AI research at ETH Zürich — neuro-symbolic AI, PDEs, symbolic discovery. Could AI understand the laws of physics? 🤖🧠
PhD candidate at the ETH AI Center & D-MATH, advised by Siddhartha Mishra and Eleni Chatzi. I build methods that pair learning with structure and symbolic reasoning to find interpretable, verifiable solutions to scientific problems Research
- Neuro-symbolic AI for scientific discovery — grammar-valid candidates → latent exploration → residual-validated refinement
- ML for PDEs & dynamical systems — operator learning, physics-informed models
- Trustworthy SciML — analytical solutions as "unit tests" for surrogate credibility
Featured work
- 🧠 SIGS — Neuro-Symbolic AI for Analytical Solutions of Differential Equations · ICML 2026 · paper · 5-min talk · site ·
SIGS - 🌊 ZeroFlood — Geospatial Foundation Model for Flood Susceptibility Mapping · paper ·
ZeroFlood - ⌚ Cross-Domain HAR — Self-Supervised Learning & Enhanced Finetuning (2024)
Find me — Scholar · Bluesky · LinkedIn · ooikonomou@ethz.ch