Saturday 12:45–14:45 in Tutorial Track 1

Straightforward introduction to Deep Learning

Mikołaj Olszewski

Audience level:
Novice

Description

This practice oriented workshop is dedicated for everyone who would like to learn the basics of Deep Learning. We’ll create a simple Deep Learning model predicting the rent price of an apartment based on it’ characteristics using the Keras framework.

Abstract

Methods based on Deep Neural Networks (DNN), called “Deep Learning”, are recently gaining more and more popularity in the Data Science industry.

They might seem overwhelming for those who never tried to use them. However, the truth is, that the basic theory behind is not that hard to understand. You might be surprised that the simplest neural network is just a linear function. You might be also surprised that learning how to create a Deep Learning model does not require a Phd. in math related subject.

With recent frameworks like TensorFlow and Keras also the effort needed to build a deep neural network is similar to the effort needed to build any other Machine Learning model.

On this practice oriented workshop you will learn the very basics of how deep neural networks work and you will implement a simple one by yourself using the Keras[1] framework. We’ll be working on a very simple, yet real world example to showcase the mechanics that powers Deep Learning. We’ll be trying to predict the rent price of an apartment based on it’ characteristics using the publicly available data originating from AirBnb[2].

Google account is needed as we will be using Google Colab for exercises.

[1] https://keras.io [2] http://insideairbnb.com

Subscribe to Receive PyData Updates

Subscribe