Epsilon-class formula for determinant de Rham cohomology

At least 26 years old · documented by

Let X→SX\to S be the family under consideration, let D{\mathcal D} be the fixed divisor, and let (E,∇,L,μ)(E,\nabla,\mathcal L,\mu) consist of an admissible absolute connection, a line bundle L\mathcal L on XX, and an isomorphism

μ:ED→∼ED⊗LD\mu:E_{\mathcal D}\xrightarrow{\sim}E_{\mathcal D}\otimes\mathcal L_{\mathcal D}

compatible with the induced connection along D{\mathcal D}. The construction associates to such a tuple a class ϵ(E,∇,L,μ)∈AD2(X)\epsilon(E,\nabla,\mathcal L,\mu)\in AD^2(X); in the distinguished case L=ω(D)\mathcal L=\omega({\mathcal D}) and μ=∇X/S,D\mu=\nabla_{X/S,\mathcal D}, write it as ϵ(E,∇)\epsilon(E,\nabla).

Epsilon-class conjecture. If ∇\nabla is an integrable, admissible, absolute connection, then

det⁡HDR/S∗(E,∇)=−f∗ϵ(E,∇).\det H^*_{DR/S}(E,\nabla)=-f_*\epsilon(E,\nabla).

This is stated as a restatement of the paper's main conjecture. The class ϵ(E,∇)\epsilon(E,\nabla) was constructed immediately beforehand and its pushforward was identified with the previously defined expression; the supplied text gives no evidence that the conjecture has been resolved.

References

Primary source

Spencer Bloch and Hélène Esnault, “Gauß-Manin determinant connections and periods for irregular connections”, arXiv:math/9912095 (1999).

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.