Kuznetsov Categories for Gauged Linear Sigma Models

David Favero, Daniel Kaplan, Tyler L. Kelly · 2025

Read Kuznetsov Categories for Gauged Linear Sigma Models by David Favero, Daniel Kaplan, Tyler L. Kelly (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-18402.txt.

Introduction

David Favero, Daniel Kaplan, and Tyler L. Kelly wrote this math paper, posted to arXiv in December 2025. It belongs to algebraic geometry, the study of shapes defined by equations. The authors define Kuznetsov and anti-Kuznetsov categories for gauged linear sigma models. These models, borrowed from physics, describe certain geometric spaces.

They show that for complete intersections of ample divisors in smooth toric varieties, the Kuznetsov category sits apart from an exceptional collection. They also prove that such spaces built from at least two divisors act as Fano visitors. The derived category of the Fano host then matches an anti-Kuznetsov category of the model. Results like these help map how related geometric spaces fit together.

Catalog record

CREATORDavid Favero, Daniel Kaplan, Tyler L. Kelly
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2512-18402

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Kuznetsov Categories for Gauged Linear Sigma Models David Favero, Daniel Kaplan, Tyler L. Kelly Abstract . We define Kuznetsov and anti-Kuznetsov categories for gauged linear sigma models. We show that for complete intersections of ample divisors in smooth projective toric varieties, the Kuznetsov category is left orthogonal to an exceptional collection. We prove that any complete intersection of r ≥ 2 r\geq 2 ample divisors in a Fano GIT quotient is a Fano visitor and the derived category of its Fano host is equivalent to an anti-Kuznetsov category of a gauged linear sigma model. 1. Introduction The derived category of coherent sheaves \operatorname{D}^{\operatorname{b}}(\operatorname{coh}X) on a variety X is an ‘almost-perfect’ invariant of X , encapsulating much of its geometry and descending to many other invariants, such as cohomology and K -theory. When \operatorname{D}^{\operatorname{b}}(\operatorname{coh}X) is decomposable, certain admissible subcategories can hold interesting information about X . First for cubic fourfolds followed by other Fano varieties, Kuznetsov pioneered identifying such admissible subcategories as (left) orthogonals to exceptional objects [ 27 , 28 , 30 , 31 ] . These components, now called Kuznetsov (or residual) components of \operatorname{D}^{\operatorname{b}}(\operatorname{coh}X) , have since enjoyed considerable attention in the literature. For example, they have been used to probe rationality [ 30 , 1 ] , study mirror symmetry [ 38 ] , explore Bridgeland stability [ 7 ] , and establish categorical Torelli theorems [ 9 , 20 , 34 ] . Despite their range of applications in many examples, there is at present no uniform definition of such categories. In this paper, we propose a uniform definition of Kuznetsov categories for gauged linear sigma models (GLSMs). Witten introduced gauged linear sigma models in physics as a generalized setting for the correspondence between Landau-Ginzburg models and Calabi-Yau manifolds [ 41 ] . A GLSM contains the data needed to construct a (well-behaved) superpotential w:X\rightarrow\operatorname{\mathbb{A}}^{1} on a stack X presented as a GIT quotient X=[V/\!\!/\!_{\theta}G] coming from a representation V of an algebraic group G . Given a GLSM, there is a corresponding absolute derived category defined by Positselski [ 35 ] generalizing the category of matrix factorizations. Choices of GIT quotients of V by G are parameterized by the characters of G . The character \theta , determines the Zariski-open, (semi)stability locus V_{\theta} , and hence the GIT quotient [V/\!\!/\!_{\theta}G] . Varying the character, one obtains relations between the various absolute derived categories (see, e.g., [ 17 , 37 , 18 , 16 , 5 , 15 , 40 ] ). We identify a distinguished character \theta_{K}\in\widehat{G} and define a Kuznetsov chamber to be a chamber in the GIT fan which contains this distinguished character \theta_{K} (Definition 2.6 ). We define a Kuznetsov category \mathcal{K} of the GLSM to be the absolute derived category associated to this new GLSM corresponding to the superpotential w_{K+\epsilon}:X_{K+\epsilon}\to\operatorname{\mathbb{A}}^{1} where X_{K+\epsilon}=[V/\!\!/\!_{\theta_{K+\epsilon}}G] and \theta_{K+\epsilon} is in the interior of said Kuznetsov chamber (Definition 2.7 ). This perspective also has the advantage of an obvious symmetry; one can define an anti-Kuznetsov category -\mathcal{K} as well using an anti-Kuznetsov chamber containing the character \theta_{-K}:=\theta_{K}^{-1} . The Kuznetsov (resp. anti-Kuznetsov) category apriori depends on the choice of chamber. Nevertheless, it follows from the work of Ballard-Favero-Katzarkov [ 3 ] that all choices of Kuznetsov (resp. anti-Kuznetsov) chambers yield equivalent Kuznetsov (resp. anti-Kuznetsov) categories (see Corollary 2.10 ). We then study Kuznetsov categories for geometric GLSMs (see Definition 2.12 ). In this case, X is a (stacky) vector bundle \mathcal{E} over some variety B and the superpotential w corresponds to a regular global section s\in\Gamma(B,\mathcal{E}^{\vee}) . The absolute derived category of the geometric GLSM is equivalent to the derived category \operatorname{D}^{\operatorname{b}}(\operatorname{coh}Z) of a complete intersection Z:=Z(s) in B . We use this approach to provide new semi-orthogonal decompositions of complete intersections in terms of their Kuznetsov categories. In particular, for geometric abelian GLSMs where \mathcal{E} is a split vector bundle consisting of anti-ample line bundles, we prove that the orthogonal to the Kuznetsov category is an exceptional collection. This has the following geometric reinterpretation: Theorem 1.1 (=Theorem 3.5 ) . Let Z be a complete intersection of ample divisors in a smooth projective toric variety (or Deligne-Mumford stack). Then, there are semiorthogonal decompositions \operatorname{D}^{\operatorname{b}}(\operatorname{coh}Z)=\langle\mathcal{K},E_{1},\dots,E_{t}\rangle (1.1) and -\mathcal{K}=\langle\operatorname{D}^{\operatorname{b}}(\operatorname{coh}Z),E_{1},\dots,E_{s}\rangle (1.2) where the E_{i} are exceptional objects and \mathcal{K} (resp. -\mathcal{K} ) is the Kuznetsov (resp. anti-Kuznetsov) category for the geometric GLSM associated to Z . This result can be viewed as a toric analogue of Orlov’s theorem [ 33 , Theorem 3.11] . As in Orlov’s theorem, more is known when we specialize to the hypersurface case: Theorem 1.2 (=Theorem 3.9 ) . Let Z be a smooth hypersurface in a smooth projective Fano toric variety X (or Deligne-Mumford stack). Assume \operatorname{\mathcal{O}}_{X}(Z)\cong\omega_{X}^{-q} for some q\in\operatorname{\mathbb{Q}}_{>0} . (i) If q\leq 1 then the Kuznetsov category \mathcal{K} in ( 1.1 ) is fractional Calabi-Yau. (ii) If q\geq 1 then the anti-Kuznetsov category -\mathcal{K} in ( 1.2 ) is fractional Calabi-Yau. (iii) If q=\frac{1}{r} for some r\in\operatorname{\mathbb{Z}} and \operatorname{\mathcal{O}}_{X}(Z) is Cartier, then \mathcal{K} in ( 1.1 ) is Calabi-Yau. In 2011, Bondal asked [ 8 ] : Which smooth projective varieties X have derived categories that are admissible subcategories of the derived category of some smooth projective Fano variety Y ? If such an X and Y exist, we call X a Fano visitor and Y a Fano host . Anti-Kuznetsov categories provide good candidates for Fano hosts. Theorem 1.3 (=Corollary 4.5 ) . Any complete intersection of r\geq 2 ample divisors in a Fano GIT quotient X is a Fano visitor and the derived category of the Fano host is equivalent to an anti-Kuznetsov category of a GLSM. This generalizes the main result of [ 24 ] which shows that any complete intersection in projective space is a Fano visitor. 2. Kuznetsov Categories for GLSMs In this section, we define the Kuznetsov category of a GLSM. Roughly, we take a chamber of the corresponding GIT problem corresponding to the canonical class and take its corresponding absolute derived category (see [ 35 ] and [ 4 , Definition 3.9] ). We expand on these notions below. The precise mathematical definition of a GLSM consists of a 5-tuple satisfying several conditions. Our definition of a Kuznetsov category can be given for the 5-tuple, without the need for any conditions. Definition 2.1 . A gauged linear sigma model (GLSM) is the data (V,\Gamma,\chi,\theta,w) consisting of: • a complex vector space V ; • a reductive algebraic group \Gamma\subset\text{GL}(V) acting linearly on V ; • a surjective character \chi:\Gamma\to\mathbb{C}^{*} ; • a character \theta:\operatorname{\mbox{ker}}\chi\rightarrow\operatorname{\mathbb{C}}^{*} such that the semistable locus with respect to \theta is \Gamma -invariant; and • a \chi - semi-invariant superpotential w:V\rightarrow\operatorname{\mathbb{A}}^{1} , that is, a regular map w satisfying w(g\cdot x)=\chi(g)w(x) , for all x\in V and g\in\Gamma . Remark 2.2 . The semi-invariance of w means that it defines a section of the equivariant line bundle \operatorname{\mathcal{O}}(\chi) on the quotient stack [X/G] , i.e., w\in\Gamma(X,\operatorname{\mathcal{O}}(\chi))^{G} . Definition 2.3 . We say a GLSM is abelian if \Gamma is an abelian group. Remark 2.4 . If a GLSM is abelian, then since \Gamma is reductive, one can choose a basis for V so that \Gamma\subseteq(\operatorname{\mathbb{G}_{m}})^{n} . Consequently, abelian GLSMs are

Read the complete text