You are on page 1of 6

Machine Learning


For Geomodeling

Hervé Gross, PhD - Reservoir Engineer


Advanced Resources and Risk Technology
hgross@ar2tech.com

© ADVANCED RESOURCES AND RISK TECHNOLOGY This document can only be distributed within UFRGS
Class objectives and outline

GOAL
Provide the essential skills to understand and apply modern data analysis tools
(machine learning) to problems commonly found in geomodeling

OUTLINE
o Dimensionality reduction
o Unsupervised learning: Clustering
o Supervised Learning: Classification and regression
o Data visualization
o Connection to AR2GEMS

PRE-REQUISITE
o Basic of programming
o A basic Python primer will be covered to ensure that everyone has the minimal
skills required to perform exercises
o Based on Scikit-learn (state-of-the-art Machine Learning in Python)

© ADVANCED RESOURCES AND RISK TECHNOLOGY 2


Organization

Day 1 Day 2 Day 3 Day 4

Student presentations
9:00 - 12:00 Intro to Python Unsupervised Learning Supervised Learning
and individual requests

12:00 - 13:00 LUNCH

Dimensionality Unsupervised Learning Supervised Learning


13:00 - 16:00 Extra time
Reduction Working Session Working Session

The class is designed to be hands-on


Exercises will be done in groups of 2-3 students
Each group will submit a report with the question answers and their source code
The same data set will be used for all exercises

© ADVANCED RESOURCES AND RISK TECHNOLOGY 3


Class Introductions

Name, department, level


Motivation for taking the class
Python expertise level (0: never done any, 5: Jedi master)
Prior experience with machine learning

© ADVANCED RESOURCES AND RISK TECHNOLOGY 4


Ex. 1: Try machine learning!

Point your web browser to http://playground.tensorflow.org

© ADVANCED RESOURCES AND RISK TECHNOLOGY 5


References

http://playground.tensorflow.org

© ADVANCED RESOURCES AND RISK TECHNOLOGY 6

You might also like