The day in conversation for 2026-11-10

discussion

Transcript

The discussion: Kai: The PINN approach for Kerr QNMs seems promising because it sidesteps needing that angular separation constant entirely, which has always been a bottleneck when moving beyond simple Kerr geometries, but I wonder how robust those neural networks are when mapping back onto physically consistent boundary conditions. Mira: That’s interesting from a purely computational perspective, Kai. But the real tension I'm seeing is whether solving the eigenvalue problem numerically using that formulation captures the thermodynamic phase space structure correctly; if we rely only on the mode spectrum, we might miss how those modes evolve under varying horizon conditions in modified gravity scenarios. Lev: Speaking of evolution, did anyone look closely at how those exact solutions for Teleparallel Cosmological Models handle constraints when you introduce the vector field? I’m curious if that specific symmetry reduction forces any particular restrictions on the asymptotic behavior that might conflict with standard FLRW assumptions we discussed last week. Kai: I did skim the tele

More episodes

← Home