CONVOLUTIONAL NEURAL NETWORKS FOR GEOPHYSICAL APPLICATIONS
(Frontal teaching)
- Language: ENGLISH
- Campus: SPAZIO ESTERNO AL POLIMI
- Enrollment: 21-04-2020to hour 12:00 on
08-05-2020
Application completed, activity in evaluation
Rate this course
Unrate this course
Do you want this course to be offered again?62 people have already voted
- Teacher in charge
- BESTAGINI PAOLO
- Credits
- 2
- Hours to attend
- 20
- Max. number of students
- 230
Description of the initiative
Non-destructive exploration of the subsurface makes use of elastic and/or electromagnetic waves that propagate and collect information about the crossed medium. The processing and interpretation of these "geophysical" data is carried out with a multidisciplinary approach and it allows to extract different kinds of information about the scenarios under investigation. However, the new challenges of geophysical imaging applications ask for new methodologies going beyond the standard and well established techniques. In this course, we will focus on geophysical imaging problems that has been recently faced using convolutional neural networks. First, we focus on the problem of seismic data interpolation through the use of convolutional autoencoders. Then, we will show how generative adversarial networks can be used to guide geophysical data interpretation.The students will learn the theoretical concepts behind the proposed techniques, and will implement some simple architectures using Python.
Duration
dal May 2020 a July 2020
Calendar
Mercoledì dalle 18 alle 20
Maggio: 13, 20, 27
Giugno: 3, 10, 17, 24
Luglio: 1, 8