-
Non-reciprocal interactions and high-dimensional chaos: comparing dynamics and statistics of equilibria in a solvable model
Authors:
Samantha J. Fournier,
Alessandro Pacco,
Valentina Ros,
Pierfrancesco Urbani
Abstract:
We investigate a model of high-dimensional dynamical variables with all-to-all interactions that are random and non-reciprocal. We characterize its phase diagram and point out that the model can exhibit chaotic dynamics. We show that the equations describing the system's dynamics exhibit a number of equilibria that is exponentially large in the dimensionality of the system, and these equilibria ar…
▽ More
We investigate a model of high-dimensional dynamical variables with all-to-all interactions that are random and non-reciprocal. We characterize its phase diagram and point out that the model can exhibit chaotic dynamics. We show that the equations describing the system's dynamics exhibit a number of equilibria that is exponentially large in the dimensionality of the system, and these equilibria are all linearly unstable in the chaotic phase. Solving the effective equations governing the dynamics in the infinite-dimensional limit, we determine the typical properties (magnetization, overlap) of the configurations belonging to the attractor manifold. We show that these properties cannot be inferred from those of the equilibria, challenging the expectation that chaos can be understood purely in terms of the numerous unstable equilibria of the dynamical equations. We discuss the dependence of this scenario on the strength of non-reciprocity in the interactions. These results are obtained through a combination of analytical methods such as Dynamical Mean-Field Theory and the Kac-Rice formalism.
△ Less
Submitted 26 March, 2025;
originally announced March 2025.
-
Triplets of local minima in a high-dimensional random landscape: Correlations, clustering, and memoryless activated jumps
Authors:
Alessandro Pacco,
Alberto Rosso,
Valentina Ros
Abstract:
We compute the distribution of triplets of stationary points in the energy landscape of the spherical p-spin model, by evaluating the quenched three-point complexity by means of the Kac-Rice formalism. We show the occurrence of transitions in the organization of stationary points in the landscape, identifying regions where local minima and saddles accumulate and cluster around other stationary poi…
▽ More
We compute the distribution of triplets of stationary points in the energy landscape of the spherical p-spin model, by evaluating the quenched three-point complexity by means of the Kac-Rice formalism. We show the occurrence of transitions in the organization of stationary points in the landscape, identifying regions where local minima and saddles accumulate and cluster around other stationary points, thus displaying the presence of correlations in the landscape. We discuss the implications of these findings for the dynamical exploration of the energy landscape in the activated regime, specifying conditions under which transitions between local minima are expected to exhibit correlated rates and when, conversely, activated jumps are likely to be memoryless.
△ Less
Submitted 11 March, 2025; v1 submitted 23 October, 2024;
originally announced October 2024.
-
Curvature-driven pathways interpolating between stationary points: the case of the pure spherical 3-spin model
Authors:
Alessandro Pacco,
Giulio Biroli,
Valentina Ros
Abstract:
This paper focuses on characterizing the energy profile along pathways connecting different regions of configuration space in the context of a prototypical glass model, the pure spherical $p$-spin model with $p=3$. The study investigates pairs of stationary points (local minima or rank-1 saddles), analyzing the energy profile along geodesic paths and comparing them with "perturbed" pathways correl…
▽ More
This paper focuses on characterizing the energy profile along pathways connecting different regions of configuration space in the context of a prototypical glass model, the pure spherical $p$-spin model with $p=3$. The study investigates pairs of stationary points (local minima or rank-1 saddles), analyzing the energy profile along geodesic paths and comparing them with "perturbed" pathways correlated to the landscape curvature. The goal is to assess the extent to which information from the local Hessian matrices around stationary points can identify paths with lower energy barriers. Surprisingly, unlike findings in other systems, the direction of softest local curvature is not a reliable predictor of low-energy paths, except in the case in which the direction of softest curvature corresponds to an isolated mode of the Hessian. However, other information encoded in the local Hessian does allow the identification of pathways associated with lower energy barriers. We conclude commenting on implications for the system's activated dynamics.
△ Less
Submitted 30 November, 2023;
originally announced November 2023.
-
Overlaps between eigenvectors of spiked, correlated random matrices: from matrix PCA to random Gaussian landscapes
Authors:
Alessandro Pacco,
Valentina Ros
Abstract:
We consider pairs of GOE (Gaussian Orthogonal Ensemble) matrices which are correlated with each others, and subject to additive and multiplicative rank-one perturbations. We focus on the regime of parameters in which the finite-rank perturbations generate outliers in the spectrum of the matrices. We investigate the statistical correlation (i.e., the typical overlap) between the eigenvectors associ…
▽ More
We consider pairs of GOE (Gaussian Orthogonal Ensemble) matrices which are correlated with each others, and subject to additive and multiplicative rank-one perturbations. We focus on the regime of parameters in which the finite-rank perturbations generate outliers in the spectrum of the matrices. We investigate the statistical correlation (i.e., the typical overlap) between the eigenvectors associated to the outlier eigenvalues of each matrix in the pair, as well as the typical overlap between the outlier eigenvector of one matrix with the eigenvectors in the bulk of the spectrum of the other matrix. We discuss implications of these results for the signal recovery problem for spiked matrices, as well as for problems of high-dimensional random landscapes.
△ Less
Submitted 9 May, 2023; v1 submitted 7 March, 2023;
originally announced March 2023.
-
Manufacturing high-Q superconducting α-tantalum resonators on silicon wafers
Authors:
D. P. Lozano,
M. Mongillo,
X. Piao,
S. Couet,
D. Wan,
Y. Canvel,
A. M. Vadiraj,
Ts. Ivanov,
J. Verjauw,
R. Acharya,
J. Van Damme,
F. A. Mohiyaddin,
J. Jussot,
P. P. Gowda,
A. Pacco,
B. Raes,
J. Van de Vondel,
I. P. Radu,
B. Govoreanu,
J. Swerts,
A. Potočnik,
K. De Greve
Abstract:
The performance of state-of-the-art superconducting quantum devices is currently limited by microwave dielectric losses at different surfaces and interfaces. α-tantalum is a superconductor that has proven effective in reducing dielectric loss and improving device performance due to its thin low-loss oxide. However, without the use of a seed layer, this tantalum phase has so far only been realised…
▽ More
The performance of state-of-the-art superconducting quantum devices is currently limited by microwave dielectric losses at different surfaces and interfaces. α-tantalum is a superconductor that has proven effective in reducing dielectric loss and improving device performance due to its thin low-loss oxide. However, without the use of a seed layer, this tantalum phase has so far only been realised on sapphire substrates, which is incompatible with advanced processing in industry-scale fabrication facilities. Here, we demonstrate the fabrication of high-quality factor α-tantalum resonators directly on silicon wafers over a variety of metal deposition conditions and perform a comprehensive material and electrical characterization study. By comparing experiments with simulated resonator loss, we demonstrate that two-level-system loss is dominated by surface oxide contributions and not the substrate-metal interface. Our study paves the way to large scale manufacturing of low-loss superconducting circuits and to materials-driven advancements in superconducting circuit performance.
△ Less
Submitted 30 November, 2022; v1 submitted 29 November, 2022;
originally announced November 2022.
-
Learning Gaussian Mixtures with Generalised Linear Models: Precise Asymptotics in High-dimensions
Authors:
Bruno Loureiro,
Gabriele Sicuro,
Cédric Gerbelot,
Alessandro Pacco,
Florent Krzakala,
Lenka Zdeborová
Abstract:
Generalised linear models for multi-class classification problems are one of the fundamental building blocks of modern machine learning tasks. In this manuscript, we characterise the learning of a mixture of $K$ Gaussians with generic means and covariances via empirical risk minimisation (ERM) with any convex loss and regularisation. In particular, we prove exact asymptotics characterising the ERM…
▽ More
Generalised linear models for multi-class classification problems are one of the fundamental building blocks of modern machine learning tasks. In this manuscript, we characterise the learning of a mixture of $K$ Gaussians with generic means and covariances via empirical risk minimisation (ERM) with any convex loss and regularisation. In particular, we prove exact asymptotics characterising the ERM estimator in high-dimensions, extending several previous results about Gaussian mixture classification in the literature. We exemplify our result in two tasks of interest in statistical learning: a) classification for a mixture with sparse means, where we study the efficiency of $\ell_1$ penalty with respect to $\ell_2$; b) max-margin multi-class classification, where we characterise the phase transition on the existence of the multi-class logistic maximum likelihood estimator for $K>2$. Finally, we discuss how our theory can be applied beyond the scope of synthetic data, showing that in different cases Gaussian mixtures capture closely the learning curve of classification tasks in real data sets.
△ Less
Submitted 14 December, 2021; v1 submitted 7 June, 2021;
originally announced June 2021.
-
Investigation of microwave loss induced by oxide regrowth in high-Q Nb resonators
Authors:
J. Verjauw,
A. Potočnik,
M. Mongillo,
R. Acharya,
F. Mohiyaddin,
G. Simion,
A. Pacco,
Ts. Ivanov,
D. Wan,
A. Vanleenhove,
L. Souriau,
J. Jussot,
A. Thiam,
J. Swerts,
X. Piao,
S. Couet,
M. Heyns,
B. Govoreanu,
I. Radu
Abstract:
The coherence of state-of-the-art superconducting qubit devices is predominantly limited by two-level-system defects, found primarily at amorphous interface layers. Reducing microwave loss from these interfaces by proper surface treatments is key to push the device performance forward. Here, we study niobium resonators after removing the native oxides with a hydrofluoric acid etch. We investigate…
▽ More
The coherence of state-of-the-art superconducting qubit devices is predominantly limited by two-level-system defects, found primarily at amorphous interface layers. Reducing microwave loss from these interfaces by proper surface treatments is key to push the device performance forward. Here, we study niobium resonators after removing the native oxides with a hydrofluoric acid etch. We investigate the reappearance of microwave losses introduced by surface oxides that grow after exposure to the ambient environment. We find that losses in quantum devices are reduced by an order of magnitude, with internal Q-factors reaching up to 7 $\cdot$ 10$^6$ in the single photon regime, when devices are exposed to ambient conditions for 16 min. Furthermore, we observe that Nb2O5 is the only surface oxide that grows significantly within the first 200 hours, following the extended Cabrera-Mott growth model. In this time, microwave losses scale linearly with the Nb$_2$O$_5$ thickness, with an extracted loss tangent tan$δ$ = 9.9 $\cdot$ 10$^{-3}$. Our findings are of particular interest for devices spanning from superconducting qubits, quantum-limited amplifiers, microwave kinetic inductance detectors to single photon detectors.
△ Less
Submitted 22 December, 2020; v1 submitted 19 December, 2020;
originally announced December 2020.
