The Abel-Jacobi map for complete intersections (Q1359388)

For a general odd-dimensional hypersurface \(X\) in \(\mathbb{P}_{\mathbb{C}}^{2m}\), \textit{M. L. Green} [J. Differ. Geom. 29, No. 3, 545-555 (1989; Zbl 0692.14003)] and \textit{C. Voisin} showed that the image of the Abel-Jacobi map \(\Phi_X: Z_{\text{hom}}^2(X)\to J^2(X)\) is contained in the torsion points of \(J^2(X)\). The author generalizes this result to the case of general odd-dimensional complete intersections \(X=V(f_0,\dots, f_r)\) in \(\mathbb{P}_{\mathbb{C}}^{2m+r}\) and classifies the exceptional cases. A less explicit, but even more general version of this theorem can be found in a paper by \textit{M. V. Nori} [see Invent. Math. 111, No. 2, 349-373 (1993; Zbl 0822.14008)]. The main technical difficulty in the proof is to find an algebraic description of the variable cohomology of \(X\). This is achieved by showing that it is isomorphic to certain graded pieces of a bigraded ring \(R\) which generalizes the Jacobi ring. More precisely, the ring \(R\) is the quotient of \(S= \mathbb{C}[x_0,\dots, x_n,y_0,\dots, y_r]\) (with \(n=2m+r\) above) such that \(\deg(x_i)= (0,1)\) and \(\deg(y_j)= (1,-\deg(f_j))\) by the Jacobi ideal \(J(F)= (\partial F/\partial x_0,\dots,\partial F/\partial x_n)\), where \(F=y_0f_0+\cdots+ y_rf_r\). When the author combines this description with a suitable ``symmetrizer lemma'' for the operation of \(S_{(1,0)}\) on \(R\), he is able to finish the proof using standard arguments.