About
Built for geoscientists who use machine learning
GeoXMate started from a simple need: to take geological and geophysical data from loading to a validated prediction without moving between separate programs and scripts, and without writing code. It is developed by Khaled Saleh.
Built around the machine-learning workflow
Every tool serves a stage of the path from raw data to a validated model and its application, in wells, seismic volumes and geological models.
Validation that holds in the next well
Models are judged on whole wells they never saw, by blind wells and leave-one-well-out, and never on randomly mixed rows.
The geoscientist stays in charge
Machine learning proposes; the interpreter checks it on sections, maps and logs, and keeps the final say.
Data acknowledgement
Volve field data
The screenshots on this site show GeoXMate working on the Volve field dataset. Equinor and the former Volve licence partners released the data under the Equinor Open Data Licence, and we gratefully acknowledge them for making it available for research and study.