Publications and Preprints

• Complex structures on S^6.
2026. [+] Video presentation
We give a self-contained exposition and proof of the construction of complex structures on the six-sphere, due to Claude under the direction of Levent Alpöge. We reinterpret the construction in more geometric terms, as a logarithmic modification of a Z-quotient of a principal G_m bundle over an elliptic surface.


• On the degree of subvarieties on abelian varieties (with Stefan Schreieder).
  2026. [+]
  We prove that on a very general principally polarized abelian g-fold, the smallest multiple of the minimal curve class which can be represented by an algebraic cycle is divisible by all primes less than g/2.
  
  
• On the non-abelian Hodge locus I (with Salim Tayou).
  Accepted to Selecta Mathematica, 2026. [+]
  We partially resolve conjectures of Deligne and Simpson concerning local systems on quasi-projective varieties that underlie a polarized variation of Hodge structure. For local systems of "compact type", we prove (1) a relative form of Deligne's finiteness theorem, for any family of quasi-projective varieties, and (2) algebraicity of the corresponding non-abelian Hodge locus.
  
  
• Polync varieties and multiparameter Kulikov models.
  2026. [+]
  "Polync varieties" are those whose singularities are locally products of normal crossing (nc) singularities. We define "colorings" of a polync variety by a set S, consisting of certain combinatorial data on the dual complex. We generalize Friedman's notion d-semistability for nc varieties to S-colored, polync varieties. Applying recent breakthroughs on the logarithmic Bogomolov-Tian-Todorov theorem, we prove the smoothability of d-semistable, K-trivial, polync varieties. We generalize the combinatorial description of Kulikov models for K3 surfaces to the polync setting, allowing for a study of degenerations of K3 surfaces with regular total space over a multiparameter base. Finally, we describe some interesting examples.
  
  
• Optimality of the Prym-Tyurin construction for $\mathcal{A}_6$ (with Olivier de Gaay Fortman, and Stefan Schreieder).
  2025. [+] Supporting code
  We prove that on a very general principally polarized abelian 6-fold, the smallest multiple of the minimal curve class which can be represented by an algebraic cycle is 6.
  
  
• Matroids and the integral Hodge conjecture for abelian varieties (with Olivier de Gaay Fortman, and Stefan Schreieder).
  2025. [+] Supporting code
  We prove that the cohomology class of any curve on a very general principally polarized abelian variety of dimension at least 4 is an even multiple of the minimal class. The same holds for the intermediate Jacobian of a very general cubic threefold. This disproves the integral Hodge conjecture for abelian varieties and shows that very general cubic threefolds are not stably rational. Our proof is motivated by tropical geometry; it relies on multivariable Mumford constructions, monodromy considerations, and the combinatorial theory of matroids.
  
  
• Combinatorics and Hodge theory of degenerations of abelian varieties: A survey of the Mumford construction (with Olivier de Gaay Fortman, and Stefan Schreieder).
  2025. [+]
  We survey the Mumford construction of degenerating abelian varieties, with a focus on the analytic version of the construction, and its relation to toric geometry. Moreover, we study the geometry and Hodge theory of multivariable degenerations of abelian varieties associated to regular matroids, and extend some fundamental results of Clemens on 1-parameter semistable degenerations to the multivariable setting.
  
  
• Boundedness of some fibered K-trivial varieties (with Stefano Filipazzi, Francois Greer, Mirko Mauri, and Roberto Svaldi).
  2025. [+]
  Whether there are finitely deformation classes of K-trivial varieties of a fixed dimension is unknown, in all dimensions greater than or equal to 3. Sometimes, a K-trivial variety admits a fibration over some lower dimensional base, in which case, the general fibers are again K-trivial varieties. We prove boundedness results, in the setting where (i) such a fibration exists, and (ii) the general fibers are either abelian or symplectic varieties. As a consequence, we resolve some open questions, such as boundedness of fibered Calabi-Yau 3-folds, and a bound on deformation classes of symplectic varieties with a Lagrangian fibration, of fixed dimension. Conditional on the generalized abundance or hyperkähler SYZ conjecture, our results prove that there are only finitely many deformation classes of hyperkähler varieties, of a fixed dimension, with second Betti number at least 5.
  
  
• On lattice-polarized K3 surfaces (with Valery Alexeev).
  Accepted to Proceedings of the Steklov Math Institute, 2025. [+]
  We correct the commonly used definitions of lattice-polarized and lattice-quasipolarized K3 surfaces.
  
  
• Shafarevich's conjecture for families of hypersurfaces over function fields (with Alice Lin and Salim Tayou).
  IMRN rnaf225, 2025. [+]
  Given a smooth quasi-projective complex algebraic variety, we prove that there are only finitely many Hodge-generic non-isotrivial families of smooth projective hypersurfaces of fixed degree and dimension over it. We prove that the finiteness is uniform and give examples where the result is sharp. We also prove similar results for certain complete intersections in projective space and more generally for algebraic varieties whose moduli space admits a period map satisfying the infinitesimal Torelli theorem.
  
  
• WEAVING QUASICRYSTALS (click for photos & description), a joint art project with Victoria Manganiello, appearing in the Bonn Math Library, Luchsinger Gallery at Greenwich Academy, and maybe coming to your mathematics institute!
  
  
• Exact enumeration of fullerenes (with Peter Smillie, and an appendix by Jan Goedgebeur).
  Duke Math Journal, Vol. 174(3): 575-613, 2025. [+]
  A fullerene, or buckyball, is a trivalent graph on the sphere with only pentagonal and hexagonal faces. We give an exact formula for the number of fullerenes with a given number of vertices. In previous work, Peter Smillie and I gave a formula which counted each fullerene by a natural weighting factor. We correct this leading term to give an exact count with no weighting factor. As input, the complicated computation requires the enumeration of fullerenes up to 400 carbon atoms. Algorithmic enumeration has a long history in computer science and computational chemistry. We employ the buckygen program, codeveloped by Jan Goedgebeur, to provide this necessary input to the paper.
  
  
• Periods of elliptic surfaces with p_g=q=1 (with Francois Greer and Abigail Ward).
  Forum of Math: Sigma, Vol. 12:e111, 2024. [+]
  We prove that the period mapping is dominant for elliptic surfaces over an elliptic curve with 12 nodal fibers, and that its degree is larger than 1.
  
  
• Compact moduli of Enriques surfaces of degree 2 (with Valery Alexeev, Zack Garza, and Luca Schaffler).
  Nagoya Math Journal, Vol. 259: 581-624, 2025. [+]
  We describe a compactification of the moduli space of Enriques surfaces with a degree 2 numerical polarization. These Enriques surfaces arise from taking a K3 double cover of P1 x P1, and then quotienting by the fixed-point free lift of an involution of P1 x P1. The image of the ramification divisor is then a canonical choice of polarizing divisor on the Enriques surface. There is an associated KSBA compactification, and we show that its normalization is also a semitoroidal compactification. The semifan is given explicitly in terms of a "folding" operation on Coxeter diagrams. Just as Kulikov models are encoded by integral-affine structures on the two-sphere, certain nice dlt models of Enriques degenerations are encoded by integral-affine structures on either the disk or RP2. We construct these dlt models and all the stable surfaces at the boundary of moduli.
  
  
• Compact moduli of K3 surfaces (with Valery Alexeev).
  Annals of Mathematics, 198(2): 727-789, 2023. [+]
  In this paper, we resolve a long-standing open problem concerning the compactification of moduli of K3 surfaces, by proving that there are (semi)toroidal modular compactifications for all degrees 2d. Our approach introduces the notion of a recognizable divisor: this is an ample divisor on the generic polarized K3 surface which is well-defined on Kulikov surfaces. We prove that the stable pair compactification associated to such a recognizable divisor is normalized by a semitoroidal compactification. Furthermore, we show that the rational curve divisor defined by the sum of the rational curves in the polarization class, is recognizable for all degrees 2d.
  
  
• Stable pair compactification of moduli of K3 surfaces of degree 2 (with Valery Alexeev and Alan Thompson).
  
  Crelle's Journal, Vol. 799: 1-56, 2023. [+]
  A stable pair compactification of the moduli space of K3 surfaces can be constructed by choosing canonically an ample divisor on each K3. For degree 2 K3s, the ramification divisor of the canonical double covering map is such a choice. We prove that up to normalization, this compactification is one of Looijenga’s semitoroidal compactifications. We explicitly identify the semifan and describe the surfaces which fiber of the boundary.
  
  
• Mixed mock modularity of special divisors (with Francois Greer and Salim Tayou).
  2023. [+]
  A well known result of Kudla-Millson proves that the generating function of special (or Heegner) divisors in an orthogonal Shimura variety is a cohomology-valued modular form. Shimura varieties admit toroidal compactifications, depending on combinatorial data called a fan. We prove that the Zariski closures of special divisors in a toroidal compactification can be corrected by an explicit linear combination of boundary divisors depending on the fan, so that the modularity property of the generating function is restored.
  
  
• The Manin-Mumford conjecture in genus 2 and rational curves on K3 surfaces (with Raju Krishnamoorthy and Daniel Litt).
  2022. [+]
  We explore the question of whether every point on a K3 surface is contained in a rational curve. Unlike the phenomenon observed by Bogomolov and Tschinkel for algebraic closures of finite fields, in characteristic zero, not all points on a Kummer K3 are contained in a rational curve whose double cover in the abelian surface has genus 2.
  
  
• Mirror symmetric compactifications of moduli spaces of K3 surfaces with a nonsymplectic involution (with Valery Alexeev).
  2022. [+]
  Following up on earlier work with Alexeev and Han, we explicitly compute the semitoroidal compactifications of the 50 moduli spaces of K3 surfaces with a nonsymplectic involution for which the fixed locus contains a curve of genus at least 2. We do so by finding Kulikov models for all Type III degenerations in all 75 moduli spaces of K3 surfaces with non-symplectic involution.
  
  
• The flex divisor of a K3 surface (with Valery Alexeev).
  IMRN rnac089, 2022. [+]
  This paper computes the degree of the flex divisor of a primitively polarized K3 surface, defined as the set of points for which there is a pencil in the linear system of the polarization whose base locus is exactly that point.
  
  
• Compactifications of moduli of elliptic K3 surfaces: stable pair and toroidal (with Valery Alexeev and Adrian Brunyate).
  Geometry & Topology, 26: 3525–3588, 2022. [+]
  We study the stable pair compactification of elliptic K3 moduli associated to the section plus the sum of the singular fibers. We prove that the normalization of this modular compactification is in fact a toroidal compactification, and compute the fan explicitly. It ends up being a refinement of a natural Coxeter fan, which results from decomposing each Coxeter chamber into nine subchambers.
  
  
• Compact moduli of K3 surfaces with a nonsymplectic automorphism (with Valery Alexeev and Changho Han).
  Transactions of the AMS, Series B 11, 2024. [+]
  We study moduli spaces of K3 surfaces with a non-symplectic automorphism fixing a curve of genus at least two. We show that such a moduli space has a canonically defined compactification by KSBA stable pairs. Furthermore, we prove that this compactification is a Looijenga semitoroidal compactification for some semifan.
  
  
• Smoothings and rational double point adjacencies for cusp singularities (with Robert Friedman).
  
  Journal of Differential Geometry 118(1): 23-100, 2021. [+]
  This paper explores smoothings of cusp singularities further. We show that the number of smoothing components is at least the number of deformation types of anticanonical pairs with the dual cycle. We also classify the adjacencies of a cusp singularity to a rational double point configuration.
  
  
• Hurwitz theory of elliptic orbifolds, I.
  
  Geometry & Topology 25: 229–274, 2021. [+]
  This paper generalizes my work with Peter Smillie to higher genera. How many triangulations of a genus g surface are there with a specified set of non-zero curvatures? It turns out the generating function is a quasi-modular form. The idea goes back to Eskin and Okounkov who connected these types of question to Hurwitz theory. The key is to count branched covers of an elliptic orbifold, i.e. a finite quotient of an elliptic curve, with specified branch profile, using representation theory of symmetric groups.
  
  
• Hurwitz theory of elliptic orbifolds, II.
  
  2018. [+]
  This paper makes some difficult technical computations which were necessary to achieve the full quasimodularity of the Hurwitz theory of elliptic orbifolds. The key step is to realize that the Hurwitz theory of the elliptic orbifold of order N is encoded by a certain tensor product of vertex operators on N copies of Fock space, multiplied the rotation operator which cyclically rotates the N tensor factors.
  
  
• Appendix to Contribution of one-cylinder square-tiled surfaces to Masur-Veech volumes (with Vincent Delecroix, Elise Goujard, Peter Zograf, and Anton Zorich).
  
  Asterisque 415: 223-274, 2020. [+]
  
  In the moduli space of abelian differentials, there is an equidistributed (with respect to the Masur-Veech volume form) subset formed by square-tiled surfaces. This paper studies which proportion of the volume is due to square-tiled surfaces formed from a fixed number of vertical or horizontal cylinders. It turns out that the one-cylinder contribution is asymptotic to the inverse of the genus. My appendix computes this contribution to the volume using some simple Hurwitz theory.
  
  
• The number of convex tilings of the sphere by triangles, squares, or hexagons (with Peter Smillie).
  
  Geometry & Topology 22: 2839-2864, 2018. [+]
  A buckyball is a spherical carbon molecule whose rings have length 6 or less. We give a formula for the number of such combinatorial arrangements with a fixed number of carbon atoms. It turns out that these numbers form the Fourier coefficients of a modular form. The key is to use a theorem of Thurston’s which identifies these triangulations with lattice points in a Hermitian hyperbolic lattice, then integrate the Siegel theta correspondence for the corresponding symmetric space.
  
  
• A proof of Looijenga's conjecture via integral-affine geometry.
  
  Journal of Differential Geometry 109(3): 467-495, 2018. [+]
  A cusp singularity is a surface singularity whose minimal resolution is a cycle of rational curves. These singularities come in dual pairs. In 1981, Looijenga proved that if a cusp singularity is smoothable, then the dual cycle appears as the anticanonical divisor on a rational surface. He conjectured the converse. This paper gives a proof of Looijenga's cusp conjecture using surgeries on integral-affine surfaces.

Selected Conference Talks

• “Geometric aspects of Hodge theory”, Colorado State University, Fort Collins, Algebraic Geometry Summer Research Institute, Summer 2025
• Algebraic Geometry Bootcamp, lecture series on K3 moduli, Colorado State University, Fort Collins, Summer 2025
• Algebraic Geometry MATRIX, Matrix Institute (Creswick, Australia), Fall 2024
• Workshop on Higher-dimensional log Calabi-Yau pairs, AIM Workshops, Pasadena, CA, Fall 2024
• School on Hodge Theory and Shimura Varieties, Essen University, Essen, Fall 2024
• A panorama of moduli spaces, Goethe University, Frankfurt am Main, Spring 2024
• Moduli, K-trivial Varieties, and Related Topics, IBS-CCG (Korea), Spring 2024
• Recent Developments in Hodge Theory, IMSA, Spring 2021, Lectures I & II
• MSRI thematic semester on flat surfaces, Fall 2019
• Teichmuller Theory, Hyperbolicity and Dynamics at IMPA, Summer 2019
• Casa Matematica Oaxaca, Flat Surfaces and Dynamics on Moduli Space, II, Summer 2019
• Reptile Conference, Spring 2019, organized with Nikon Kurnosov
• Workshop on Modularity in GW Theory, Michigan, Spring 2019
• Workshop on Holomorphic Differentials, Stonybrook, Spring 2019
• Character Varieties and TQFT, University of Auckland, Fall 2018
• Flat Surfaces and Algebraic Curves, Oberwolfach, Fall 2018
• Algebraic and Combinatorial Aspects of Tropical Geometry (AMS Sectional), Ohio State, Spring 2018 
• AGNUBC miniconference, Fall 2017
• Poisson Geometry and Stacks, Fields Institute, Summer 2017
• Conference on moduli spaces, mirror symmetry, and enumerative geometry, ICTP, Summer 2016
• Robert Friedman’s birthday, Perspectives on Complex Algebraic Geometry, Spring 2015
• IAS thematic year on topology of algebraic varieties, Fall 2014

Notes for Teaching

• Notes on 4-manifolds (Spring 2021)

Links

• My Hypocrite Reader article on BWV 849 by Bach (sadly most links broken)
• My Hypocrite Reader article on self-reference
• My Hypocrite Reader article on the integers
• Some of my math friends: Salim Tayou, Peter Smillie, Raju Krishnamoorthy, Michael McBreen, Francois Greer, Eduard Duryev (who made the cool background image)
• Some new Italian math friends: Stefano Filipazzi, Mirko Mauri, Roberto Svaldi
• My K3 comrade: Valery Alexeev