Milnor-type formula for the non-acyclicity class

Let kk be a perfect field, let XX be a smooth scheme over kk, let FDctf(X,Λ)\mathcal F\in D_{\rm ctf}(X,\Lambda), and let SS(F)SS(\mathcal F) be its singular support. Let ZXZ\subseteq X be a closed subscheme and let f:XY=Speck[t]f:X\to Y={\rm Spec}k[t] be a morphism that is SS(F)SS(\mathcal F)-transversal outside ZZ. The refined Gysin pull-back of the characteristic cycle by the section fdt:XTXf^*dt:X\to T^*X is

ccX/Y/kZ(F):=(fdt)!(CC(F)),cc^Z_{X/Y/k}(\mathcal F):=(f^*dt)^!(CC(\mathcal F)),

a zero-cycle class on ZZ. Milnor-type conjecture. One has

C~X/Y/kZ(F)=cl~(\ccX/Y/kZ(F))inHZ0(X,KX/Y/k),\widetilde{C}^Z_{X/Y/k}(\mathcal F)=\widetilde{{\rm cl}}(\cc^Z_{X/Y/k}(\mathcal F))\quad\text{in}\quad H^0_Z(X,\mathcal K_{X/Y/k}),

where cl~\widetilde{{\rm cl}} is the composition of the cycle class map with the natural map to HZ0(X,KX/Y/k)H^0_Z(X,\mathcal K_{X/Y/k}). This is the expected characteristic-cycle formula for non-isolated singular or characteristic points; the supplied text does not establish its resolution status.

Sources & referencesView supporting material

Primary source

Enlin Yang and Yigeng Zhao, “Cohomological Milnor formula and Saito's conjecture on characteristic classes”, arXiv:2209.11086 (2025).

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.