Dual and double canonical bases of quantum groups

This paper proves that the dual canonical bases of Drinfeld double quantum groups coincide with Berenstein–Greenstein's double canonical bases by utilizing the geometry of NKS quiver varieties to reinterpret their algebraic construction, thereby establishing their positivity and invariance under braid group actions.

The Gist

Imagine you are trying to organize a massive, chaotic library of mathematical objects called Quantum Groups. These aren't books you can hold; they are complex algebraic structures that describe symmetries in the universe, but they are so intricate that mathematicians have struggled to find a "clean" way to list their contents.

For decades, mathematicians have been trying to build a Canonical Basis—think of this as the "perfect index" or the "standard catalog" for this library. If you have a good index, you can easily find any item, understand how items relate to one another, and predict how they behave when mixed together.

This paper, written by Ming Lu and Xiaolong Pan, solves a long-standing mystery: Two different teams of mathematicians built two different indexes for the same library, and they thought they were different. This paper proves they are actually the exact same thing.

Here is the breakdown using simple analogies:

1. The Two Competing Maps

Imagine two explorers trying to map a vast, foggy mountain range (the Quantum Group).

• Explorer A (Lusztig & Qin): They used Geometry. They looked at the landscape, the shapes of the terrain, and the "perverse sheaves" (which you can think of as special, glowing markers placed on the mountains). By studying the shape of the land, they created a map called the Dual Canonical Basis. This map is famous because it has "positive" numbers on it, meaning it's very stable and predictable.
• Explorer B (Berenstein & Greenstein): They used Algebra. They didn't look at the landscape; instead, they took the existing maps of the mountain's two halves (the "positive" side and the "negative" side) and stitched them together using a very complicated, intricate sewing machine. They called their result the Double Canonical Basis.

The Problem: Explorer B's map was built with such complex algebraic stitching that no one knew if it actually matched the clean, geometric map of Explorer A. They suspected they were the same, but they couldn't prove it. They also had many guesses (conjectures) about whether their map would behave nicely (e.g., would it stay the same if you twisted the mountain?).

2. The Breakthrough: The "NKS" Lens

The authors of this paper, Lu and Pan, decided to look at Explorer B's complicated algebraic stitching through a new lens: Quiver Varieties.

Think of a Quiver as a simple diagram of dots and arrows. A Quiver Variety is a geometric shape built from these diagrams. The authors realized that the "intricate sewing machine" used by Explorer B was actually just a specific type of geometric movement on these shapes.

They used a special tool called a $C^*$-action. Imagine this as a giant, magical spotlight rotating over the mountain range.

• When the spotlight spins, it reveals the "fixed points"—the parts of the mountain that don't move.
• The authors showed that the complicated algebraic steps Explorer B took were actually just pulling back information along the paths created by this spinning spotlight.

3. The "Aha!" Moment

By using this geometric spotlight, the authors proved two massive things:

1. The Maps are Identical: The "Double Canonical Basis" (the algebraic stitch) and the "Dual Canonical Basis" (the geometric map) are the exact same set of numbers. The complicated algebraic sewing was just a roundabout way of describing the same geometric shapes.
2. The Conjectures are True: Because the maps are the same, all the nice properties Explorer A knew about their geometric map automatically apply to Explorer B's algebraic map.
   • Positivity: The numbers in the index are all positive (no messy negative signs).
   • Symmetry: If you twist the mountain (apply a "braid group action"), the index stays the same.
   • Invariance: The map survives various mathematical transformations, like flipping the mountain upside down.

4. Why Does This Matter?

In the world of math, having two different definitions for the same thing is like having two different spellings for "color" and "colour" that you aren't sure are interchangeable. It creates confusion and makes it hard to build new theories.

• Unification: This paper unifies two major approaches (Geometry and Algebra). It tells us that the "shape" of the universe (geometry) and the "rules" of the universe (algebra) are telling the same story.
• New Tools: Because the algebraic map is now proven to be the same as the geometric one, mathematicians can now use the powerful tools of geometry to solve hard algebraic problems, and vice versa.
• The "Sl2" Example: The paper even includes a specific, detailed example for a smaller version of the mountain (called $sl_2$), writing out the exact formulas. This serves as a "proof of concept," showing exactly how the two maps align in a concrete case.

Summary

Think of this paper as the moment two cartographers realize they were drawing the same island from different angles. One drew it from a satellite (Geometry), and the other drew it by walking the trails and measuring steps (Algebra).

Lu and Pan built a bridge between the satellite view and the walking path. They proved that the "Double" map and the "Dual" map are the same island. This confirms that the island is stable, symmetrical, and beautifully structured, settling decades of debate and opening the door for future discoveries in the "quantum" world.

Technical Summary

Here is a detailed technical summary of the paper "Dual and Double Canonical Bases of Quantum Groups" by Ming Lu and Xiaolong Pan.

1. Problem Statement

The paper addresses a fundamental question in the representation theory of quantum groups: the relationship between two distinct constructions of canonical bases for the Drinfeld double quantum group $\tilde{U}$ (and its $\imath$-quantum group generalization).

• The Dual Canonical Basis: Established geometrically by Qin (building on Lusztig's work) using perverse sheaves on cyclic quiver varieties. This basis is defined for the entire quantum group $\tilde{U}$ and is known to have integral and positive structure constants.
• The Double Canonical Basis: Constructed algebraically by Berenstein and Greenstein (BG17) by combining the dual canonical bases of the positive ($U^+$) and negative ($U^-$) halves of the quantum group via intricate operations involving the bar-involution and triangular decompositions.

The Core Question: Do these two bases coincide? Specifically, does the geometrically defined dual canonical basis of $\tilde{U}$ match the algebraically defined double canonical basis? Resolving this would settle several conjectures in [BG17] regarding the positivity of structure constants and invariance under braid group actions.

2. Methodology

The authors employ a geometric approach utilizing Nakajima-Keller-Scherotzke (NKS) quiver varieties to reinterpret the algebraic construction of the double canonical basis.

• Geometric Realization: They utilize Qin's geometric realization of $\tilde{U}$ via the quantum Grothendieck ring of perverse sheaves on NKS quiver varieties (specifically cyclic quiver varieties). The dual canonical basis corresponds to the intersection cohomology (IC) complexes of these varieties.
• $C^*$-Actions and Fixed Points: A crucial technical tool is the introduction of a $\mathbb{C}^*$-action on the quiver varieties (following Nakajima). The authors analyze the fixed-point subvarieties under this action. They demonstrate that these fixed points correspond to graded quiver varieties associated with the "regular" and "singular" parts of the NKS category.
• Restriction Functors and Convolution: The authors reinterpret the algebraic operations used by Berenstein and Greenstein (specifically the "Lusztig's Lemma" steps used to define the double basis) as pullbacks along convolution diagrams.
  • They show that the restriction functors in the geometric setting correspond to the algebraic embeddings $\iota_\pm$ used in the double basis construction.
  • They prove that the pullbacks along these diagrams preserve IC sheaves, thereby inducing embeddings of quantum Grothendieck rings.
• Upper Triangularity: By leveraging the intrinsic properties of IC sheaves (specifically their behavior under restriction and the stratification of the varieties), they derive the upper triangular relations required to prove the coincidence of the bases.

3. Key Contributions

1. Proof of Coincidence: The primary contribution is the rigorous proof that the dual canonical basis (geometric) and the double canonical basis (algebraic) of the Drinfeld double $\tilde{U}$ are identical.
2. Geometric Interpretation of Algebraic Constructions: The paper provides a geometric explanation for the intricate algebraic steps in Berenstein and Greenstein's construction, linking them to the geometry of NKS quiver varieties and $\mathbb{C}^*$-fixed points.
3. Resolution of Conjectures: The coincidence result immediately validates several open conjectures from [BG17]:
   • Positivity: The structure constants of the double canonical basis lie in $\mathbb{N}[v^{1/2}, v^{-1/2}]$.
   • Invariance: The basis is invariant under Lusztig's braid group actions, the anti-involution $*$, and the Chevalley involution $\omega$.
   • Transition Coefficients: The transition matrix coefficients from the PBW basis (and Hall basis) to the double canonical basis belong to $\mathbb{N}[v^{1/2}, v^{-1/2}]$.
4. Explicit Formulas for $\tilde{U}_v(\mathfrak{sl}_2)$: The authors derive explicit formulas for the dual (double) canonical basis of the quantum group of type $A_1$ ($\tilde{U}_v(\mathfrak{sl}_2)$), comparing them with previous results and confirming the coincidence in this specific case.

4. Main Results

• Theorem A (Theorem 4.16): The double canonical basis of $\tilde{U}$ coincides with the dual canonical basis. Consequently, it is invariant under Lusztig's braid group actions, and its structure constants lie in $\mathbb{N}[v^{1/2}, v^{-1/2}]$.
• Corollaries:
  • The basis is invariant under the anti-involution $*$ and the Chevalley involution $\omega$ (Proposition 4.21, Corollary 4.22).
  • The transition coefficients from the PBW basis $\{F^a E^c K^\mu K'^\nu\}$ to the double canonical basis are in $\mathbb{N}[v^{1/2}, v^{-1/2}]$ (Corollary 4.23). This generalizes Lusztig's result for $U^+$ to the entire quantum group.
• Explicit Formulas (Section 5): For $\tilde{U}_v(\mathfrak{sl}_2)$, the paper provides a closed-form expression for the basis elements $L(v, w)$ in terms of the generators $E, F, K, K'$, confirming the match with the basis constructed by Berenstein-Greenstein and Shen.

5. Significance

• Unification of Approaches: The paper bridges the gap between the geometric categorification approach (perverse sheaves) and the algebraic combinatorial approach (Hall algebras and PBW bases) for quantum groups. It confirms that the "dual" and "double" perspectives are not just related but identical.
• Validation of Positivity: By establishing the coincidence, the paper proves the positivity of the structure constants for the double canonical basis, a property that was conjectured but not previously proven for the full quantum group.
• Foundation for Future Work: The geometric framework developed here, particularly the interpretation of restriction functors and $\mathbb{C}^*$-actions on NKS varieties, provides a robust toolkit for studying $\imath$-quantum groups and potentially the coproduct structure (which the authors note is a direction for future geometric realization).
• Connection to Cluster Algebras: The paper notes the similarity between this basis and bases constructed via cluster algebras (Shen), suggesting deep underlying connections between quantum groups, quiver varieties, and cluster structures.

In summary, this paper resolves a major open problem in the theory of quantum groups by proving that the geometrically constructed dual canonical basis and the algebraically constructed double canonical basis are one and the same, thereby confirming their desirable structural properties (positivity, invariance) through geometric means.