0:005:08

Recursos

Resumen

Take a Hodge class on a power Aᵐ of an abelian variety with an imaginary-quadratic action, and say which part of the proof makes each piece of it algebraic. That is the skill this class builds, following the paper Weil classes and Hodge classes on abelian powers.

Why powers are harder [0:07]

Divisor classes are the easy case. Weil's determinant plane is the first hard one: on a variety of dimension 2n with signature (n, n), divisors do not span it. On Aᵐ, Hodge classes are tensors in first cohomology that couple several copies, and at special periods, where the Hodge group drops, extra tensors appear. The paper proves the rational Hodge conjecture in every codimension on every power for two families [0:52]:

  • sixfolds with an imaginary-quadratic action whose polarization gives a hyperbolic (3,3) Hermitian form;
  • every variety of dimension at most five with such an action, with no polarization fixed in advance.

The procedure [1:26]

  1. Lift the class to a balanced tensor on m copies of H = H¹(A, ℚ). Diagonal pullback brings it back.
  2. Invariant theory for Sp, SO and SL breaks invariant tensors into pairings and top exterior lines (determinants). Pairings are divisor classes.
  3. Balance [2:07]: each opposite pair D(W) ⊗ D(W̄) comes from its pair tensor. The balance equations then leave one determinant, the one for a whole eigenspace of the quadratic action.
  4. Supply it [2:37]: pad A with 6 − n CM elliptic curves, giving a split (3,3) sixfold. Its Weil plane is algebraic (Markman's secant-sheaf construction, Theorem 3.1). Project onto A's slots and contract each curve's pairing. The companion CM theorem makes the Hodge lifts algebraic.

Quick check [3:34]: a (2,2) fourfold needs two curves, one on each side, to reach (3,3).

The Weil-plane input [4:00] starts from a genus-three Jacobian. You get an ideal sheaf from Abel orbits, then a Fourier–Mukai transform and duality, and its Chern character carries a nonzero Weil component. Semiregularity and a Baire argument then spread algebraicity to every period. Next, read Section 4 of the paper, where this seed is built.

Hodge conjecture on abelian variety powers

OpenAI Math Preprints

Clase 63 / 100

Ver clases