Skip to main content

Combining Tectonics and Machine Learning to Constrain Plate Reconstruction Models Through Time

Project description

Neural networks aid plate tectonics modelling

The constraints of the Earth’s lithosphere net rotation with respect to the underlying mantle cannot be determined from oceanic crust, as it is destroyed by plate tectonics. On the other hand, plate tectonics have a self-organising and statistically predictable nature that can be taught to neural networks. The TEMPO project will produce estimates for net rotation by training networks to use current data on the Earth, synthetic data, and the rules of physics. They will thus generate proposals of tectonic motion and mantle convection that will be tested against geological data, contributing to further research and the modelling of plate tectonics.


Net EU contribution
€ 196 707,84
45, Rue D'ulm
75230 Paris Cedex 05

See on map

Ile-de-France Ile-de-France Paris
Activity type
Higher or Secondary Education Establishments
Other funding
€ 0,00