-
Assessment of Uncertainty Quantification in Universal Differential Equations
Authors:
Nina Schmid,
David Fernandes del Pozo,
Willem Waegeman,
Jan Hasenauer
Abstract:
Scientific Machine Learning is a new class of approaches that integrate physical knowledge and mechanistic models with data-driven techniques for uncovering governing equations of complex processes. Among the available approaches, Universal Differential Equations (UDEs) are used to combine prior knowledge in the form of mechanistic formulations with universal function approximators, like neural ne…
▽ More
Scientific Machine Learning is a new class of approaches that integrate physical knowledge and mechanistic models with data-driven techniques for uncovering governing equations of complex processes. Among the available approaches, Universal Differential Equations (UDEs) are used to combine prior knowledge in the form of mechanistic formulations with universal function approximators, like neural networks. Integral to the efficacy of UDEs is the joint estimation of parameters within mechanistic formulations and the universal function approximators using empirical data. The robustness and applicability of resultant models, however, hinge upon the rigorous quantification of uncertainties associated with these parameters, as well as the predictive capabilities of the overall model or its constituent components. With this work, we provide a formalisation of uncertainty quantification (UQ) for UDEs and investigate important frequentist and Bayesian methods. By analysing three synthetic examples of varying complexity, we evaluate the validity and efficiency of ensembles, variational inference and Markov chain Monte Carlo sampling as epistemic UQ methods for UDEs.
△ Less
Submitted 13 June, 2024;
originally announced June 2024.
-
Topological signatures of a p-wave superconducting wire through light
Authors:
Frederick Del Pozo,
Karyn Le Hur
Abstract:
We show how the $\mathbb{Z}_{2}$ topological index of a one-dimensional topological p-wave superconductor can be revealed when driving with a classical vector potential i.e. an electromagnetic wave, through the quasiparticles inter-band transition rates. As a function of driving frequency $ω$, it is possible to obtain a measure of this topological invariant from the resonance envelope classifying…
▽ More
We show how the $\mathbb{Z}_{2}$ topological index of a one-dimensional topological p-wave superconductor can be revealed when driving with a classical vector potential i.e. an electromagnetic wave, through the quasiparticles inter-band transition rates. As a function of driving frequency $ω$, it is possible to obtain a measure of this topological invariant from the resonance envelope classifying the two distinct topological phases of the short-range Kitaev wire. We also propose to probe the topological phase transition in the model through the responses of the global capacitance in the presence of the light field and also through the Josephson current between the wire and the proximity coupled bulk superconductor. The system may also be implemented on the Bloch sphere allowing alternative ways to measure the $\mathbb{Z}$ and $\mathbb{Z}_2$ topological invariants through circuit or cavity quantum electrodynamics.
△ Less
Submitted 31 May, 2024; v1 submitted 25 January, 2024;
originally announced January 2024.
-
Fractional Topology in Interacting 1D Superconductors
Authors:
Frederick del Pozo,
Loïc Herviou,
Karyn Le Hur
Abstract:
We investigate the topological phases of two one-dimensional (1D) interacting superconducting wires and propose topological markers directly measurable from ground state correlation functions. These quantities remain powerful tools in the presence of couplings and interactions. We show with the density matrix renormalization group that the double critical Ising (DCI) phase discovered in [1] is a f…
▽ More
We investigate the topological phases of two one-dimensional (1D) interacting superconducting wires and propose topological markers directly measurable from ground state correlation functions. These quantities remain powerful tools in the presence of couplings and interactions. We show with the density matrix renormalization group that the double critical Ising (DCI) phase discovered in [1] is a fractional topological phase with gapless Majorana modes in the bulk, and a one-half topological invariant per wire. Using both numerics and quantum field theoretical methods, we show that the phase diagram remains stable in the presence of an inter-wire hop** amplitude $t_{\bot}$ at length scales below $\sim 1/t_{\bot}$. A large inter-wire hop** amplitude results in the emergence of two integer topological phases, stable also at large interactions. They host one edge mode per boundary shared between both wires. At large interactions, the two wires are described by Mott physics, with the $t_{\bot}$ hop** amplitude resulting in a paramagnetic order.
△ Less
Submitted 16 March, 2023; v1 submitted 10 October, 2022;
originally announced October 2022.
-
Emergence of network features from multiplexity
Authors:
Alessio Cardillo,
Jesús Gómez-Gardeñes,
Massimiliano Zanin,
Miguel Romance,
David Papo,
Francisco del Pozo,
Stefano Boccaletti
Abstract:
Many biological and man-made networked systems are characterized by the simultaneous presence of different sub-networks organized in separate layers, with links and nodes of qualitatively different types. While during the past few years theoretical studies have examined a variety of structural features of complex networks, the outstanding question is whether such features are characterizing all si…
▽ More
Many biological and man-made networked systems are characterized by the simultaneous presence of different sub-networks organized in separate layers, with links and nodes of qualitatively different types. While during the past few years theoretical studies have examined a variety of structural features of complex networks, the outstanding question is whether such features are characterizing all single layers, or rather emerge as a result of coarse-graining, i.e. when going from the multilayered to the aggregate network representation. Here we address this issue with the help of real data. We analyze the structural properties of an intrinsically multilayered real network, the European Air Transportation Multiplex Network in which each commercial airline defines a network layer. We examine how several structural measures evolve as layers are progressively merged together. In particular, we discuss how the topology of each layer affects the emergence of structural properties in the aggregate network.
△ Less
Submitted 4 March, 2013; v1 submitted 10 December, 2012;
originally announced December 2012.
-
Topological Measure Locating the Effective Crossover between Segregation and Integration in a Modular Network
Authors:
A. Ajdari Rad,
I. Sendiña-Nadal,
D. Papo,
M. Zanin,
J. M. Buldú,
F. del Pozo,
S. Boccaletti
Abstract:
We introduce an easily computable topological measure which locates the effective crossover between segregation and integration in a modular network. Segregation corresponds to the degree of network modularity, while integration is expressed in terms of the algebraic connectivity of an associated hyper-graph. The rigorous treatment of the simplified case of cliques of equal size that are gradually…
▽ More
We introduce an easily computable topological measure which locates the effective crossover between segregation and integration in a modular network. Segregation corresponds to the degree of network modularity, while integration is expressed in terms of the algebraic connectivity of an associated hyper-graph. The rigorous treatment of the simplified case of cliques of equal size that are gradually rewired until they become completely merged, allows us to show that this topological crossover can be made to coincide with a dynamical crossover from cluster to global synchronization of a system of coupled phase oscillators. The dynamical crossover is signaled by a peak in the product of the measures of intra-cluster and global synchronization, which we propose as a dynamical measure of complexity. This quantity is much easier to compute than the entropy (of the average frequencies of the oscillators), and displays a behavior which closely mimics that of the dynamical complexity index based on the latter. The proposed toplogical measure simultaneously provides information on the dynamical behavior, sheds light on the interplay between modularity vs total integration and shows how this affects the capability of the network to perform both local and distributed dynamical tasks.
△ Less
Submitted 15 June, 2012;
originally announced June 2012.
