Harvard/MIT Algebraic Geometry Seminar
Spring 2019
Tuesdays at 3 pm
The Harvard/MIT Algebraic Geometry Seminar will alternate between MIT
(2147) and Harvard (Science Center 507).

Feb 52018
SC 507 Harvard
Mihai Fulger, University of Connecticut
Seshadri constants for bundles
abstract±
For line bundles, Seshadri constants are classical local positivity invariants at a point. They detect ampleness in several ways, and they measure jet separation at the point in an asymptotic sense. A generalization of Seshadri constants to higher rank was first studied explicitly by Hacon. In our work, we extend the classical rank one results. We also give three applications. First, we offer a Seshadri characterization of projective space as the only Fano manifold whose tangent bundle has positive Seshadri constant for at least one point. We conjecture that the Fano assumption can be removed, and offer the case of surfaces as evidence. Second, we make partial progress towards a conjecture on the shape of the nef cone of the self product of a very general curve of high genus. Third, we observe that Seshadri constants control jet separation for direct images of pluricanonical sheaves. This is all in joint work with Takumi Murayama.

Feb 122019
SC 507 Harvard
Dan Abramovich, Brown University
Moduli technique in resolution of singularities
abstract±
Semistable reduction, a form of resolution of singularities of families,
is often the first step in constructing compactified moduli spaces, and
can be used to discover their properties. I will describe workinprogress
with Michael Temkin and Jaroslaw Wlodarczyk in which we prove functorial
semistable reduction for families of varieties in characteristic 0,
refining work with Karu from 2000. Techniques developed for moduli spaces
enter in unexpected ways.

Feb 192018
2147 MIT
Maksym Fedorchuk, Boston College
Standard models of low degree del Pezzo fibrations
abstract±
A del Pezzo fibration is one of the natural outputs of the
Minimal Model Program for threefolds. At the same time, geometry of
an arbitrary del Pezzo fibration can be unsatisfying due to the
presence of nonintegral fibers and terminal singularities of an
arbitrarily large index. In 1996, Corti developed a program of
constructing `standard models' of del Pezzo fibrations within a fixed
birational equivalence class. Standard models enjoy a variety of
desired properties, one of which is that all of their fibers are
QGorenstein integral del Pezzo surfaces. Corti proved the existence
of standard models for del Pezzo fibrations of degree d\geq 2, with
the case of d=2 being the most difficult. The case of d=1 remained a
conjecture. In 1997, Kollár recast and improved the Corti’s result in
degree d=3 using ideas from the Geometric Invariant Theory for cubic
surfaces. I will present a generalization of Kollár’s approach in
which we develop notions of stability for families of low degree
(d\leq 2) del Pezzo fibrations in terms of their Hilbert points (i.e.,
low degree equations cutting out del Pezzos). A correct choice of
stability and a bit of enumerative geometry then leads to (very good)
standard models in the sense of Corti. This is a joint work with
Hamid Ahmadinezhad and Igor Krylov.

Feb 262019
SC B10 Harvard
Ethan Cotterill, Universidade Federal Fluminense (**Note unusual room**)
Real inflection points of real linear series on real (hyper)elliptic curves (joint with I. Biswas and C. Garay López)
abstract±
According to Plucker's formula, the total inflection of a linear series (L,V) on a complex algebraic curve C is fixed by numerical data, namely the degree of L and the dimension of V. Equipping C and (L,V) with compatible real structures, it is more interesting to ask about the total real inflection of (L,V). The topology of the real inflectionary locus depends in a nontrivial way on the topology of the real locus of C. We study this dependency when C is hyperelliptic and (L,V) is a complete series. We first use a nonarchimedean degeneration to relate the (real) inflection of complete series to the (real) inflection of incomplete series on elliptic curves; we then analyze the real loci of Wronskians along an elliptic curve, and formulate some conjectural quantitative estimates.

Mar 52019
2147 MIT
Sho Tanimoto, Kumamoto University
Sections on del Pezzo fibrations over P^1
abstract±
We prove movable bend and break lemma, i.e., a free curve of high degree breaks into the union of two free curves as a stable map for sections on del Pezzo fibrations over P^1. We use this to study some inductive arguments proving the irreducibility of the space of sections. Our approach is inspired by Manin’s Conjecture and also has applications to it. This is joint work with Brian Lehmann.

ThurMar 7
2449 MIT
Yongbin Ruan, University of Michigan (**Special seminar at 1 PM; Note unusual room/date/time**)
The structure of higher genus GromovWitten theory of quintic 3folds
abstract±
One of biggest and most difficult problems in the subject of GromovWitten
theory is to compute higher genus GromovWitten theory of compact CalabiYau 3fold.
There have been a collection of remarkable conjectures from physics for so called 14 oneparameter
models, simplest compact CalabiYau 3folds similar to the quintic 3folds. The backbone of
this collection of conjectures are four structural conjectures: (1) YamaguchiYau finite generation;
(2) Holomorphic anomaly equation; (3) Orbifold regularity and (4) Conifold gap condition. In
the talk, I will present background and our approach to the problem.
This is a joint work with F. Janda and S. Guo. Our proof is based on certain localization formula from
log GLSM theory developed by Q. Chen, F. Janda and myself.

ThurMar 7
2449 MIT
Felix Janda, University of Michigan (**Special seminar at 2:30 PM; Note unusual room/date/time**)
Holomorphic anomaly equations for quintic threefolds
abstract±
The generating series F_g of GromovWitten invariants of CalabiYau threefolds are conjectured to behave like quasimodular forms. Furthermore, there are holomorphic anomaly equations that control in which way F_g fails to be modular. These conjectures have been used very successfully in the physics literature, for example in the computation of GromovWitten invariants of quintics up to genus 51. In my talk I will explain how to use the new technique of log GLSM (under development with Q. Chen and Y. Ruan) to give a proof of these conjectures in the case of quintic threefolds. Based on joint work with S. Guo and Y. Ruan.

Mar 122019
SC 507 Harvard
Jake Levinson, University of Washington
BoijSöderberg Theory for Grassmannians
abstract±
The Betti table of a graded module over a polynomial ring encodes much of its structure and that of the corresponding sheaf on projective space. In general, it is hard to tell which tables of numbers can arise as Betti tables. An easier problem is to describe such tables up to positive scalar multiple: this is the "cone of Betti tables". The BoijSöderberg conjectures, proven by EisenbudSchreyer, gave a beautiful description of this cone and, as a bonus, a "dual" description of the cone of cohomology tables of sheaves.
I will describe some extensions of this theory, joint with Nicolas Ford and Steven Sam, to the setting of GLequivariant modules over coordinate rings of matrices. Here, the dual theory (in geometry) concerns sheaf cohomology on Grassmannians. The main result is an equivariant analog of the BoijSöderberg pairing between Betti tables and cohomology tables. This is a bilinear pairing of cones, with output in the cone coming from the "base case" of square matrices, which we also fully characterize.

Mar 192019
2449 MIT
Luc Illusie, Université ParisSud, Talk 1, 34 (**Note two seminar talks and unusual room**)
The de RhamWitt complex: review and prospects
abstract±
I will recall the historical background, some of the main results, and discuss open questions raised by the new approach of Bhatt, Lurie and Mathew.

Mar 192019
2449 MIT
Luc Illusie, Université ParisSud, Talk 2, 4:305:30
Remarks on the cotangent complex and the Nygaard filtration: DI revisited
abstract±
I'll revisit decompositions of de Rham complexes in positive characteristic (DeligneIllusie), by discussing some relations between the cotangent complex, liftings mod $p^2$, and the de RhamWitt complex.

Mar 262019
SC 507 Harvard
Changho Han, Harvard University
'Almost K3' stable log surfaces and curves of genus 4
abstract±
The geometry of compact moduli spaces of log surfaces is mysterious in general. Thus, describing an example with its geometric properties is already valuable. To do so, we consider an 'almost K3' stable log surface, which is a pair where the log canonical divisor is positive but very close to 0. We study compactified moduli spaces of such log surfaces, constructed using the techniques of Kollár, ShepherdBarron, Alexeev, and Hacking. We completely describe such a compactification of the moduli space of (X, D) where X is a quadric surface and D is a canonical genus 4 curve, obtaining a new birational model of the moduli space M_4 of smooth curves of genus 4. This is joint work with Anand Deopurkar.

Apr 22019
2147 MIT
Tomasz Szemberg, Pedagogical University of Cracow
Postulation in projective spaces and unexpected hypersurfaces
abstract±
Given a subscheme X in projective space, it is a very classical problem to determine the dimension of the vector space of all homogeneous polynomials of some fixed degree d vanishing along X. Even in the simplest case, when X is supported on a finite set of points in the projective plane, a complete answer is not known and it is subject to open conjectures due to Nagata (1959) and SegreHarbourneGimiglianoHirschowitz (196478). Recently Cook II, Harbourne, Migliore and Nagel observed that the situation becomes even more interesting if one replaces the space of all polynomials of certain degree by its carefully chosen subspaces. I will focus on examples exhibiting these new phenomena. This is based on joint work with Bauer, Malara and Szpond.

Apr 92019
SC 507 Harvard
Aaron Pixton, MIT
abstract±

Apr 162019
2147 MIT
Junho Peter Whang, MIT
abstract±

Apr 232019
SC 507 Harvard
Kalina Mincheva, Yale University
abstract±

Apr 302019
2147 MIT
Daniel HalpernLeistner, Cornell University
abstract±

May 72019
SC 507 Harvard
TBD
abstract±

May 142019
2147 MIT
Daniel Bragg, Berkeley
abstract±
This seminar is organized by Joe Harris (Harvard), Davesh Maulik (MIT), Brooke Ullery (Harvard). This seminar is supported in part by grants from the NSF. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.