Zero-divisor algebras of graph functions: quantum caging, entangling routing and stochastic first-passage exclusion
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.
Kai: I'm Kai.
Mira: And I'm Mira, here with Lev.
Kai: We're looking at the paper: "Zero-divisor algebras of graph functions".
Mira: Here's the gist of the paper. Zero-divisor algebras of graph functions: quantum caging, entangling routing and stochastic first-passage exclusion constructs Lie-bialgebraic differential structures on vertex functions of a finite graph using coefficients in commutative algebras with zero divisors to derive exact Jacobi criteria and classify solutions Over Cq, the general solution is a superposition of such clique layers The canonical cobracket selects matching layers from the wider Jacobi-compatible class, and this structure makes the weighted graph Laplacian an inner derivation yielding an incidence-type vertex–edge calculus with a positive squared-Laplacian factorization.
Kai: Representing the idempotent channels by internal-state projectors gives exact quantum caging and channel-controlled transfer that creates path–channel entanglement, and the ordered real realization gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable crossover to escape under weak channel switching.
Mira: Algebraic Structure and Rigidity The paper develops a graph bracket defined by
f, g: j = Xk wjk(fjgk − fkgj), which is an inner derivation of the weighted graph Laplacian
f, 1: = -∆w f.
Kai: The Jacobi identity for this bracket is equivalent to the incident-annihilation condition wi jwik = 0 (j, k) for distinct incident edges at the same vertex, and this condition is sufficient for Jacobi compatibility.
Title and authors: Kai: So what does that mean for us? If we can define these algebraic rules and find this structure over C q, it means we can select specific types of geometries.
Mira: They specifically show that the canonical cobracket selects a much stricter subclass of Jacobi-compatible classes, which is more restrictive than just having clique layers. This selection happens because they look at the local relation where weights are zero, which is w j k w j l equals zero for distinct incident edges k and l.
Lev: So, if we have this selection mechanism, it means we can move beyond just any clique layer structure that Jacobi allows, into something more specific that has these nice algebraic properties. What does that look like practically?
Kai: It leads directly to the idea of an overlapping-clique weighting. This weighting separates the two compatibility levels, showing that it defines a Lie algebra and a Lie coalgebra but not quite a full Lie bialgebra yet. They use this to select exactly what you need for their physical models.
Mira: And this selection is key because it’s what allows them to get the exact quantum caging results. This isn't just assuming there are two-vertex channels; it's derived from the compatibility rules of the cobracket, which they call the canonical cobracket δ(ej) equals X k w j k e j ∧ e k.
The paper's summary: Kai: So, let’s look at what they actually found in their summary. They spend a lot of time on the physical realizations, which is where it gets really interesting for experimentalists.
Mira: In the quantum case, they represent those idempotent channels using orthogonal internal-state projectors. This gives them exact caging and channel-controlled entangling routing that creates path–channel entanglement between different parts of the system.
Lev: Entangling routing sounds complicated to build on hardware. What does that entangling routing actually entail? Is it just correlated outputs, or is there a specific coherent transfer happening?
Kai: It means a coherent transfer where distinct internal channel states get correlated with distinct spatial outputs. It’s a way of controlling the path and the channel simultaneously in a quantum setting.
Mira: Then they have the stochastic model, which is where things get really applied to diffusion problems. The graph itself becomes the configuration space for a graph Fokker–Planck equation, and this leads to exact first-passage exclusion.
Lev: So, if it’s a Fokker-Planck equation on the graph configuration space, how does that translate into something we can measure? What does that first-passage exclusion actually look like in a simulation or experiment?
Kai: It means the vertex coordinate representation of this Markov evolution yields exact first-passage exclusion. And they also get a controlled crossover to escape under weak channel switching, which is useful for designing robust systems.
The paper's improvements: Mira: Now let’s talk about how this work improves what came before. They show that the canonical cobracket actually separates the broader Jacobi-compatible clique-layer class from the incident-annihilating matching subclass.
Kai: That separation is important because it explains why exact realizations decompose into invariant two-vertex channels rather than just being an assumption we impose on top of a model.
Lev: So, instead of assuming you have these specific two-vertex channels, the algebra itself dictates that they must be there in this selected class. How does that change how we think about error correction overhead?
Mira: The algebraic structure derived from the canonical cobracket induces vertex–edge identities, like delta w equals partial delta and two w equals two delta dagger delta. This identifies the square of the inner derivative with a positive edge form after mixed channel products have vanished.
Kai: That’s significant because it leads to a common transport operator, DG which is related to ad1, and its second-order composition is identified as two over m times the Dirac delta dagger delta.
Lev: So if we have this operator D G and T over G, how does that relate to the actual physical dynamics of a system on hardware? What’s the overhead for running this type of transport operator?
Conclusion: Kai: So, wrapping up the whole thing, they've shown that by using zero-divisor algebras and this specific cobracket structure, you can get exact quantum caging and entanglement routing.
Mira: And on the stochastic side, you get exact first-passage exclusion and controlled escape under weak channel switching. The whole point is that this algebraic selection process is what makes the physical results come out exactly instead of just being approximations.
Lev: From my side, I see this as a way to rigorously define the transport operator T over G such that it conserves total R-valued mass through the condition one T M equals zero. That’s a nice piece of conservation law that we can check against our simulations.
Kai: So, in short, they use these algebraic tools to select the right geometry from a larger class of possibilities based on compatibility rules derived from the cobracket. It shows how deep algebraic structure can lead to these exact physical phenomena without needing extra assumptions.
Mira: It’s about showing that the compatibility hierarchy is what dictates whether you get simple clique layers or this more specific matching-layer subclass, which then gives you the precise physical properties of caging and routing we need.
Lev: It’s a solid piece of work connecting these abstract graph theory concepts to concrete transport operators and conservation laws. We should see how this algebraic definition of the operator D G translates into actual Hamiltonian terms on a real system next.
HUN-REN Wigner Research Centre for Physics · Institute for Particle and Nuclear Physics
math-ph, math.CO, math.MP, math.RA
Submitted: 2023-11-20
Updated: 2026-11-10
Comments: 29 pages
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 86/100
The gist: Zero-divisor algebras of graph functions: quantum caging, entangling routing and stochastic first-passage exclusion constructs Lie-bialgebraic differential structures on vertex functions of a finite…
Key concepts
- Graph Bracket
- A mathematical operation defined as [f, g]j = Xk wjk(fjgk − fkgj) that acts as an inner derivation of the weighted graph Laplacian. Its Jacobi identity imposes conditions on incident edges, which are crucial for determining when a structure is compatible with the underlying graph geometry.
- Canonical Cobracket
- This structure defines a Lie bialgebra and acts as a selection mechanism. It distinguishes between different types of compatibility, selecting 'matching layers' from the wider class of Jacobi-compatible clique layers, which is key to deriving the exact physical results.
- Quantum Caging
- A physical effect where quantum states are confined or 'caged' within specific paths or channels. This is achieved by representing internal state projectors using idempotents selected by the canonical cobracket, leading to exact caging and channel-controlled transfer.
- Graph Fokker–Planck Equation
- A stochastic model where the graph represents a configuration space. The second-order operator derived from the algebraic structure leads to this equation, which accurately describes first-passage exclusion and how a process escapes under weak channel switching.
Terminology
Summary
Zero-divisor algebras of graph functions: quantum caging, entangling routing and stochastic first-passage exclusion constructs Lie-bialgebraic differential structures on vertex functions of a finite graph using coefficients in commutative algebras with zero divisors to derive exact Jacobi criteria and classify solutions Over Cq, the general solution is a superposition of such clique layers The canonical cobracket selects matching layers from the wider Jacobi-compatible class, and this structure makes the weighted graph Laplacian an inner derivation yielding an incidence-type vertex–edge calculus with a positive squared-Laplacian factorization. p. 0 Keywords: graph Lie algebras, zero divisors, Lie-bialgebra rigidity, clique layers, quantum caging, graph Fokker–Planck equation,
first-passage processes
Abstract
We construct Lie-bialgebraic differential structures on vertex functions of a finite graph using
coefficients in commutative algebras with zero divisors. We derive the exact Jacobi criterion for the
graph bracket and classify its solutions. Representing the idempotent channels by internal-state projectors gives exact quantum caging and
channel-controlled transfer that creates path–channel entanglement. The ordered real realization
gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable
crossover to escape under weak channel switching.Where the paper says it
Where the paper says it
Algebraic Structure and Rigidity
The paper develops a graph bracket defined by [f, g]j = Xk wjk(fjgk − fkgj), which is an inner derivation of the weighted graph Laplacian [f, 1] = -∆w f. The Jacobi identity for this bracket is equivalent to the incident-annihilation condition wi jwik = 0 (j, k) for distinct incident edges at the same vertex, and this condition is sufficient for Jacobi compatibility. p. 0 A four-vertex diamond built from overlapping triangle layers shows that Jacobi
compatibility is strictly broader than the matching geometry generated by proper edge colouring. For the canonical cobracket, we prove a rigidity theorem over commutative 2-torsion-free rings:
Lie-bialgebra compatibility is equivalent to the local annihilation condition wi jwik = 0 for distinct
incident edges. We derive the exact Jacobi criterion for the
graph bracket and classify its solutions. Over an integral domain, each connected component of the
nonzero support is a uniformly weighted clique; over C
q
, the general solution is a superposition of
such clique layers.Where the paper says it
Where the paper says it
Bialgebraic Selection Mechanism
The canonical cobracket δ(ej) = Xk wjkej ∧ ek defines a Lie bialgebra, and over a commutative 2-torsion-free ring, Lie-bialgebra compatibility is equivalent to the local relation wjkwjl = 0 (k, l). This condition ensures that the canonical bialgebra selects matching layers from the wider Jacobi-compatible class, which is stricter than Jacobi alone permitting clique layers. p. 0 For the canonical cobracket, we prove a rigidity theorem over commutative 2-torsion-free rings:
Lie-bialgebra compatibility is equivalent to the local annihilation condition wi jwik = 0 for distinct
incident edges. Hence the canonical bialgebra selects matching layers from the wider
Jacobi-compatible class. Conversely, if every channel support is a matching, distinct incident weights have
disjoint idempotent support and therefore annihilate. □
The overlapping-clique weighting (16) separates the two compatibility levels explicitly. □
The overlapping-clique weighting (16) separates the two compatibility levels explicitly. It defines a Lie
algebra, and Lemma 3.Where the paper says it
Where the paper says it
Where the paper says it
Physical Realizations
The paper provides exact quantum and stochastic realizations of the bialgebra-selected matching geometry. In the quantum model, idempotents are represented as orthogonal internal-state projectors, giving exact caging and channel-controlled entangling routing that creates path–channel entanglement, where entangling routing means a coherent transfer that correlates distinct internal channel states with distinct spatial outputs. p. 1 In the quantum model the idempotents are represented as orthogonal internal-state projectors, giving
exact caging and channel-controlled entangling routing. Here entangling routing means a coherent transfer
that correlates distinct internal channel states with distinct spatial outputs and thereby creates path–channel
entanglement. The controlled direct-sum form is standard in quantum
information [3]; the point here is that it is selected by the same compatibility theorem that governs the graph
bracket and cobracket. In the stochastic model the graph is the configuration space of a graph Fokker–Planck
equation; its vertex-coordinate representation is a finite Markov evolution, and the selected sectors give exact
first-passage exclusion and a controlled crossover to metastable escape under weak channel switching.Where the paper says it
Where the paper says it
Transport Operators and Conservation
The canonical cobracket induces the vertex–edge identities ∆w = ∂δ, ∆2 w = 2δ † δ under the selected annihilation law, which identifies the square of the inner derivative with the positive edge form after mixed channel products have vanished. This leads to a common transport operator DG = i∆w = i∂δ = -i ad1, and its second-order composition TˆG = -ħ2/2m D2G is identified as ħ2/2m δ†δ. p. 6 Corollary 3. 9 (Common transport operator) With the conventions used below,
DG = i∆w = i∂δ = −i ad1, (43)
TˆG = −
ℏ
2
2m
D
2
G =
ℏ
2
m
δ
†
δ, (44)
L = −
1
2
∆w =
1
2
ad1, L
2 =
1
2
δ
†
δ. In the stochastic model the graph is the configuration space of a graph Fokker–Planck
equation; its vertex-coordinate representation is a finite Markov evolution, and the selected sectors give exact
first-passage exclusion and a controlled crossover to metastable escape under weak channel switching. The
applications connect the construction to graph state transfer [8, 10], first-passage and switching processes
[24, 6, 15, 9], and algebraic caging distinct from the interference mechanisms of Abelian and multicomponent
non-Abelian Aharonov–Bohm cages [27, 18].Where the paper says it
Where the paper says it
Stochastic Caging and Routing
The stochastic caging theorem states that for a fixed channel α, the effective transition graph of M(α) is contained in the colour-α matching, meaning a process initially at a vertex remains in the matching component containing at most two vertices. This leads to an exact stochastic router where conditional on C = α, P(Xτ = uβ C = α) = δαβ, which is an error-free stochastic demultiplexer. p. 0 The ordered real realization
gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable
crossover to escape under weak channel switching. These models are deliberately reducible; their
role is to exhibit the quantum and stochastic consequences of the matching geometry selected by
canonical bialgebra compatibility.Where the paper says it
Conclusion
The canonical cobracket separates the broader Jacobi-compatible clique-layer class from the incident-annihilating matching subclass, explaining why exact realizations decompose into invariant two-vertex channels rather than being imposed as an assumption, and this selection is derived from canonical bialgebra compatibility. The resulting structure yields exact quantum caging, channel-controlled path–channel entanglement, stochastic first-passage exclusion, and controlled escape under channel switching. p. 0 Representing the idempotent channels by internal-state projectors gives exact quantum caging and
channel-controlled transfer that creates path–channel entanglement. The ordered real realization
gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable
crossover to escape under weak channel switching.Where the paper says it
Acknowledgments
I acknowledge Mihaly B´ anyai for useful discussions at the very beginning of this project and I am grateful to my colleagues Andras Telcs, P´ eter ´ Erdi, Zolt ´ an Somogyv ´ ari, Tam ´ as Kiss, L´ aszl ´ o Zal ´ anyi, Marcell Stippinger and Attila Bencze for convincing me to demonstrate the applicability of the approach.
References
[1] Anderson D F and Livingston P S 1999 The zero-divisor graph of a commutative ring J. Algebra 217 434–447 (doi: 10. p. 1 Second, we prove the exact canonicalbialgebra rigidity theorem and show that, over C
q
, it selects matching layers. The resulting complex Lie
algebra decomposes channelwise into two-dimensional affine factors and central isolated directions.Where the paper says it
[2] Anderson W J 1991 Continuous-Time Markov Chains: An Applications-Oriented Approach (New York: Springer)
[3] Barenco A, Bennett C H, Cleve R, DiVincenzo D P, Margolus N, Shor P, Sleator T, Smolin J A and Weinfurter H 1995 Elementary gates for quantum computation Phys. Rev. A 52 3457–3467 (doi: 10.1103/PhysRevA.52.3457)
[4] Bazso F and L´ abos E 2006 Boolean–Lie algebras and the Leibniz rule J. Phys. A: Math. Gen. 39 6871–6876 (doi: 10.1088/0305-4470/39/22/005)
[5] Beck I 1988 Coloring of commutative rings J. Algebra 116 208–226 (doi: 10. p. 1 Second, we prove the exact canonicalbialgebra rigidity theorem and show that, over C
q
, it selects matching layers. The resulting complex Lie
algebra decomposes channelwise into two-dimensional affine factors and central isolated directions.Where the paper says it
[6] Benichou O, Loverdo C, Moreau M and Voituriez R 2011 Intermittent search strategies Rev. Mod. Phys. 83 81–129 (doi: 10.1103/RevModPhys.83.81)
[7] Berkolaiko G and Kuchment P 2013 Introduction to Quantum Graphs (Mathematical Surveys and Monographs vol 186) Providence, RI: American Mathematical Society (doi: 10.1090/surv/186)
[8] Bose S 2003 Quantum communication through an unmodulated spin chain Phys. Rev. Lett. 91 207901 (doi: 10.1103/PhysRevLett.91.207901)
[9] Bressloff P C 2024 Truncated stochastically switching processes Phys. Rev. E 109 024103 (doi: 10.1103/PhysRevE.109.024103)
[10] Christandl M, Datta N, Ekert A and Landahl A J 2004 Perfect state transfer in quantum spin networks Phys. Rev. Lett. 92 187902 (doi: 10.1103/PhysRevLett.92.187902)
[11] Chung F 2005 Laplacians and the Cheeger inequality for directed graphs Ann. Comb. 9 1–19 (doi: 10.1007/s00026-005-0237-z)
[12] Diestel R 2017 Graph Theory 5th edn Graduate Texts in Mathematics vol 173 Berlin: Springer (doi: 10.1007/978-3-662-53622-3)
[13] Dimakis A and Muller-Hoissen F 1994 Discrete differential calculus: graphs, topologies, and gauge theory J. Math. Phys. 35 6703–6735 (doi: 10.1063/1.530638)
[14] Farinati M A and Jancsa A P 2018 Lie bialgebra structures on 2-step nilpotent graph algebras J. Algebra 505 70–91 (doi: 10. p. 1 Second, we prove the exact canonicalbialgebra rigidity theorem and show that, over C
q
, it selects matching layers. The resulting complex Lie
algebra decomposes channelwise into two-dimensional affine factors and central isolated directions.Where the paper says it
[15] Grebenkov D S 2019 A unifying approach to first-passage time distributions in diffusing diffusivity and switching diffusion models J. Phys. A: Math. Theor. 52 174001 (doi: 10.1088/1751-8121/ab0dae)
[16] Kauffman L H 2004 Non-commutative worlds New J. Phys. 6 173 (doi: 10.1088/1367-2630/6/1/173)
[17] Kosmann-Schwarzbach Y 2004 Lie bialgebras, Poisson–Lie groups and dressing transformations Integrability of Nonlinear Systems (Lecture Notes in Physics vol 638) ed Y Kosmann-Schwarzbach, B Grammaticos and K M Tamizhmani Berlin: Springer pp 107–173 (doi: 10.1007/978-3-540-40962-5 5)
[18] Li S, Xue Z-Y, Gong M and Hu Y 2020 Non-Abelian Aharonov–Bohm caging in photonic lattices Phys. Rev. A 102 023524 (doi: 10.1103/PhysRevA.102.023524)
[19] Majid S 2013 Noncommutative Riemannian geometry on graphs J. Geom. Phys. 69 74–93 (doi: 10.1016/j.geomphys.2013.02.004)
[20] Moudgalya S and Motrunich O I 2022 Hilbert space fragmentation and commutant algebras Phys. Rev. X 12 011050 (doi: 10.1103/PhysRevX.12.011050)
[24] Redner S 2001 A Guide to First-Passage Processes (Cambridge: Cambridge University Press) doi: 10.17/CBO9780511606014
[25] Rovelli C and Zatloukal V 2019 Natural discrete differential calculus in physics Found. Phys. 49 693–699 (doi: 10.1007/s10701-019-00271-1)
[26] Sunada T 2008 Discrete geometric analysis Analysis on Graphs and its Applications Proc. Symp. Pure Math. vol 77 ed P Exner, J P Keating, P Kuchment, T Sunada and A Teplyaev Providence, RI: American Mathematical Society pp 51–83 (doi: 10.1090/pspum/077/2459864)
Discussion
The canonical cobracket separates the broader Jacobi-compatible clique-layer class from the incident-annihilating matching subclass, explaining why exact realizations decompose into invariant two-vertex channels rather than being imposed as an assumption. The factorization ∆w = ∂δ is of incidence/Hodge type [13, 26], but here the edge map is simultaneously a Lie cobracket compatible with the vertex bracket, while the vertex operator is generated internally as − ad1. p. 6 The identity ∆w = ∂δ has the form of the standard incidence/Hodge factorization of a graph Laplacian
[13, 26]. The additional statement is not that incidence factorization is new: here the edge map δ is also a Lie
cobracket satisfying co-Jacobi and the 1-cocycle identity, while the vertex operator is generated internally as
− ad1. The canonical cocycle condition supplies the corresponding plus-sign relations and, in a 2-torsion-free ring,
removes these remaining closed configurations. This is the local reason that Jacobi permits clique layers
whereas canonical bialgebra compatibility selects matching layers.Where the paper says it
Where the paper says it
Scope and Further Directions
Within commutative 2-torsion-free coefficient rings, the canonical cobracket cannot support non-matching Lie-bialgebra dynamics of this form, as Theorem 3.2 rules it out, and richer compatible geometries may arise from a modified cobracket or noncommutative coefficients. p. 1 Their block
reducibility is not an independent solvability ansatz: it is the consequence of the canonical bialgebra selection
theorem. Richer non-matching dynamics with a compatible coalgebra would require changing the canonical
cobracket or leaving at least one of the commutative, symmetric, 2-torsion-free hypotheses. The ordered real realization
gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable
crossover to escape under weak channel switching. These models are deliberately reducible; their
role is to exhibit the quantum and stochastic consequences of the matching geometry selected by
canonical bialgebra compatibility.Where the paper says it
Where the paper says it
Conclusions
The paper determines the compatibility hierarchy for zero-divisor-valued graph brackets, showing that over an integral domain nonzero support components are uniformly weighted cliques but over Cq Jacobi-compatible weights are superpositions of clique layers. Over a commutative 2-torsion-free ring, compatibility with the canonical cobracket is equivalent to annihilation of distinct incident weights, reducing the clique-layer family precisely to matching layers. p. 1 For the canonical cobracket
δ(ej) =
X
k
wjkej ∧ ek, (2)
and over a commutative 2-torsion-free coefficient ring, Lie-bialgebra compatibility is equivalent to the local
relation
wjkwjℓ = 0 (k , ℓ). (3)
Thus Jacobi alone permits clique layers, whereas the 1-cocycle requirement for the canonical cobracket selects
the matching-layer subclass. Representing the idempotent channels by internal-state projectors gives exact quantum caging and
channel-controlled transfer that creates path–channel entanglement. The ordered real realization
gives an intrinsic graph Fokker–Planck equation, exact first-passage exclusion, and a solvable
crossover to escape under weak channel switching.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI systems can perform exact quantum caging by using an internal-state projector representation of idempotents, which is selected by canonical bialgebra compatibility rather than being an imposed solvability assumption.
-
The system can execute channel-controlled transfer that creates path-channel entanglement, as described in the paper, by leveraging the algebraic structure derived from the graph bracket and cobracket.
-
Stochastic systems can exhibit exact first-passage exclusion and controlled escape under weak channel switching by modeling the graph as a configuration space of a graph Fokker-Planck equation derived from the second-order operator delta dagger delta.
-
The AI can model complex transport phenomena where mixed-colour path products vanish due to the zero-divisor rule, leading to deterministic internal-state dependent absorbing routing in stochastic processes.
Abstract
We show how algebraic constraints on weighted graphs yield exact localization and controlled transport in quantum and stochastic systems. Antisymmetrizing a noncommutative, nonassociative product of vertex functions defines a Lie bracket. For a specified Lie cobracket over commutative rings without nonzero 2-torsion, compatibility holds if and only if products of distinct incident edge weights vanish. When the coefficient ring is a finite product of integral domains, the edges with nonzero weights in each component of the ring form a matching. These matchings are determined by the edge weights and need not coincide across the components. These matchings yield exact two-vertex caging in fixed quantum channels; separately specified channel-selective controls enable coherent routing and path--internal-state entanglement. For matching-supported stochastic jump rates, they enforce first-passage exclusion, while prescribed internal-state switching opens inaccessible paths and yields a singular weak-switching escape limit. The bracket--cobracket factorization identifies a common edge operator underlying quantum kinetic Hamiltonians and Markov generators, subject to distinct sign and positivity conditions. The algebra fixes the channel structure; controlled motion requires additional dynamical operations.
Related papers
- Existence of a fluid phase for the two-dimensional one-component plasma
- The Huang-Yang conjecture for the low-density Fermi gas
- Anderson localisation in infinite and finite resonator crystals with long-range interactions
- Free energy and spectral edge of the SYK model
- Anderson localization for 1-d quasi-periodic Schr"odinger operators with degenerate weights
- Irregular SLE(4) martingales and isomonodromic deformations