The day in conversation for 2026-11-15

discussion

Transcript

The discussion: Mira: That optimized distillation method for correlation functions, modulating those Laplacian eigenmodes—it seems like a crucial step for getting better overlap with the ground state when dealing with excited states in these larger volumes. Kai: Yes, I’m interested in how that optimization specifically handles the noise inherent in the stochastic sources we used earlier; did you find that modifying the eigenmodes helped suppress those high-frequency artifacts more effectively than just increasing ensemble statistics? Lev: I noticed that while the distillation improves operator overlap, I still struggle to interpret how robust those improvements are when we try to extract flavor decomposition for the axial charges. We got good agreement for the isovector ones, but what about the flavor-diagonal charges? That depends heavily on how well those specific bilinears decouple from excited state contamination. Mira: Regarding that decoupling, I wonder if applying a stricter symmetry constraint during the distillation phase could isolate those diagonal contributions better, or if we are limited by the quality of the initial gauge field configurations we started with? Kai: If we tighten the constraints, we might introduce more systematic bias related to lattice artifacts near the continuum limit; I'm concerned that pushing for perfect decoupling might require configurations so fine-tuned they become computationally intractable for running a full analysis chain. Lev: True, but let’s pivot slightly—the PDF calculation using

More episodes

← Home