Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:2607.08677

Twofold universality of large-NNN melonic random tensors

Carlos I. Pérez-Sánchez

Published Jul 9, 2026• 1 citation
arXivHTMLPDF

Abstract

We construct a measure that exhibits two aspects of a new type of universality and dramatically simplifies the integration of tensors $T_{a_1,a_2,\ldots,a_D} \in \mathbb{C}$ ($a_1,\ldots,a_D=1,\ldots,N$) at large $N$. In contrast to matrix integration, in which matrix traces canonically yield the integrand, tensors need additional information (equivalent to a $D$-coloured graph $B$) to contract their indices and form a tensor trace $B(T)$. We show that, whenever each $B_1,\ldots, B_n$ can be obtained by a recursive construction known as melonicity, then the leading order in $N$ of the integral of $ {B_1}(T) {B_2}(T) \cdots {B_n}(T) $ is independent of the -- often intricate -- combinatorics of the traces $B_i$, but also, to our surprise, independent of $D$ as far as $D\geq 3$. Instead, at large $N$, these integrals are some functions (indexed by $n$) of the number of vertices $2p_i$ of $B_i$ which we call melonic polynomials. Melonic traces cumulants with respect to any ('interacting') measure \[ \exp\Big\{-N^{D-1} \sum_{i=1}^m g_i {B_i}(T)\Big\} \mathrm{d}μ_0(T) \quad (g_1,\ldots,g_m \in \mathbb{R}, \mathrm{d}μ_0(T) =\text{the tensor Gaussian}) \] with each $B_i$ melonic, can be computed with our universal measure that replaces each $B_i$ by a canonical trace depending only on $p_i$. We prove that any two melonic tensor models are indistinguishable at large-$N$, independently of the number of tensor indices (first universality aspect), and of the fine-grainedness of their interactions (second universality), being a sufficient condition that the couplings (the parameters $g_i$ above) agree and their respective traces are monomials with the same degree in $T$.

DOI
10.48550/arxiv.2607.08677

Related papers

Redundant moments of non-melonic random tensor models: simplifying the tensor bootstrap

Carlos I. Pérez-Sánchez

We prove the redundancy of a family (the so-called melonic family) of large-$N$ moments in arbitrary tensor models. This reduces the number of independent entries in the positive semidefinite matrices that build the core of the tensor bootstrap (for tensor models there are already several generalizations of the positive semidefinite Toeplitz and Hankel matrices of lattice gauge theory and random matrix bootstrap, respectively). More concretely, we prove that any melonic operator $C$ at large-$N$ has an expectation value that depends on $C$ exclusively through its degree. A similar statement is shown here to hold for the Schwinger-Dyson equations of melonic moments. The strength of our result is also its scope, which is not limited to melonic tensor models.

arXiv:2607.28518•Jul 30, 2026•0 citations

Tensor models with generalized melonic interactions

Valentin Bonzom

Tensor models are natural generalizations of matrix models. The interactions and observables in the case of unitary invariant models are generalizations of matrix traces. Some notable interactions in the literature include the melonic ones, the tetrahedral one as well as the planar ones in rank three, or necklaces in even ranks. Here we introduce generalized melonic interactions which generalize the melonic and necklace interactions. We characterize them as tree-like gluings of quartic interactions. We also completely characterize the Feynman graphs which contribute to the large $N$ limit. For a subclass of generalized melonic interactions called totally unbalanced interactions, we prove that the large $N$ limit is Gaussian and therefore the Feynman graphs are in bijection with trees. This result further extends the class of tensor models which fall into the Gaussian universality class. Another key aspect of tensor models with generalized melonic interactions is that they can be written as matrix models without increasing the number of degrees of freedom of the original tensor models. In the case of totally unbalanced interactions, this new matrix model formulation in fact decreases the number of degrees of freedom, meaning that some of the original degrees of freedom are effectively integrated. We then show how the large $N$ Gaussian behavior can be reproduced using a saddle point analysis on those matrix models.

arXiv:1905.01903•May 6, 2019•7 citations

Universality for random tensors and cycle graphs with multiple edges

Nana Kanbe

We consider the universality for the trace invariants of $c_1 N \times \cdots \times c_D N$ tensors with i.i.d. complex random elements. In the case $c_1 = \cdots = c_D$, Gurau derived the universality in the limit $N \to \infty$ by representing the average trace invariants in terms of the corresponding colored graphs. Moreover he could explicitly calculate the asymptotic forms of the average trace invariants, if the corresponding graphs were special ones called "melonic graphs". In this paper we study another kind of special graphs, cycle graphs with multiple edges. One can construct these graphs by using simple cycle graphs and melonic graphs. Counting the number of the components, we can explicitly calculate the asymptotic forms of the average trace invariants.

arXiv:1603.08707•Mar 29, 2016•0 citations

Melonic Large N Limit of 5-Index Irreducible Random Tensors

Sylvain Carrozza, Sabine Harribey

Abstract We demonstrate that random tensors transforming under rank-5 irreducible representations of $$\mathrm {O}(N)$$ O ( N ) can support melonic large N expansions. Our construction is based on models with sextic (5-simplex) interaction, which generalize previously studied rank-3 models with quartic (tetrahedral) interaction (Benedetti et al. in Commun Math Phys 371:55, 2019. arXiv:1712.00249 ; Carrozza in JHEP 06:039, 2018. arXiv:1803.02496 ). Beyond the irreducible character of the representations, our proof relies on recursive bounds derived from a detailed combinatorial analysis of the Feynman graphs. Our results provide further evidence that the melonic limit is a universal feature of irreducible tensor models in arbitrary rank.

arXiv:2104.03665•Communications in Mathematical Physics•Jan 18, 2022•19 citations

The SYK model and random tensors: Gaussian universality

Thomas Krajewski, Matteo Laudonio, Romain Pascalie +1 more

The SYK model involves N Majorana fermions in 1+0 dimensions with quenched Gaussian dis- order and proves to be exactly solvable in the large N limit at strong coupling. It has been initially proposed by Sachdev and Ye as a model of condensed matter and later gained some interest as a toy model of AdS/CFT correspondence, thanks to the work of Kitaev. On the other side, random tensors are generalisations of random matrices to objets that carry more than two indices. It turns out that the SYK model and random tensors involve a special class of Feynman graphs known as "melons". We will briefly review both constructions. Then, we will show how non Gaussian disorder can be reduced to a Gaussian one, treating the coupling as a random tensor, thanks to Gurau’s Gaussian universality result.

Corfu Summer Institute School and Workshops on Elementary Particle Physics and Gravity (CORFU)•Aug 18, 2020•1 citation

Tensor models from the viewpoint of matrix models: the cases of loop models on random surfaces and of the Gaussian distribution

Valentin Bonzom, Frédéric Combes

Two direct connections between random tensors and random matrices are discussed in this article. In the rst part, we introduce U( \tau ) matrix models which generate fully packed, oriented loops on random surfaces. e latter are found to be in bijection with a set of regular edge-colored graphs. It is shown that the expansion in the number of loops is organized like the 1/ N expansion of rank-three tensor models. Recent results on tensor models are applied in this context. For example, congurations which maximize the number of loops are precisely the melonic graphs of tensor models and a scaling limit which projects onto themelonic sector is found. is approach is generalized to higher-rank tensor models, which generate loops with fugacity \tau on triangulations in dimension d–1 . In the second part, we introduce singular value decompositions to evaluate the expectations of polynomial observables of Gaussian random tensors. Performing the integrals over the unitary group leads to a notion of eective observables which expand onto regular trace invariants. We show that both asymptotic and exact new calculations of expectations can be performed this way.

Annales de l’Institut Henri Poincaré D Combinatorics Physics and their Interactions•Apr 8, 2015•12 citations
We construct a measure that exhibits two aspects of a new type of universality and dramatically simplifies the integration of tensors Ta1,a2,…,aD∈CT_{a_1,a_2,\ldots,a_D} \in \mathbb{C}Ta1​,a2​,…,aD​​∈C (a1,…,aD=1,…,Na_1,\ldots,a_D=1,\ldots,Na1​,…,aD​=1,…,N) at large NNN. In contrast to matrix integration, in which matrix traces canonically yield the integrand, tensors need additional information (equivalent to a DDD-coloured graph BBB) to contract their indices and form a tensor trace B(T)B(T)B(T). We show that, whenever each B1,…,BnB_1,\ldots, B_nB1​,…,Bn​ can be obtained by a recursive construction known as melonicity, then the leading order in NNN of the integral of B1(T)B2(T)⋯Bn(T){B_1}(T) {B_2}(T) \cdots {B_n}(T)B1​(T)B2​(T)⋯Bn​(T) is independent of the -- often intricate -- combinatorics of the traces BiB_iBi​, but also, to our surprise, independent of DDD as far as D≥3D\geq 3D≥3. Instead, at large NNN, these integrals are some functions (indexed by nnn) of the number of vertices 2pi2p_i2pi​ of BiB_iBi​ which we call melonic polynomials. Melonic traces cumulants with respect to any ('interacting') measure [ \exp\Big{-N^{D-1} \sum_{i=1}^m g_i {B_i}(T)\Big} \mathrm{d}μ_0(T) \quad (g_1,\ldots,g_m \in \mathbb{R}, \mathrm{d}μ_0(T) =\text{the tensor Gaussian}) ] with each BiB_iBi​ melonic, can be computed with our universal measure that replaces each BiB_iBi​ by a canonical trace depending only on pip_ipi​. We prove that any two melonic tensor models are indistinguishable at large-NNN, independently of the number of tensor indices (first universality aspect), and of the fine-grainedness of their interactions (second universality), being a sufficient condition that the couplings (the parameters gig_igi​ above) agree and their respective traces are monomials with the same degree in TTT.
We prove the redundancy of a family (the so-called melonic family) of large-NNN moments in arbitrary tensor models. This reduces the number of independent entries in the positive semidefinite matrices that build the core of the tensor bootstrap (for tensor models there are already several generalizations of the positive semidefinite Toeplitz and Hankel matrices of lattice gauge theory and random matrix bootstrap, respectively). More concretely, we prove that any melonic operator CCC at large-NNN has an expectation value that depends on CCC exclusively through its degree. A similar statement is shown here to hold for the Schwinger-Dyson equations of melonic moments. The strength of our result is also its scope, which is not limited to melonic tensor models.
Tensor models are natural generalizations of matrix models. The interactions and observables in the case of unitary invariant models are generalizations of matrix traces. Some notable interactions in the literature include the melonic ones, the tetrahedral one as well as the planar ones in rank three, or necklaces in even ranks. Here we introduce generalized melonic interactions which generalize the melonic and necklace interactions. We characterize them as tree-like gluings of quartic interactions. We also completely characterize the Feynman graphs which contribute to the large NNN limit. For a subclass of generalized melonic interactions called totally unbalanced interactions, we prove that the large NNN limit is Gaussian and therefore the Feynman graphs are in bijection with trees. This result further extends the class of tensor models which fall into the Gaussian universality class. Another key aspect of tensor models with generalized melonic interactions is that they can be written as matrix models without increasing the number of degrees of freedom of the original tensor models. In the case of totally unbalanced interactions, this new matrix model formulation in fact decreases the number of degrees of freedom, meaning that some of the original degrees of freedom are effectively integrated. We then show how the large NNN Gaussian behavior can be reproduced using a saddle point analysis on those matrix models.
We consider the universality for the trace invariants of c1N×⋯×cDNc_1 N \times \cdots \times c_D Nc1​N×⋯×cD​N tensors with i.i.d. complex random elements. In the case c1=⋯=cDc_1 = \cdots = c_Dc1​=⋯=cD​, Gurau derived the universality in the limit N→∞N \to \inftyN→∞ by representing the average trace invariants in terms of the corresponding colored graphs. Moreover he could explicitly calculate the asymptotic forms of the average trace invariants, if the corresponding graphs were special ones called "melonic graphs". In this paper we study another kind of special graphs, cycle graphs with multiple edges. One can construct these graphs by using simple cycle graphs and melonic graphs. Counting the number of the components, we can explicitly calculate the asymptotic forms of the average trace invariants.
Abstract We demonstrate that random tensors transforming under rank-5 irreducible representations of O(N)\mathrm {O}(N)O(N) O ( N ) can support melonic large N expansions. Our construction is based on models with sextic (5-simplex) interaction, which generalize previously studied rank-3 models with quartic (tetrahedral) interaction (Benedetti et al. in Commun Math Phys 371:55, 2019. arXiv:1712.00249 ; Carrozza in JHEP 06:039, 2018. arXiv:1803.02496 ). Beyond the irreducible character of the representations, our proof relies on recursive bounds derived from a detailed combinatorial analysis of the Feynman graphs. Our results provide further evidence that the melonic limit is a universal feature of irreducible tensor models in arbitrary rank.
The SYK model involves N Majorana fermions in 1+0 dimensions with quenched Gaussian dis- order and proves to be exactly solvable in the large N limit at strong coupling. It has been initially proposed by Sachdev and Ye as a model of condensed matter and later gained some interest as a toy model of AdS/CFT correspondence, thanks to the work of Kitaev. On the other side, random tensors are generalisations of random matrices to objets that carry more than two indices. It turns out that the SYK model and random tensors involve a special class of Feynman graphs known as "melons". We will briefly review both constructions. Then, we will show how non Gaussian disorder can be reduced to a Gaussian one, treating the coupling as a random tensor, thanks to Gurau’s Gaussian universality result.
Two direct connections between random tensors and random matrices are discussed in this article. In the rst part, we introduce U( \tau ) matrix models which generate fully packed, oriented loops on random surfaces. e latter are found to be in bijection with a set of regular edge-colored graphs. It is shown that the expansion in the number of loops is organized like the 1/ N expansion of rank-three tensor models. Recent results on tensor models are applied in this context. For example, congurations which maximize the number of loops are precisely the melonic graphs of tensor models and a scaling limit which projects onto themelonic sector is found. is approach is generalized to higher-rank tensor models, which generate loops with fugacity \tau on triangulations in dimension d–1 . In the second part, we introduce singular value decompositions to evaluate the expectations of polynomial observables of Gaussian random tensors. Performing the integrals over the unitary group leads to a notion of eective observables which expand onto regular trace invariants. We show that both asymptotic and exact new calculations of expectations can be performed this way.