Orestis Oikonomou

PhD Candidate @ ETH AI Center.
An enthousiast of learning new things, with a background in applied mathematics.

[Orestis Oikonomou @oroikonomou]

Find out more about SIGS here: https://t.co/YcGDFLC9EB 🤖🧠 #AIforScience #PhysicsML #ETHZurich #PhysicalAI #OpenScience #NeuralPhysics #HybridAI #AI4Science

[Orestis Oikonomou @oroikonomou]

Surrogate accuracy isn’t the same as verification. Analytical solutions act as unit tests when you need credibility. The motivation behind the updated SIGS: grammar-valid candidates → latent exploration → residual-validated refinement (incl. coupled systems). @ETH_AI_Center

[Orestis Oikonomou @oroikonomou]

Beyond single PDEs, new version of SIGS includes results on nonlinear + coupled PDE systems (e.g., shallow water / Euler-type settings) and robustness studies (KdV). #AI4Science #compositional_learning