Condition-number conjecture for cryo-EM covariance matrix blocks

For frequency indices k1k_1 and k2k_2, let L^k1,k2\hat L^{k_1,k_2} be the corresponding block of L^\hat L, and write κ(L^k1,k2)\kappa(\hat L^{k_1,k_2}) for its condition number. Condition-number conjecture.

κ(L^k1,k2)1.4818+0.8524min(k1,k2).\kappa(\hat L^{k_1,k_2})\leq 1.4818+0.8524\min(k_1,k_2).

This follows heuristically from the proven lower bound λmin(L^k1,k2)1/(2π)\lambda_{\min}(\hat L^{k_1,k_2})\geq 1/(2\pi) together with the preceding conjectured linear growth of the maximal eigenvalue. The bound remains conjectural because the maximal-eigenvalue estimate is supported by numerical experiments rather than a theoretical proof.

Sources & referencesView supporting material

Primary source

Gene Katsevich, Alexander Katsevich and Amit Singer, “Covariance Matrix Estimation for the Cryo-EM Heterogeneity Problem”, arXiv:1309.1737 (2014).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.