Report describing the result of the machine learning benchmark carried out during the WP DONUT: EURAD-DONUT Deliverable 4.8

Due to recent technological developments, the fields of artificial intelligence and machine learning methods (ML) are growing at a very fast pace. The DONUT scientific community has recently started using ML for a) accelerating numerical simulations, b) multiscale and multiphysics couplings, c) uncertainty quantification and sensitivity analysis. There are first evidences, which suggest an overall acceleration of calculations between one to four orders of magnitude. Within DONUT a benchmark was designed to coordinate activities and test a variety of ML techniques relevant to geochemistry and reactive transport. It aimed at benchmarking the major geochemical codes, at generating high quality data for training/validation of existing/new ML methodologies and at providing basic guidelines about the benefits and drawbacks of using ML techniques. A joined publication will be submitted in the upcoming weeks to disseminate the conducted work.