Bayesian Inference and Neural Networks in Formation Evaluation (PP12)


DATE	LOCATION	FEES		REGISTRATION


Dec 13-17, 2010	Vienna, Austria		Course Fee: EUR 2475 plus VAT Computer Fee: EUR 300 plus VAT

Instructor

Rudolf K. Fruhwirth

Course Level: Intermediate

This course will introduce the participants to the basic methods and techniques of modern Bayesian inference and Neural Network technology and their application in formation evaluation and reservoir characterisation. Bayesian inference allows dealing with complex problems in a conceptually simple and unified way. Neural Networks are an advanced tool for data mining due to their capability to learn and understand complicated data relationships. Typical applications will be covered including (missing) data generation, data QC, outlier detection, and model homogenisation. cVision^TM, a fully automatically working general purpose neural network software, will be used throughout the course and will be provided for all participants along with a 3 months evaluation license.

WHAT YOU WILL LEARN

Principles of Bayesian statistics

Principles of Neural Networks

Relationship between Neural Networks and Bayesian inference

Principles of classification

Use of Bayesian statistics and Neural Networks

OUTLINE

Bayesian statistics
– Statistical distributions
– Fundamentals of Bayesian statistics
– Bayesian rules
– Bayesian classification
– Prior and posterior probabilities
– Examples

Principles of Neural Networks technology
– Overview of artificial intelligence
– Artificial neurons
– Learning and the technology behind
– Relationship between Bayesian statistics and Neural Networks
– Efficient training neural networks
– Avoiding pitfalls
– Examples

Features
– Principles of feature extraction
– Feature selection by sequential search techniques

Use of the technology in the oil & gas industry
– Facies classification Facies generation Facies QC determination of outliers, internally consistent models
– Log calibration QC, outliers, consistent models
– Generation of missing logs shear wave logs porosity logs permeability logs

WHO SHOULD ATTEND
Petrophysicists, geophysicists, geologists, engineers and other asset team members who want to understand Bayesian statistics and Neural Networks and how to apply theses novel technologies to their particular problems in formation evaluation, seismic processing/interpretation and reservoir characterisation. A basic understanding of Neural Networks is preferential.

COURSE VENUE:

Vienna
Renaissance Vienna Hotel
Linke Wienzeile/Ullmannstr. 71
A-1150 Vienna
AUSTRIA

INSTRUCTOR

Rudolf K. Fruhwirth