Construction of PBW and canonical bases in the folding extension algebras for symmetrizable types

Yumeng Wu

REVIEW 3 major objections 4 minor 16 references

Working in the folding extension algebra of a finite-type quiver with an admissible automorphism, this paper proves that the classes of indecomposable projective modules (canonical basis) are obtained from the dual standard modules (PBW bas

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · deepseek-v4-flash

2026-08-01 02:17 UTC pith:5E4EQPEV

load-bearing objection A genuine extension of Kato and Varagnolo-Vasserot to the folded setting, with a triangularity theorem that is likely right; the real gaps are in the base-ring bookkeeping and the compressed proof of the projective resolution. the 3 major comments →

arxiv 2607.25495 v1 pith:5E4EQPEV submitted 2026-07-28 math.QA math.RT

Construction of PBW and canonical bases in the folding extension algebras for symmetrizable types

Yumeng Wu This is my paper

classification math.QA math.RT MSC 16G2017B37

keywords folding extension algebrasquiver Hecke algebrascanonical basisPBW basisskew group algebraperverse sheavessymmetrizable typesGrothendieck groups

verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper extends the geometric construction of quantum-group bases to quivers equipped with an admissible automorphism a. It defines the folding extension algebra R(ν)⟨a⟩ as a skew group algebra, and constructs standard, dual standard, projective, and simple modules over it. In finite type, it proves that the dual standard modules form a PBW basis of the Grothendieck group of projectives, while the indecomposable projective modules form the canonical basis, and that the change-of-basis matrix from PBW to canonical is upper triangular with all diagonal entries equal to 1. The coefficients are Laurent polynomials in t with coefficients in Z[ζ], ζ a primitive n-th root of unity. This gives a geometric categorification of the PBW and canonical bases and of their relation in the symmetrizable setting, and recovers the known symmetric case when a is the identity.

Core claim

The core discovery is a unipotent triangular relation between two bases of the Grothendieck group of projective modules over the folding extension algebra. For finite type, for each a-invariant orbit λ, μ, the class of the indecomposable projective module (P_μ, φ̂_μ) is expressed as Σ_λ p_{μ,λ}(t)[(K̃_λ, γ_λ)], with p_{μ,λ} ∈ Z[ζ][t,t^{-1}], p_{μ,λ}=0 whenever μ<λ, and p_{λ,λ}=1. Equivalently, the matrix expressing canonical-basis classes in terms of PBW-basis classes is upper triangular with ones on the diagonal. The proof constructs the dual standard modules as Ext-algebra modules associated to the geometric complexes C_λ, builds a projective resolution from weight filtrations, and establi

What carries the argument

The central object is the folding extension algebra R(ν)⟨a⟩ — the skew group algebra formed from the Ext-algebra R(ν) of the perverse-sheaf complexes on quiver representation spaces and the cyclic group generated by an admissible automorphism a (an automorphism of the quiver with no arrows inside any vertex orbit, and a^n=1). This object encodes a-invariant (folded) modules as R(ν)-modules with a compatible automorphism action. The argument that carries the main theorem combines: (1) identifying indecomposable projectives with Ext-modules of a-invariant simple perverse sheaves, with ζ-eigenvalue and trace-zero summands; (2) constructing dual standard modules (K̃_λ, γ_λ) from the geometric co

Load-bearing premise

The central claim collapses if the filtration of a folded standard module (K_λ,β_λ) contains a simple summand (L_μ, ζ^r θ_μ) with μ<λ — the paper cites this order property from the unfolded case and does not prove it for the ζ-eigenspace and trace-zero summands that appear after folding.

What would settle it

Compute the composition factors of the folded standard module (K_λ,β_λ) in a finite-type quiver with a nontrivial admissible automorphism (e.g., the A_3 quiver with arrows 1→2, 3→2 and a swapping 1 and 3). If any simple summand (L_μ, ζ^r θ_μ) with μ<λ occurs, then the coefficient p_{μ,λ}(t) is nonzero for μ<λ, the transition matrix is not upper triangular with diagonal ones, and Theorem 3.16/Remark 3.17 is false.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

In finite type, the Grothendieck group K(P_ν) of projective modules over R(ν)⟨a⟩ carries two explicit bases — canonical (indecomposable projectives) and PBW (dual standard modules) — and the transition matrix between them is upper triangular with diagonal entries one.
Consequently, the class of every indecomposable projective module is a finite Laurent-polynomial combination of dual standard modules, with no contributions from lower orbits.
The bialgebra structure on ⊕_ν K(P_ν), induced by induction and restriction functors, identifies it with the negative part of the quantum group of the folded datum, so the construction categorifies these bases geometrically.
The functor Ext•( , L) preserves the signed basis and the crystal structure, so the folded setting retains the standard structural features of the unfolded symmetric case.
When a = id, the results reduce to the known symmetric extension-algebra statement; for nontrivial a, the coefficients remain Laurent polynomials but lose positivity.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

A direct check of the filtration order property in the folded category — for example, computing the composition factors of (K_λ,β_λ) for an A_3 quiver with arrows 1→2, 3→2 and automorphism swapping 1 and 3 — would test the upper-triangular claim independently of the unfolded statement cited in the paper.
The paper's skew-group construction is set up for the cyclic group generated by the admissible automorphism; the same formalism should extend to arbitrary finite groups acting admissibly on the quiver, producing analogous triangular PBW/canonical relations for more general foldings.
Because the dual standard modules are realized from the complexes C_λ, the transition coefficients p_{μ,λ}(t) are natural candidates for q-analogues of intersection-cohomology Poincaré polynomials of folded orbit closures; the ζ-eigenvalue bookkeeping may account for a signed version when a is nontrivial.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Desk Editor's Note

A genuine extension of Kato and Varagnolo-Vasserot to the folded setting, with a triangularity theorem that is likely right; the real gaps are in the base-ring bookkeeping and the compressed proof of the projective resolution.

read the letter

Short version: this is a real paper with a plausible main theorem, but the write-up is a sketch in exactly the places that matter. The folded triangular transition matrix between PBW and canonical bases is new and worth taking seriously; the paper is honest that much of the framework comes from McNamara and Kato. I would not desk-reject it, but it needs revision before I'd trust the details.

What's actually new: the definition of standard and dual standard modules in the folded extension algebra, the pairing identity (Prop 3.15), and the triangular relation (Thm 3.16 + Remark 3.17). Theorem 3.8 is explicitly said to recover McNamara's folding categorification, so the credit line is right. The geometric setup is standard, and the bialgebra/crystal part is a routine extension once the skew group algebra is in place.

Soft spots, in proportion: the base ring is a real ambiguity. The paper fixes O=Z[ζ] for Grothendieck groups but then uses 1/n in Lemma 3.4 and Prop 3.10 without saying whether the skew group algebra is over Q_l(ζ) (where n is invertible) or over O. As written, Lemma 3.4's Fourier decomposition is not valid over Z[ζ] for n>1. This is fixable, but it needs to be stated.

The bigger issue is that the proof of the projective resolution (Thm 3.14) is a compressed version of Kato's method, with an unproved dimension-vector splitting. A referee will need the details filled in before the main theorem can be certified. The stress-test concern about Remark 3.17 is less damning than it first appears: the underlying R(ν)-module K_λ is still Kato's K_λ, so any a-invariant simple constituent (L_μ, ζ^r θ_μ) of the folded module has L_μ as an R(ν)-composition factor, and Kato's μ≥λ bound applies. The remark should say this explicitly; as written it reads like an unproved transfer of a filtration statement, but the filtration does not actually need to be β_λ-stable. That said, the paper still needs to prove the coefficient of the diagonal is exactly 1, not just up to a scalar, and that is not fully demonstrated.

Who it's for: people working on geometric categorification of quantum groups, especially KLR/folding extensions. It deserves a serious referee, but the referee should ask for a rewritten proof of Section 3.3 with the base ring made explicit and the gaps filled. I'd encourage engagement rather than desk rejection.

Referee Report

3 major / 4 minor

Summary. The paper constructs, for a quiver with admissible automorphism a, the folding extension algebra R(ν)⟨a⟩ as a skew group algebra over the Ext-algebra of Lusztig sheaves, following McNamara's folding and Varagnolo–Vasserot's geometric realization. It defines folded standard and dual standard modules, studies their duality, and introduces Grothendieck groups with a trace-like pairing. The main claim is that, in finite type, the classes of indecomposable projectives (the canonical basis) are related to the dual standard modules (the PBW basis) by an upper-triangular transition matrix with diagonal entries one (Theorem 3.16 and Remark 3.17). The paper also claims that Ext^•(−, L) preserves signed bases and crystal structures (Theorem 3.8).

Significance. If the main theorem were fully proved, the paper would give a geometric categorification of the PBW-to-canonical transition in the symmetrizable/folding setting, extending work of Kato and Varagnolo–Vasserot. The organizational plan is sensible, and the paper makes explicit use of substantial prior machinery: Lusztig's a-equivariant sheaf theory, McNamara's folded KLR algebras, Kato's Ext-algebra computations, and Varagnolo–Vasserot's identification of KLR algebras with extension algebras. The construction of the skew group algebra and the definition of the two bases are natural. However, the central triangularity claim rests on a nontrivial filtration statement that is asserted rather than proved, and there is a base-ring ambiguity involving 1/n that affects several structural results. These are not cosmetic issues; they are load-bearing for the main theorem. The paper is therefore promising but not yet complete.

major comments (3)

[Remark 3.17 / Theorem 3.16] The assertion that every simple constituent (L_μ, ζ^r θ_μ) of the folded standard module (K_λ, β_λ) satisfies μ ≥ λ is exactly the step that makes the transition matrix upper triangular. Remark 3.17 cites [4] for this, but [4] proves the analogous filtration statement only for the unfolded R(ν)-module K_λ. The object (K_λ, β_λ) is an R(ν)⟨a⟩-module, and a composition series in that category has subquotients (L_μ, ζ^r θ_μ); β_λ is not shown to preserve any R(ν)-filtration whose subquotients are the unfolded L_μ. Unless one proves that the folded Jordan–Hölder constituents are controlled by the same orbit-order condition, the conclusions of Remark 3.17 — vanishing for μ<λ and the diagonal entry one — do not follow. Since Theorem 3.16 and the abstract's central claim rest on this, the proof is incomplete.
[Section 3.1–3.2, Lemma 3.4 and Proposition 3.10] The paper fixes O=Z[ζ] in Section 3, but Lemma 3.4 decomposes R(ν)⟨a⟩ into n eigencomponents and Proposition 3.10 explicitly uses 1/n when defining the splitting si_N(x)=(1/n)Σ a^{-k} i_N(a^k x). Such idempotent decompositions and splittings require n to be invertible in the coefficient ring. As written, the displayed components in Lemma 3.4 are not direct summands of R(ν)⟨a⟩ over O=Z[ζ], and the proof of Proposition 3.10 is invalid without a base change. This affects Corollary 3.5 (classification of indecomposable projectives) and the finite-global-dimension statement used later. The manuscript should either work over a base ring containing 1/n and then justify the integrality of the basis coefficients in O[t,t^{-1}], or provide a genuinely integral argument.
[Theorem 3.14 and Proposition 3.15] The construction of the projective resolution (Q(C_λ), d) of the dual standard module is only sketched. Lemma 3.12 and the diagram chase are presented in condensed form, and the passage from the unfolded Kato resolution to a resolution in R(ν)⟨a⟩-mod with projective and traceless summands needs a detailed verification. Likewise, Proposition 3.15 asserts that the induced a-action on the one-dimensional Hom spaces Hom_{R(ν)}(eK_λ, DK_λ) is the identity; this is justified by self-duality of (P_λ, φ_λ), but a direct proof in the folded setting is required because it underlies the dual-basis relation used in the matrix argument of Theorem 3.16.

minor comments (4)

[Remark 3.18] There is a typo: 'coeifficients' should be 'coefficients'.
[Definition 3.7] The terms 'traceless' and 'trace-zero elements' are used repeatedly, but the definition of a permutation morphism/traceless element is informal. A precise definition would improve readability and prevent ambiguity in relation (4).
[Proposition 3.10] The chain of inequalities involving projective dimensions is terse. In particular, the first inequality proj.dim_{R(ν)} M ≥ proj.dim_{R(ν)⟨a⟩}(R(ν)⟨a⟩⊗_{R(ν)} M) is not immediate and should be justified with a reference or a short argument.
[Section 3.1] The definition of the automorphism a on R(ν) via the diagram involving (a∗)^{-1}(φ_0[n]^{-1}∘ f∘ φ_0) is not fully explained. A sentence clarifying the degree shifts and the role of the cyclicity condition would help.

Circularity Check

0 steps flagged

No load-bearing circularity: the transition-matrix theorem is reduced to Kato's independent filtration result; only a minor non-load-bearing self-citation ([6]) appears.

full rationale

The claimed derivation of upper-triangularity (Theorem 3.16 + Remark 3.17) is not circular. Theorem 3.16 identifies the PBW-to-canonical coefficients with the standard-module-to-simple coefficients by inverting the two dual bases supplied by the pairing (Proposition 3.15). This is a standard algebraic identity, not an assumption of the conclusion. The triangularity and diagonal-one statements are then imported from Kato's external theorem [4], which concerns the unfolded R(ν)-modules Kλ and Lμ. The folded standard module (Kλ, βλ) has the same underlying R(ν)-module Kλ, and any R(ν)⟨a⟩-composition series of it forgets to an R(ν)-composition series of Kλ; hence a folded constituent (Lμ, ζ^r θμ) forces Lμ to be a constituent of Kλ, and Kato's bound μ ≥ λ applies. Thus Remark 3.17 is a genuine reduction, though it compresses the forgetting argument. The only self-citation is the attribution of Cλ as "the geometric realization of PBW basis in [6]" (Lan–Wu–Xiao, including the present author). Since Cλ is explicitly constructed in the same paragraph and [6] is not used in the proof of the transition-matrix theorem, this citation is not load-bearing. Non-circular correctness concerns are the 1/n idempotent decomposition in Lemma 3.4 and Proposition 3.10 over O = Z[ζ], and the unproved assertion in Proposition 3.15 that the induced a-action on the Hom space is the identity; these are gaps or possible errors, not definitional or fitted circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 7 axioms · 0 invented entities

The central claim relies on Lusztig's semisimple sheaves, Varagnolo–Vasserot's KLR identification, Kato's extension-algebra results, and Lusztig's a-equivariant lemmas. Two additional assumptions are introduced in this paper without full proof: the extension of Kato's filtration statement to the folded standard modules, and the invertibility of n in the base ring. No free parameters are fitted; no new entities are invented.

axioms (7)

standard math Decomposition theorem: L_ν = (π_ν,Ω)_! Q_l[dim F_ν,Ω] is a semisimple complex on E_{V,Ω} (BBD [2, Thm 5.4.5]).
Invoked after Definition 2.2 to form Q_V; all later Ext-algebra constructions assume semisimplicity of Weil sheaves.
domain assumption R(ν) is the KLR algebra in the simply-laced case and Ext^•(P,L) are indecomposable projectives (Varagnolo–Vasserot [16, Thm 4.4]).
Used in Corollary 3.5 for indecomposability of folded projectives and in Prop 3.10 via finite global dimension.
domain assumption Kato's results on extension algebras: finite global dimension and Ext^k_{R(ν)}(K̃_λ, DK_μ)=0 for k≠0 or μ≠λ ([4, Thm 2.7, Prop 3.9]).
Prop 3.10 and Prop 3.15 import these facts directly; the folded statements are reduced to them.
domain assumption Lusztig's a-equivariant structure: a^*L_{aν} ≅ L_ν, the map φ_0, and the restriction formula Res^ν_{ν',ν''}(L_ν, φ_0) = Σ (L_ν',φ_0)⊠(L_ν'',φ_0)[M(ν',ν'')] ⊕ T ([8, Lem 12.3.2, Lem 12.3.3, Prop 12.6.3]).
These supply the periodic structure used to define the skew group algebra and the bialgebra/comultiplication in Theorem 3.8.
domain assumption Weight-filtration facts for mixed perverse sheaves: existence of a weight filtration and vanishing of Hom between objects of different weights ([1, Thm 5.4.16, Lem 5.4.14]).
Used in Lemma 3.12 and Corollary 3.13 to construct F^{<l}K and the graded pieces gr^l C_λ; these are external to the paper.
ad hoc to paper The folded standard module (K_λ,β_λ) has filtration factors only (L_μ, ζ^r θ_μ) with μ ≥ λ and exactly one top (L_λ,θ_λ).
Remark 3.17 asserts this for the folded case citing [4]; no proof is given. This is load-bearing for the upper triangularity conclusion.
ad hoc to paper n is invertible in the coefficient ring for the skew group algebra.
Lemma 3.4's decomposition and Prop 3.10's averaging operator require 1/n; O=Z[ζ] does not contain it, so the proof depends on an unstated localization/field change.

pith-pipeline@v1.3.0-alltime-deepseek · 16969 in / 31203 out tokens · 295719 ms · 2026-08-01T02:17:05.688847+00:00 · methodology

0 comments

read the original abstract

We study extension algebras arising from Lusztig's construction of perverse sheaves on quiver representation spaces equipped with an admissible automorphism $a$, leading to folding Khovanov--Lauda--Rouquier (KLR) algebras. Extending the method of M. Varagnolo and E. Vasserot to Lusztig's symmetrizable setting, we construct the skew group algebra $R(\nu)\langle\mathbf{a}\rangle$, where $\mathbf{a}$ is induced by the admissible automorphism $a$, and show that $\Ext^{\bullet}(\ ,L)$ preserves the signed basis and the crystal structure. In this framework, for finite types, we define standard and dual standard modules, analyze their behavior under dualities, and study the resulting Grothendieck groups. We prove that the transition matrix from the classes of indecomposable projective modules, which give the canonical basis, to the dual standard modules, which give the PBW basis, is upper triangular with diagonal entries equal to one. This provides a geometric categorification of these bases and of their relation in the symmetrizable setting.

discussion (0)

Reference graph

Works this paper leans on

16 extracted references · 1 linked inside Pith

[1]

P. N. Achar.Perverse sheaves and applications to representation theory, volume 258 ofMathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 2021

2021
[2]

A. A. Be ˘ ılinson, J. Bernstein, and P. Deligne. Faisceaux pervers. InAnalysis and topology on singular spaces, I (Luminy, 1981), volume 100 ofAst´ erisque, pages 5–171. Soc. Math. France, Paris, 1982

1981
[3]

Y. Bi. The product of simple modules over KLR algebras and quiver Grassmannians.Representation Theory, 28(17):552–592, 2024

2024
[4]

S. Kato. An algebraic study of extension algebras.American Journal of Mathematics, 139(3):567–615, 2017

2017
[5]

Khovanov and A

M. Khovanov and A. D. Lauda. A diagrammatic approach to categorification of quantum groups i.Represent. Theory, 13:309–347, 2009

2009
[6]

Y. Lan, Y. Wu, and J. Xiao. Structure coefficients for quantum groups, 2025

2025
[7]

G. Lusztig. Canonical bases arising from quantized enveloping algebras.Journal of the American Mathematical Society, 3(2):447–498, 1990

1990
[8]

Lusztig.Introduction to quantum groups, volume 110 ofProgress in Mathematics

G. Lusztig.Introduction to quantum groups, volume 110 ofProgress in Mathematics. Birkh¨ auser Boston, Inc., Boston, MA, 1993. 27

1993
[9]

G. Lusztig. Canonical bases and Hall algebras.Representation theories and algebraic geometry, pages 365–399, 1998

1998
[10]

Y. Ma, T. Shoji, and Z. Zhou. Foldings of KLR algebras.Journal of Algebra, 639:60–98, 2024

2024
[11]

P. J. McNamara. Folding KLR algebras.Journal of the London Mathematical Society, 100(2):447–469, 2019

2019
[12]

Reiten and C

I. Reiten and C. Riedtmann. Skew group algebras in the representation theory of artin algebras.Journal of Algebra, 92(1):224–282, 1985

1985
[13]

Rouquier

R. Rouquier. 2-Kac-Moody algebras, 2008. arXiv:0812.5023

Pith/arXiv arXiv 2008
[14]

Rouquier

R. Rouquier. Quiver Hecke algebras and 2-Lie algebras.Algebra Colloquium, 19(02):359–410, 2012

2012
[15]

Schiffmann

O. Schiffmann. Lectures on canonical and crystal bases of Hall algebras. InGeometric methods in representation theory. II, volume 24 ofS´ emin. Congr., pages 143–259. Soc. Math. France, Paris, 2012

2012
[16]

Varagnolo and E

M. Varagnolo and E. Vasserot. Canonical bases and KLR-algebras.Journal f¨ ur die reine und angewandte Mathematik, 2011(659):67–100, 2011. Beijing International Center for Mathematical Research, Beijing 100871, P. R. China Email address:2506397175@pku.edu.cn (Y. Wu)

2011

[1] [1]

P. N. Achar.Perverse sheaves and applications to representation theory, volume 258 ofMathematical Surveys and Monographs. American Mathematical Society, Providence, RI, 2021

2021

[2] [2]

A. A. Be ˘ ılinson, J. Bernstein, and P. Deligne. Faisceaux pervers. InAnalysis and topology on singular spaces, I (Luminy, 1981), volume 100 ofAst´ erisque, pages 5–171. Soc. Math. France, Paris, 1982

1981

[3] [3]

Y. Bi. The product of simple modules over KLR algebras and quiver Grassmannians.Representation Theory, 28(17):552–592, 2024

2024

[4] [4]

S. Kato. An algebraic study of extension algebras.American Journal of Mathematics, 139(3):567–615, 2017

2017

[5] [5]

Khovanov and A

M. Khovanov and A. D. Lauda. A diagrammatic approach to categorification of quantum groups i.Represent. Theory, 13:309–347, 2009

2009

[6] [6]

Y. Lan, Y. Wu, and J. Xiao. Structure coefficients for quantum groups, 2025

2025

[7] [7]

G. Lusztig. Canonical bases arising from quantized enveloping algebras.Journal of the American Mathematical Society, 3(2):447–498, 1990

1990

[8] [8]

Lusztig.Introduction to quantum groups, volume 110 ofProgress in Mathematics

G. Lusztig.Introduction to quantum groups, volume 110 ofProgress in Mathematics. Birkh¨ auser Boston, Inc., Boston, MA, 1993. 27

1993

[9] [9]

G. Lusztig. Canonical bases and Hall algebras.Representation theories and algebraic geometry, pages 365–399, 1998

1998

[10] [10]

Y. Ma, T. Shoji, and Z. Zhou. Foldings of KLR algebras.Journal of Algebra, 639:60–98, 2024

2024

[11] [11]

P. J. McNamara. Folding KLR algebras.Journal of the London Mathematical Society, 100(2):447–469, 2019

2019

[12] [12]

Reiten and C

I. Reiten and C. Riedtmann. Skew group algebras in the representation theory of artin algebras.Journal of Algebra, 92(1):224–282, 1985

1985

[13] [13]

Rouquier

R. Rouquier. 2-Kac-Moody algebras, 2008. arXiv:0812.5023

Pith/arXiv arXiv 2008

[14] [14]

Rouquier

R. Rouquier. Quiver Hecke algebras and 2-Lie algebras.Algebra Colloquium, 19(02):359–410, 2012

2012

[15] [15]

Schiffmann

O. Schiffmann. Lectures on canonical and crystal bases of Hall algebras. InGeometric methods in representation theory. II, volume 24 ofS´ emin. Congr., pages 143–259. Soc. Math. France, Paris, 2012

2012

[16] [16]

Varagnolo and E

M. Varagnolo and E. Vasserot. Canonical bases and KLR-algebras.Journal f¨ ur die reine und angewandte Mathematik, 2011(659):67–100, 2011. Beijing International Center for Mathematical Research, Beijing 100871, P. R. China Email address:2506397175@pku.edu.cn (Y. Wu)

2011