Unitary Yang–Baxter operators: classification by three generating sources, localization, and virtual abelianness of braid images

The constant quantum Yang–Baxter equation asks for an invertible operator $R:V\otimes V\to V\otimes V$ satisfying $(R\otimes I)(I\otimes R)(R\otimes I)=(I\otimes R)(R\otimes I)(I\otimes R)$; a unitary solution (unitary Yang–Baxter operator, YBO) yields braid group representations relevant to knot invariants, exactly solved lattice models, and topological quantum computation. Apart from a complete classification in local dimension 2 (Hietarinta, 1992) and restricted families (set-theoretic, involutive, two-eigenvalue), no classification is known. Galindo and Rowell (2026) proposed a conjectural classification: up to a unit scalar and a local unitary change of basis, every unitary YBO is generated from three algebraic sources — monomial solutions, group-type solutions from Yetter–Drinfeld modules, and solutions from twisted group-algebra towers — using external products, direct sums, and admissible internal orthogonal sums. Two companion conjectures, originating with Rowell–Wang (2010) and Galindo–Rowell (2014), predict that the braid representations of a simple object in a unitary braided fusion category are localizable by a unitary YBO exactly when the squared Frobenius–Perron dimension is an integer, and that all braid images of unitary YBOs are virtually abelian. All three conjectures are currently open. The problem is to prove or refute each of them.

Background

The (constant, quantum) Yang–Baxter equation (YBE) is the cubic identity $(R\otimes I)(I\otimes R)(R\otimes I)=(I\otimes R)(R\otimes I)(I\otimes R)$ for an operator $R\in\mathrm{End}(V\otimes V)$ on a finite-dimensional complex Hilbert space $V$. It originated in Baxter's solution of two-dimensional exactly solved lattice models and in the quantum inverse scattering method. A solution $R$ (a Yang–Baxter operator, YBO) equips $V$ with the structure of a braided vector space: placing a copy $R_i$ of $R$ on adjacent tensor factors $(i,i+1)$ defines a representation $\rho_{R,n}:B_n\to U(V^{\otimes n})$ of the $n$-strand braid group $B_n$, since the YBE is exactly the braid relation $R_iR_{i+1}R_i=R_{i+1}R_iR_{i+1}$; the representation is unitary when $R$ is unitary.

In this program two YBOs are considered equivalent if they differ by a unit scalar and a local unitary change of basis, $R\mapsto\lambda(S\otimes S)^{-1}R(S\otimes S)$ with $\lambda\in U(1)$ and $S\in U(V)$. Classification up to this equivalence is complete only in local dimension 2 (Hietarinta, 1992). Among unitary solutions, the involutive ones ($R^2=I$) were classified by Lechner, Pennig and Wood (2019) via extremal characters of the infinite symmetric group, and Lechner (2026) classified unitary R-matrices with exactly two eigenvalues up to one exceptional class $[e^{\pi i/3},1/2,2d]$ whose existence was open; Galindo and Rowell (2026) proved this class is nonempty in local dimension 4. The general classification problem in local dimension at least 3 remains open and is the subject of this problem.

Three algebraic sources of unitary YBOs are known. Monomial solutions have, in a suitable basis, exactly one nonzero entry per row and column; they combine set-theoretic solutions (permutation-type data, classified as algebraic objects by Etingof–Schedler–Soloviev) with multiplicative cocycle data. Group-type solutions arise from Yetter–Drinfeld modules over discrete groups — vector spaces carrying a group action and a compatible group-grading. Twisted group-algebra towers, developed by Gustafson, Kimball, Rowell and Zhang, build $R$ from finite algebraic data: a finite abelian group (typically a metric group, i.e. an abelian group with a quadratic form), a normalized 2-cocycle, a bicharacter, and a coefficient function; the braid relation becomes a finite identity for the coefficients. This source contains the classical Gaussian (metaplectic) solutions associated with finite quadratic forms, which Goldschmidt and Jones used to construct link invariants.

Galindo and Rowell (2026) define a closure class $\mathcal{G}$ of realized local-tower presentations: it contains the braid-image presentations of all monomial and group-type unitary YBOs and all presentations arising from unitary braid-compatible coefficient elements in twisted group-algebra towers, and it is closed under three operations — external (tensor) products, direct sums, and admissible internal orthogonal sums (a replacement operation combining two orthogonal partial braid elements inside an already generated tower). Every operator $R_{\mathbf{d}}$ with $\mathbf{d}\in\mathcal{G}$ is a unitary YBO; the classification conjecture asserts the converse.

A group is virtually abelian if it has an abelian subgroup of finite index. Braid images of unitary YBOs control the computational use of the corresponding anyon models: a virtually abelian braid image is far from a universal gate set and the associated link invariants are efficiently evaluable, whereas dense braid images underlie universal topological quantum computation. Galindo and Rowell (2014) conjectured that the braid image $\rho_{R,n}(B_n)$ of every unitary braided vector space is virtually abelian (and finite when $R$ has finite order); this is verified for Gaussian solutions, group-type solutions, the metric families (which have finite images), and the dimension-8 cyclic families of the 2026 paper, but no general proof exists.

The second companion conjecture connects YBOs to braided fusion categories — the categories modeling anyon systems in topological phases, whose simple objects $X$ carry a Frobenius–Perron dimension $\mathrm{FPdim}(X)$, the categorical dimension given by the Perron–Frobenius eigenvalue of the fusion matrix of $X$ (the matrix whose entries are the multiplicities in the tensor-product decompositions of simple objects). A sequence of unitary braid representations $(\rho_n,V_n)$ is localizable (Rowell–Wang) if it is uniformly equivalent to the tensor-power representations of a single unitary YBO: there is a braided vector space $(W,R)$ whose local copies reproduce the action of $B_n$ on $W^{\otimes n}$ with faithful module structure for all $n$. This formalizes when anyonic braiding gates can be realized directly in the circuit model, where gates act locally on a few adjacent copies of a Hilbert space. Rowell and Wang conjectured that for a simple object $X$ the representations on $\mathrm{End}(X^{\otimes n})$ are localizable if and only if $\mathrm{FPdim}(X)^2$ is an integer. One direction is proven under the hypothesis that the braid image generates the centralizer algebras, and the criterion is verified for the Jones representation at $q=e^{\pi i/\ell}$ (localizable exactly for $\ell\in\{3,4,6\}$; the Jones representation is the braid-group action obtained from the two-dimensional simple object of the $SU(2)$ quantum-group categories and produces the Jones polynomial), for the metaplectic $SO(N)_2$ categories (shown Gaussian by Rowell–Wenzl), and — by the 2026 paper — for $SO(4)_2$ (a quotient of the Birman–Murakami–Wenzl algebra, the diagram algebra governing braid representations of the orthogonal and symplectic quantum groups) and $SU(3)_3$ (a tower of Hecke algebras — the Hecke algebra being the $q$-deformation of the symmetric group algebra — realized via a quaternionic twisted-tower operator).

The 2026 paper supports the three conjectures computationally: searches within Clifford groups (the normalizers of the Pauli group, the group of generalized clock-and-shift operators on the local Hilbert space) and a polynomial-in-Pauli ansatz in local dimensions 2 through 11 find no unitary YBO outside the conjectured class except representatives subsequently identified as twisted-tower family members; the metric families have finite braid images; the dimension-8 cyclic families have virtually abelian, sometimes infinite, images; and after reduction to the finite field $\mathbb{F}_{1201}$ the two-ququint (local dimension 5) search outputs match family data — a finite-field observation explicitly not lifted to characteristic zero. The authors state that the three conjectures remain the main organizing open questions.

Problem Statement

The problem is to settle the classification program for unitary Yang–Baxter operators proposed by Galindo and Rowell (arXiv:2608.16865), consisting of three conjectures, each of which must be proved or refuted in full generality:

1. Classification (Conjecture 1.1 of the source). Let $\mathcal{G}$ be the closure class of Definition 6.1 of the source: the smallest class of realized local-tower presentations containing the braid-image presentations of all monomial and group-type unitary YBOs and all presentations arising from unitary braid-compatible coefficient elements in twisted group-algebra towers, and closed under external products, direct sums, and admissible internal orthogonal sums. Prove or refute: every unitary Yang–Baxter operator $R$, on a finite-dimensional Hilbert space of any local dimension, is equivalent (up to a unit scalar and a local unitary change of basis) to the operator $R_{\mathbf{d}}$ of some $\mathbf{d}\in\mathcal{G}$.

2. Localization (Conjecture 1.2 of the source, after Rowell–Wang 2010). Prove or refute: for every simple object $X$ of a unitary braided fusion category, the sequence of braid group representations $(\rho^X_n,\mathrm{End}(X^{\otimes n}))$ is localizable by a unitary Yang–Baxter operator if and only if $\mathrm{FPdim}(X)^2\in\mathbb{Z}$.

3. Virtual abelianness (Conjecture 1.3 of the source, after Galindo–Rowell 2014). Prove or refute: for every unitary Yang–Baxter operator $R$ and every $n\ge2$, the braid image $\rho_{R,n}(B_n)$ is virtually abelian.

The three questions form one coherent objective — a conceptual classification of unitary YBOs together with its structural consequences — and a complete answer must resolve each of them. Their relationship is explicit: Conjecture 1.1 is the classification core; Conjecture 1.3 is the structural property that any complete classification must account for, since all three sources and all three closure operations are currently known only to produce virtually abelian braid images; Conjecture 1.2 identifies which fusion-category braid representations the classified operators realize, and each of its verified instances is supplied by an operator from the conjectured generating class. A proof of one conjecture alone, or a refutation of one alone, is a partial result that does not by itself complete the objective, though a valid refutation of any one of the three does resolve that conjecture negatively. No restriction on local dimension, spectrum, ansatz, model class, or answer form is part of the problem.

References

1. César Galindo and Eric C. Rowell, Unitary Yang–Baxter Operators: Towards a Classification, preprint (2026), arXiv:2608.16865, https://arxiv.org/abs/2608.16865
2. Eric C. Rowell and Zhenghan Wang, Localization of unitary braid group representations, Communications in Mathematical Physics 311 (2012), no. 3, 595–615, doi:10.1007/s00220-011-1386-7, arXiv:1009.0241, https://arxiv.org/abs/1009.0241
3. César Galindo and Eric C. Rowell, Braid representations from unitary braided vector spaces, Journal of Mathematical Physics 55 (2014), no. 6, 061702, doi:10.1063/1.4880196, arXiv:1312.5557, https://arxiv.org/abs/1312.5557
4. Gandalf Lechner, The classification problem for unitary R-Matrices with two eigenvalues, preprint (2026), arXiv:2603.20158, https://arxiv.org/abs/2603.20158
5. Paul P. Martin, Eric C. Rowell, and Fiona Torzewska, A Categorical Perspective on Braid Representations, preprint (2025), arXiv:2506.07950, https://arxiv.org/abs/2506.07950
6. Yasmeen S. Baki and Padmini Veerapen, Gaussian solutions to the Yang–Baxter equation and their twists, preprint (2026), arXiv:2607.13876, https://arxiv.org/abs/2607.13876
7. Jarmo Hietarinta, All solutions to the constant quantum Yang–Baxter equation in two dimensions, Physics Letters A 165 (1992), no. 3, 245–251, doi:10.1016/0375-9601(92)90044-M, https://doi.org/10.1016/0375-9601(92)90044-M
8. Paul Gustafson, Andrew Kimball, Eric C. Rowell, and Qing Zhang, Braid group representations from twisted tensor products of algebras, Peking Mathematical Journal 3 (2020), no. 2, 103–130, doi:10.1007/s42543-020-00023-5, arXiv:1906.08153, https://arxiv.org/abs/1906.08153
9. Gandalf Lechner, Ulrich Pennig, and Simon Wood, Yang–Baxter representations of the infinite symmetric group, Advances in Mathematics 355 (2019), 106769, doi:10.1016/j.aim.2019.106769, arXiv:1707.00196, https://arxiv.org/abs/1707.00196
10. Paul Martin and Eric C. Rowell, Classification of spin-chain braid representations, Communications in Mathematical Physics 407 (2026), no. 10, article 201, doi:10.1007/s00220-026-05716-z, arXiv:2112.04533, https://doi.org/10.1007/s00220-026-05716-z
11. Eric C. Rowell and Hans Wenzl, SO(N)_2 braid group representations are Gaussian, Quantum Topology 8 (2017), no. 1, 1–33, doi:10.4171/QT/85, arXiv:1401.5329, https://arxiv.org/abs/1401.5329
12. Pavel Etingof, Travis Schedler, and Alexandre Soloviev, Set-theoretical solutions to the quantum Yang-Baxter equation, Duke Mathematical Journal 100 (1999), no. 2, 169–209, doi:10.1215/s0012-7094-99-10007-x, https://doi.org/10.1215/s0012-7094-99-10007-x