To find a tailor-made solution for the customers’ problems, Geosaic tries to use know-how from AI research projects and knowledge from previous projects.  

​

Which architecture and data?

​

A plan and an architecture are designed and individually adapted for each problem. Data can be 1D, 2D, 3D, but also multidimensional. Eventually, almost all geoscientific data can be processed.

​

Which algorithms?

​

Both supervised and unsupervised algorithms are used for this, depending on the problem. Regarding unsupervised algorithms, clustering algorithms are mostly experienced, regarding supervised algorithms, it is CNN algorithms . However, new algorithms are constantly being evaluated and developed.

​

Data preparation?

​

One of the most important steps is data preparation which can be done by the customer (with instructions) or by Geosaic. This step can be very time-consuming and, in individual cases, even account for more than 50% of the project volume.

 

Where does the data storage take place?

​

If desired, data can be stored on local hard drives or on cloud servers in Austria or Germany.

​

Is AI useful for my data?

​

Free consultation is useful to answer this question. It may always be the case that AI is not the right solution for you. However, in most cases it is. Therefore, free consultation is definitely recommended beforehand.