Deep Learning with PyTorch - July 1st to 5th

Date:

July 1st to 5th. Morning, 9:00 AM to 12:00 PM

Classroom:

PC1

Instructors

I work as Senior Data Scientist in Research Team in Alpha (Telefonica). I have extensive experience in both industry and academic word. On one hand I was working as Data Scientist for Messaging Team in Schibsted (Fotocasa, Leboncoin, Infojobs, Segundamano, etc) helping building data driven and experimentation driven products like, for example, smart replies in chat. In SCRM (digital hub of Lidl) I was setting up data driven environment, building recommender systems and forecasting demand. In King using data I was helping build the best games. On the other hand, I was also working as associate professor doing research in Dynamical Systems and teaching graduate and undergraduate courses. Last few years I have been also given courses on machine learning, deep learning and data management.

Iñaki Estella Aguerri

Iñaki Estella Aguerri received the B.Sc. and M.Sc. degrees in telecommunication engineering from the Universitat Politècnica de Catalunya, Barcelona, Spain, in 2008 and 2011, respectively, and the Ph.D. degree from Imperial College London, London, U.K., in 2015. From 2008 to 2013, he was a Research Assistant with the Center Tecnològic de Telecomunicacions de Catalunya, Barcelona, Spain. He was a Visiting Scholar with Stanford University in 2012 and with the Biomedical Image Analysis Group, Eindhoven University of Technology, in 2008. From 2014 to 2018, he joined the Mathematical and Algorithmic Sciences Laboratory, France Research Center, Huawei Technologies Company, Ltd., Boulogne-Billancourt, France. Since September 2018 he is with the Health moonshot of Telefónica Innovación Alpha, Barcelona, Spain.
His primary interests are in the intersection of the fields of statistics, optimization and information theory and their applications to machine learning, signal processing and communications systems.

Language

English

Description

Why is it worth to learn Deep Learning? Because it is very effective in tasks like: understanding images, understanding natural languages, translations, driving cars and more.

Why is it worth to know pytorch? Pytorch is a machine/deep learning library for python mainly developed by Facebooks's AI Research Group. Since pytorch's models are just like other python programs, it quite easy to write and debug them. Hence pytorch is great for research and experiments. Moreover, with new version 1.0, pytorch introduce "script mode" that aims to become also a tool for writing production code.

In this workshop we give an introduction to deep learning using pytorch. You will have an opportunity to build, train and evaluate few models. We start with a simple single layer neural networks, go though convolutional ones and finish with recurrent neural network. We will train algorithms to predict sales, recognise objects on images and classify texts.

Course goals

Learn what is Deep Learning and its main types of networks: Fully Connected Neural Network, Convolutional Neural Network, Recurrent Neural Network. Being able to construct and test neural network.

Course contents

What Deep Learning is and why we need it
Introduction to pytorch.
Forward and backward propagation.
Fully Connected Neural Networks
Convolutional Neural Network
Recurrent Neural Network

Prerequisites

Good knowledge of python and previous exposure to Machine Learning. Familiarity with github would be helpful.

Targeted at

Everyone interested in practical machine learning.

Evaluation

Project with a model trained, tested, served.

Computer class or student's laptop?

Student's laptop

Software requirements

Python 3.6 with numpy, pandas, pytorch. Instructions will be provided