arXiv subjects

Jonas Maziero

Publications and source records attributed to Jonas Maziero.

At least 37 records · Page 2Linked to original sources

arXiv2025

MixFunn: A Neural Network for Differential Equations with Improved Generalization and Interpretability

We introduce MixFunn, a novel neural network architecture designed to solve differential equations with enhanced precision, interpretability, and generalization capability. The architecture comprises two key components: the mixed-function neuron, which integrates multiple parameterized nonlinear functions to improve representational flexibility, and the second-order neuron, which combines a linear transformation of its inputs with a quadratic term to capture cross-combinations of input variables. These features significantly enhance the expressive power of the network, enabling it to achieve comparable or superior results with drastically fewer parameters and a reduction of up to four orders of magnitude compared to conventional approaches. We applied MixFunn in a physics-informed setting to solve differential equations in classical mechanics, quantum mechanics, and fluid dynamics, demonstrating its effectiveness in achieving higher accuracy and improved generalization to regions outside the training domain relative to standard machine learning models. Furthermore, the architecture facilitates the extraction of interpretable analytical expressions, offering valuable insights into the underlying solutions.

cs.LG↗

arXiv2025

Learning to learn with an evolutionary strategy applied to variational quantum algorithms

Variational Quantum Algorithms (VQAs) employ parameterized quantum circuits optimized using classical methods to minimize a cost function. While VQAs have found broad applications, certain challenges persist. Notably, a significant computational burden arises during parameter optimization. The prevailing ``parameter shift rule'' mandates a double evaluation of the cost function for each parameter. In this article, we introduce a novel optimization approach named ``Learning to Learn with an Evolutionary Strategy'' (LLES). LLES unifies ``Learning to Learn'' and ``Evolutionary Strategy'' methods. ``Learning to Learn'' treats optimization as a learning problem, utilizing recurrent neural networks to iteratively propose VQA parameters. Conversely, ``Evolutionary Strategy'' employs gradient searches to estimate function gradients. Our optimization method is applied to two distinct tasks: determining the ground state of an Ising Hamiltonian and training a quantum neural network. The obtained results underscore the efficacy of this novel approach. Additionally, we identify a key hyperparameter that significantly influences gradient estimation using the ``Evolutionary Strategy'' method.

quant-ph↗

arXiv2025

Quantum fluctuation theorem in a curved spacetime

The interplay between thermodynamics, general relativity and quantum mechanics has long intrigued researchers. Recently, important advances have been obtained in thermodynamics, mainly regarding its application to the quantum domain through fluctuation theorems. In this letter, we apply Fermi normal coordinates to report a fully general relativistic detailed quantum fluctuation theorem based on the two point measurement scheme. We demonstrate how the spacetime curvature can produce entropy in a localized quantum system moving in a general spacetime. The example of a quantum harmonic oscillator living in an expanding universe is presented. This result implies that entropy production is strongly observer dependent and deeply connects the arrow of time with the causal structure of the spacetime.

gr-qc↗

arXiv2024

An Updated Quantum Complementarity Principle

Bohr's complementarity principle has long been a fundamental concept in quantum mechanics, positing that, within a given experimental setup, a quantum system (or quanton) can exhibit either its wave-like character, denoted as $W$, or its particle-like character, denoted as $P$, but not both simultaneously. Modern interpretations of Bohr's complementarity principle acknowledge the coexistence of these aspects in the same experiment while introducing the constraint $W + P \le α$. Notably, estimations of $W$ or $P$ frequently rely on indirect retrodiction methods, a practice that has led to the claim of the violation of Bohr's complementarity principle. By taking a different route, recent advancements demonstrate that quantum complementarity relations can be rigorously derived from the axioms of quantum mechanics. To reconcile these observations and eliminate potential paradoxes or violations, we propose an updated formulation for the quantum complementarity principle, which is stated as follows: \textit{For a given quantum state preparation $ρ_t$ at a specific instant of time $t$, the wave and particle behaviors of a quanton are constrained by a complementarity relation $\mathfrak{W}(ρ_t) + \mathfrak{P}(ρ_t) \le α(d)$, which is derived directly from the axioms of quantum mechanics.}

quant-ph↗

arXiv2024

Non-binary artificial neuron with phase variation implemented on a quantum computer

The first artificial quantum neuron models followed a similar path to classic models, as they work only with discrete values. Here we introduce an algorithm that generalizes the binary model manipulating the phase of complex numbers. We propose, test, and implement a neuron model that works with continuous values in a quantum computer. Through simulations, we demonstrate that our model may work in a hybrid training scheme utilizing gradient descent as a learning algorithm. This work represents another step in the direction of evaluation of the use of artificial neural networks efficiently implemented on near-term quantum devices.

quant-ph↗

arXiv2024

Hybrid model of the kernel method for quantum computers

The field of quantum machine learning is a promising way to lead to a revolution in intelligent data processing methods. In this way, a hybrid learning method based on classic kernel methods is proposed. This proposal also requires the development of a quantum algorithm for the calculation of internal products between vectors of continuous values. In order for this to be possible, it was necessary to make adaptations to the classic kernel method, since it is necessary to consider the limitations imposed by the Hilbert space of the quantum processor. As a test case, we applied this new algorithm to learn to classify whether new points generated randomly, in a finite square located under a plane, were found inside or outside a circle located inside this square. It was found that the algorithm was able to correctly detect new points in 99% of the samples tested, with a small difference due to considering the radius slightly larger than the ideal. However, the kernel method was able to perform classifications correctly, as well as the internal product algorithm successfully performed the internal product calculations using quantum resources. Thus, the present work represents a contribution to the area, proposing a new model of machine learning accessible to both physicists and computer scientists.

quant-ph↗

arXiv2024

Using quantum computers to identify prime numbers via entanglement dynamics

Recently, the entanglement dynamics of two harmonic oscillators initially prepared in a separable-coherent state was demonstrated to offer a pathway for prime number identification. This article presents a generalized approach and outlines a deterministic algorithm making possible the implementation of this theoretical concept on scalable fault-tolerant qubit-based quantum computers. We prove that the diagonal unitary operations employed in our algorithm exhibit a polynomial-time complexity of degree two, contrasting with the previously reported exponential complexity of general diagonal unitaries.

quant-ph↗

arXiv2024

Efficient fidelity estimation: Alternative derivation and related applications

In [Phys. Rev. A 107, 012427 (2023)], A. J. Baldwin and J. A. Jones proved that Uhlmann-Jozsa's fidelity between two quantum states $ρ$ and $σ$, i.e., $F(ρ,σ)~:=~(Tr\sqrt{\sqrtρσ\sqrtρ})^2$, can be written in a simplified form as $F(ρ,σ) = (Tr\sqrt{ρσ})^2$. In this article, we give an alternative proof of this result, using a function power series expansion and the properties of the trace function. Our approach not only reinforces the validity of the simplified expression but also facilitates the exploration of novel dissimilarity functions for quantum states and more complex trace functions of a density operator.

quant-ph↗

arXiv2024

A differentiable programming framework for spin models

We introduce a novel framework for simulating spin models using differentiable programming, an approach that leverages the advancements in machine learning and computational efficiency. We focus on three distinct spin systems: the Ising model, the Potts model, and the Cellular Potts model, demonstrating the practicality and scalability of our framework in modeling these complex systems. Additionally, this framework allows for the optimization of spin models, which can adjust the parameters of a system by a defined objective function. In order to simulate these models, we adapt the Metropolis-Hastings algorithm to a differentiable programming paradigm, employing batched tensors for simulating spin lattices. This adaptation not only facilitates the integration with existing deep learning tools but also significantly enhances computational speed through parallel processing capabilities, as it can be implemented on different hardware architectures, including GPUs and TPUs.

cond-mat.stat-mech↗

arXiv2024

Restricting to the chip architecture maintains the quantum neural network accuracy

In the era of noisy intermediate-scale quantum devices, variational quantum algorithms (VQAs) stand as a prominent strategy for constructing quantum machine learning models. These models comprise both a quantum and a classical component. The quantum facet is characterized by a parametrization $U$, typically derived from the composition of various quantum gates. On the other hand, the classical component involves an optimizer that adjusts the parameters of $U$ to minimize a cost function $C$. Despite the extensive applications of VQAs, several critical questions persist, such as determining the optimal gate sequence, devising efficient parameter optimization strategies, selecting appropriate cost functions, and understanding the influence of quantum chip architectures on the final results. This article aims to address the last question, emphasizing that, in general, the cost function tends to converge towards an average value as the utilized parameterization approaches a $2$-design. Consequently, when the parameterization closely aligns with a $2$-design, the quantum neural network model's outcome becomes less dependent on the specific parametrization. This insight leads to the possibility of leveraging the inherent architecture of quantum chips to define the parametrization for VQAs. By doing so, the need for additional swap gates is mitigated, consequently reducing the depth of VQAs and minimizing associated errors.

quant-ph↗

arXiv2023

The irreversibility of relativistic time-dilation

The fluctuation relations, which characterize irreversible processes in Nature, are among the most important results in non-equilibrium physics. In short, these relations say that it is exponentially unlikely for us to observe a time-reversed process and, thus, establish the thermodynamic arrow of time pointing from low to high entropy. On the other hand, fundamental physical theories are invariant under time-reversal symmetry. Although in Newtonian and quantum physics the emergence of irreversible processes, as well as fluctuation relations, is relatively well understood, many problems arise when relativity enters the game. In this work, by considering a specific class of spacetimes, we explore the question of how the time-dilation effect enters into the fluctuation relations. We conclude that a positive entropy production emerges as a consequence of both the special relativistic and the gravitational (enclosed in the equivalence principle) time-dilation effects.

quant-ph↗

arXiv2023

Influence of spin on tunneling times in the super-relativistic regime

For the relativistic tunneling effect described using Dirac's equation, in [Phys. Rev. A 70, 052112 (2004)] the authors presented the deduction of a general result that allows for the determination of the phase time (group delay) as the sum of the particle dwell time inside a potential barrier and of the self-interference delay associated with the incident and reflected wave functions interaction. In this article, a mathematical model is derived through a construction analogous to the proposal mentioned above, but based on an alternative representation for Dirac's equation. This representation is similar to the one introduced in [Found. Phys. 45, 1586 (2015)]. Thus, from the application of this model in the study of the tunneling effect in the absence of an external magnetic field, the influence of spin on the tunneling times is described. More specifically, the tunneling time is obtained as the sum of the dwell times inside the potential barrier for particles with spin up and spin down and the self-interaction time associated with the incident and reflected wave functions for particles with spin up.

quant-ph↗

arXiv2023

The quantum cost function concentration dependency on the parametrization expressivity

Although we are currently in the era of noisy intermediate scale quantum devices, several studies are being conducted with the aim of bringing machine learning to the quantum domain. Currently, quantum variational circuits are one of the main strategies used to build such models. However, despite its widespread use, we still do not know what are the minimum resources needed to create a quantum machine learning model. In this article, we analyze how the expressiveness of the parametrization affects the cost function. We analytically show that the more expressive the parametrization is, the more the cost function will tend to concentrate around a value that depends both on the chosen observable and on the number of qubits used. For this, we initially obtain a relationship between the expressiveness of the parametrization and the mean value of the cost function. Afterwards, we relate the expressivity of the parametrization with the variance of the cost function. Finally, we show some numerical simulation results that confirm our theoretical-analytical predictions. To the best of our knowledge, this is the first time that these two important aspects of quantum neural networks are explicitly connected.

quant-ph↗

arXiv2023

Simulating noisy quantum channels via quantum state preparation algorithms

In Refs. [Phys. Rev. A 96, 062303 (2017)] and [Sci. China Phys. Mech. Astron. 61, 70311 (2018)], the authors reported an algorithm to simulate, in a circuit-based quantum computer, a general quantum channel (QC). However, the application of their algorithm is limited because it entails the solution of intricate non-linear systems of equations in order to obtain the quantum circuit to be implemented for the simulation. Motivated by this issue, in this article we identify and discuss a simple way to implement the simulation of QCs on any $d$-level quantum system through quantum state preparation algorithms, that have received much attention in the quantum information science literature lately. We exemplify the versatility of our protocol applying it to most well known qubit QCs, to some qudit QCs, and to simulate the effect of Lorentz transformations on spin states. We also regard the application of our protocol for initial mixed states. Most of the given application examples are demonstrated using IBM's quantum computers.

quant-ph↗

arXiv2023

Evolution strategies: Application in hybrid quantum-classical neural networks

With the rapid development of quantum computers, several applications are being proposed for them. Quantum simulations, simulation of chemical reactions, solution of optimization problems and quantum neural networks (QNNs) are some examples. However, problems such as noise, limited number of qubits and circuit depth, and gradient vanishing must be resolved before we can use them to their full potential. In the field of quantum machine learning, several models have been proposed. In general, in order to train these different models, we use the gradient of a cost function with respect to the model parameters. In order to obtain this gradient, we must compute the derivative of this function with respect to the model parameters. One of the most used methods in the literature to perform this task is the parameter-shift rule method. This method consists of evaluating the cost function twice for each parameter of the QNN. A problem with this method is that the number of evaluations grows linearly with the number of parameters. In this work we study an alternative method, called Evolution Strategies (ES), which are a family of black box optimization algorithms which iteratively update the parameters using a search gradient. An advantage of the ES method is that in using it one can control the number of times the cost function will be evaluated. We apply the ES method to the binary classification task, showing that this method is a viable alternative for training QNNs. However, we observe that its performance will be strongly dependent on the hyperparameters used. Furthermore, we also observe that this method, alike the parameter shift rule method, suffers from the problem of gradient vanishing.

quant-ph↗

arXiv2023

Simulation of positive operator-valued measures and quantum instruments via quantum state preparation algorithms

In Ref. [Phys. Rev. A 100, 062317 (2019)], the authors reported an algorithm to implement, in a circuit-based quantum computer, a general quantum measurement (GQM) of a two-level quantum system, a qubit. Even though their algorithm seems right, its application involves the solution of an intricate non-linear system of equations in order to obtain the angles determining the quantum circuit to be implemented for the simulation. In this article, we identify and discuss a simple way to circumvent this issue and implement GQMs on any $d$-level quantum system through quantum state preparation algorithms. Using some examples for one qubit, one qutrit and two qubits, we illustrate the easy of application of our protocol. Besides, we show how one can utilize our protocol for simulating quantum instruments, for which we also give an example. All our examples are demonstrated using IBM's quantum processors.

quant-ph↗

arXiv2023

Quantum coherence versus interferometric visibility in a biased Mach-Zehnder interferometer

The double-slit interferometer and the Mach-Zehnder interferometer (MZI) with balanced beam splitters are prototypical setups for investigating the quantum wave-particle duality. These setups induced a quantitative association of interferometric visibility (IVI) with the wave aspect of a single quantum system (WAQ). Recently, it was realized that quantum coherence (QC) can be better suited than IVI for quantifying the WAQ in complementarity relations. In this article, we investigate a MZI with biased beam splitters both in the input and the output, and we show that in some cases the IVI is not adequate to quantify the WAQ since it does not reflect the behavior of the quantum coherence, even for a bi-dimensional closed quantum system. Using IBM quantum computers, we experimentally verify our theoretical findings by doing a full quantum simulation of the optical MZI with biased beam splitters.

quant-ph↗

arXiv2023

Quantum simulation of the generalized-entangled quantum eraser and the related complete complementarity relations

We utilize IBM's quantum computers to perform a full quantum simulation of the optical quantum eraser (QE) utilizing a Mach-Zehnder interferometer with a variable partially-polarizing beam splitter (VPPBS) at the input. The use of the VPPBS motivates us to introduce the entangled quantum eraser, for which the path information is erased using a Bell-basis measurement. We also investigate the behavior of the wave aspect, i.e., the quantum coherence, as well as the particle character, represented by the predictability and entanglement, as delineated in complete complementarity relations (CCRs). As we show in this article, the utilization of the VPPBS uncover interesting aspects of the QE and CCRs. For instance, we can recover the full wave-behavior by the erasure procedure even when we have only partial knowledge about the path through entanglement.

quant-ph↗