Introduction to Machine learning (ML) for Geophysics

Trainer(s): Jaap Mondt
Duration: one day (F2F) or one month (E-Learning)

Business context

More and more Machine Learning will play a role not only in society in general but also in the geosciences. Machine Learning resorts under the overall heading of Artificial Intelligence. In this domain often the word "Algorithms" is used to indicate that computer algorithms are used to obtain results. This is then associated with the idea that it is all black-box work and not fully understood by those who are applying it, let alone those confronted with decisions based on these algorithms. Also, "Big Data" is often mentioned, indicating that these algorithms need an enormous amount of input data to produce useful results.

Many scientists mention "Let the data speak for itself" when referring to machine learning, indicating that hidden or latent relationships between observations and classes of (desired) outcomes can be derived using these algorithms. This is obviously very useful when there is no clear physics (equations) to describe that relationship. A clear example is in the field of Quantitative Interpretation. For clastics we have a reasonable understanding in which cases known rock properties expressed in equations can be used to predict say pore fluids. But for carbonates it is often an enigma and we have to resort to statistical relationships. Then Machine Learning enters into the game. If we have many wells with known drilling results, the algorithms can derive non-linear relationships between seismic observations and the known well results (supervised learning). But sometimes it is already useful if an algorithm can define separate classes (say seismic facies), which then can be more easily interpreted (unsupervised learning).

The ultimate case is where the data speaks for itself under the constraint that the relationship should still to some degree satisfy a physics equation in addition.

Remains the questions "How can these algorithms learn?" This is discussed in the course by explaining so-called "Backward propagation". 

Who should attend

All those interested in understanding the impact Machine Learning will have on the Geosciences and then specifically the impact on geophysical processing and interpretation. Hence, geologists, geophysicists and petroleum and reservoir engineers, involved in exploration and development of hydrocarbon fields, but also those working in shallow-surface geophysics.

Course content

09:00-09:15        Welcome, Program, (1) Biography, (2) Intro ML

09:15-09:30        (3) ML Tutorial, (4) ML Open Source Software, (5) Weka

09:30-10:00        Exercises 1-2

10:00-11:15        Refreshments

10:15-11:00        (6) DNN, (7) ML First Arrival Picking, (8) ML Trace Interpolation

11:00-12:00        Exercises 3-5

12:00-13:00        (9) Activation Functions, (10) Forward and Backward Propagation

13:00-14:00        Lunch

14:00-14:30        Videos: Geophysical Inversion versus ML, Deeplizard

14:30-15:00        Exercises: 6-8

15:00-15:15        Refreshments

15:15-15:45        (11) ML Fluid Substitution

15:45-16:45        Exercises 9-11

16:45-17:00        (12) Future of ML in Geophysics

Learning, methods and tools

At the end of the course participants will have an idea how Machine learning, being part of Artificial Intelligence will impact the future of Geosciences. This will be evident from the few examples shown in processing and (quantitative) interpretation. The course uses a mixture of lectures, practical exercises and direct (workshop-like) participant involvement in discussions.

Use of laptops for exercises and WIFI internet access in the classroom is mandatory.

The course can be customized to meet specific needs of participants.