The day in conversation for 2026-11-19

discussion

Transcript

The discussion: Ines: That finding about fractional orders affecting parameter estimation in Caputo SIR models is interesting, especially when we consider how standard integer-order models simplify the transition dynamics between infection stages. Marcus: I'm more focused on that hierarchical tensor framework they proposed; it seems like a necessary structure if we are trying to scale multi-omics data integration beyond simple correlation matrices. What specific constraints are they placing on the dimensionality reduction when mapping those complex biological systems onto tensors? Yuki: I read the temporal tracking piece as well, and I wonder if the observed temporal structures in those step-to-step perturbations models might actually be artifacts introduced by the recursive extrema extraction method rather than inherent system behavior. Ines: If it's an artifact of the data reduction, does that imply that standard time-series analysis methods used in earlier days for parameter fitting are being misled by the way those moments are extracted? Marcus: That brings up a point about Ab-initio cryo-EM reconstruction from denoised images—if the signal fidelity is compromised by aggressive noise reduction, how reliable are those moment calculations for determining structural connectivity in neuronal networks? Yuki: The paper on cryo-EM

More episodes

← Home