Announcement_8

Paper accepted at NeurIPS 2026:

• The paper Permutation-Invariant Spectral Learning via Dyson Diffusion, with Tassilo Schwarz, Cai Dieball, Constantin Kogler, Renaud Lambiotte, Arnaud Doucet and Aljaz Godec was accepted as a Poster at NeurIPS 2026. In it we design a diffusion model focused on Dyson’s Brownian motion to learn spectra without the need for permutation-equivariant learning architectures, while preserving all remaining information.

---