probability

Primary LanguageJupyter NotebookMIT LicenseMIT

Prob&Stats with TensorFlow Probability

GitHub contributors GitHub last commit GitHub

This project is conceptualised to supplement ISTD 50.034 (Introduction to Probability and Statistics) with some nice visualisations and code things to play with. It does not, and is not intended to, cover all the subject areas of 50.034.

This repository contains Jupyter Notebooks that are designed to run on Google Colaboratory, a free Jupyter environment that runs in the cloud. This means that there's nothing to install on your computer!

Most of the code here will utilise the TensorFlow Probability library. If you're viewing the Notebook on Colab, there's nothing you have to worry about since everything you need is already configured.

Contents

S/N Notebook Title Link
01 Introduction and Normal Distribution Open In Colab
02 Special Distributions Open In Colab
03 Bayesian Approach to Coin-flip Open In Colab
04 Sampling from Distributions Open In Colab
05 CLT and Approximating Distributions Open In Colab
06 WIP WIP

Do note that you'll get Colab's usual warning about running an untrusted notebook. That is expected behavior for opening a third-party notebook on Colab. Just click past that.

What's in the Notebooks

Most of the notebooks contain concise explanations of key concepts, as well as visualisations. You will be able to interact with the visualisations using sliders to see the effect of varying parameters.

slider_gif

Local Install

If you want to run the notebooks locally (not recommended if you don't already have a TensorFlow environment set up):

git clone https://github.com/OpenSUTD/ProbStats-with-TFP
cd ProbStats-with-TFP

# install dependencies
pip install tensorflow
pip install -r requirements.txt

# run Jupyter Notebook (or JupyterLab)
jupyter notebook

Contributing

Any relevant contributions are welcome! Feel free to open a pull request.

How To Contribute

You can fork this repository into your own GitHub account, then clone it into your Google Drive folder on your local computer. This way, you'll be able to push commits to your own fork, and submit a Pull Request to this repository.

Simple Requirements

  1. Only use "common" libraries
  2. Design to run on Google Colaboratory
  3. Make an effort to keep code simple and understandable

Core Maintainer

Timothy Liu / @tlkh