From data to knowledge
June 28 to 30, 2017
Notre offre de cours s’étoffe constamment
Aussi susceptible de vous intéresser :
Professionals who want to develop the data analysis skills required to extract meaningful information from raw datasets. The course will be given in English.
Basic knowledge of linear algebra and of programming (the course uses Python) is required.
In 2012, the Harvard Business School called data scientist the sexiest job of the 21st century, and in 2015 McKinsey projected that “by 2018, the U.S. alone may face a 50-60 percent gap between supply and requisite demand of deep analytic talent“. Why has data scientist become a most in-demand job ?
Today massive amounts of data are available in all areas of science, government and industry. Exploited sensibly, these raw data can significantly improve the efficiency of research, services and industries in as many fields as healthcare, engineering, finance, telecommunications or urban developments just to name a few.
How are powerful companies like Google, Facebook, IBM or Apple using data analysis techniques ? What are the most important algorithms to know and how to apply them to your projects ?
- Learn the top algorithms used in data analysis
- Understand how leading companies are using data analysis techniques
- Be able to apply data analysis algorithms to real-world datasets
- Introduction to Data Science :
What is data science ? Examples of data (text, image, sound) and networks (Internet, Facebook, transportation). Data pre-processing (cleaning, normalization).
- Introduction to Python :
Python and its large computing ecosystem for data science. Application to TensorFlow, the Google AI software for deep learning.
- Unsupervised Learning :
Data clustering with K-means (text clustering). Network partitioning with normalized cuts. Community detection with Louvain and Nibble. Spectral network analysis (brain functionality analysis).
- Supervised Learning :
Classification with support vector machine (document classification). Artificial neural networks (deep learning) with application to image classification.
- Recommender Systems :
PageRank for web ranking (Google search engine), collaborative filtering (YouTube, Netflix) and content filtering (Amazon) for data recommendation. Application to movie recommendation.
- Feature Extraction :
Standard and sparse PCA (gene analysis), Lasso regression techniques (visual human analysis), matrix factorization (music analysis).
- Visualization :
Robust PCA (faces), LapEigenmap model (social networks), t-SNE technique (music visualization).
- Big Data Optimization :
Online and distributed optimizations. Scalable computing platforms (Hadoop MapReduce and Dato GraphLab) for massive datasets.
- Good balance of theory and practice (through coding demonstrations and case studies)
- Each lecture will be followed by practical exercices
A certificate of attendance will be delivered at the end of the course.
- Ecole Polytechnique Fédérale de Lausanne (EPFL), School of Engineering (STI), Signal Processing Laboratory 2 (LTS2)
- Prof. Pierre Vandergheynst
Professor and head of the Signal Processing Laboratory 2, EPFL
- Prof. Xavier Bresson
Professor of Data Science, NTU, Singapore
Dates and schedule
- Wednesday, June 28, 2017 – from 8:30 am to 5:30 pm
- Thursday, June 29, 2017 – from 8:30 am to 5:30 pm
- Friday, June 30, 2017 – from 8:30 am to 5:30 pm
EPFL, Lausanne, Switzerland
Course fee *
1’900.- Swiss Francs** (includes course material, lunches and refreshments)
* based on the price of the last edition
** 10% special discount for contributing members of EPFL Alumni
Participants should bring their own laptop to use during this session.
to be determined