Simuladores quânticos têm tido um importante papel no estudo e desenvolvimento da computação quântica ao longo dos anos. A simulação de algoritmos quânticos em computadores clássicos é computacionalmente difícil, principalmente devido à natureza paralela dos sistemas quânticos. Para acelerar essas simulações, alguns trabalhos propõem usar hardware paralelo programável como FPGAs, o que diminui consideravelmente o tempo de execução. Contudo, essa abordagem tem três problemas principais: pouca escalabilidade, já que apenas transfere a complexidade do domínio do tempo para o domínio do espaço; a necessidade de re-síntese a cada mudança no algoritmo; e o esforço extra ao projetar o código RTL para simulação. Para lidar com esses problemas, uma arquitetura de um co-processador SIMD é proposta, cujas operações das portas quânticas está baseada no modelo Network of Butterflies. Com isso, eliminamos a necessidade de re-síntese com mudanças pequenas no algoritmo quântico simulado, e eliminamos a influência de um dos fatores que levam ao crescimento exponencial do uso de recursos da FPGA. Adicionamente, desenvolvemos uma ferramenta para geração automática do código RTL sintetizável do co-processador, reduzindo assim o esforço extra de projeto. / Quantum simulators have had a important role on the studying and development of quantum computing throughout the years. The simulation of quantum algorithms on classical computers is computationally hard, mainly due to the parallel nature of quantum systems. To speed up these simulations, some works have proposed to use programmable parallel hardware such as FPGAs, which considerably shorten the execution time. However this approach has three main problems: low scalability, since it only transfers the complexity from time domain to space domain; the need of re-synthesis on every change on the algorithm; and the extra effort on designing the RTL code for simulation. To handle these problems, an architecture of a SIMD co-processor is proposed, whose operations of quantum gates are based on Network of Butterflies model. Thus, we eliminate the need of re-synthesis on small changes on the simulated quantum algorithm, and we eliminated the influence of one of the factors that lead to the exponential growth on the consumption of FPGA resources. Aditionally, we developed a tool to automatically generate the synthesizable RTL code of the co-processor, thus reducing the extra design effort.

[pt] A dissertação apresenta detalhadamente o algoritmo de fatoração de Shor, tanto em termos de sua execução passo a passo como mediante sua representação em forma de circuito, abordando aspectos tanto de sua parte clássica como de sua parte quântica. Inicialmente são apresentados aspectos de teoria dos números indispensáveis para a compreensão do algoritmo e em seguida são desenvolvidos conceitos e propriedades de mecânica quântica e de informação quântica pertinentes. Em atenção ao caráter eminentemente estocástico do algoritmo realiza-se um estudo de sua fonte estocástica e demonstram-se os principais teoremas que embasam a avaliação de sua probabilidade de sucesso. Desenvolvem-se exemplos de simulação clássica do algoritmo. Finalmente, a eficiência do algoritmo de fatoração de Shor é comparada com a de algoritmos clássicos. / [en] The dissertation presents in detail Shor s factoring algorithm, including its execution step by step and its representation in the form of a circuit, addressing aspects of both its classical and its quantum parts. Aspects of number theory indispensable to understand the algorithm are presented, followed by a development of concepts and properties of quantum mechanics and quantum information. Considering the eminently stochastic character of the algorithm, a study of its stochastic source is carried out and the main theorems that support the evaluation of its probability of success are proved. Examples of classical simulation of the algorithm are developed. Finally, the efficiency of Shor s factoring algorithm is compared with that of classical algorithms.

Simuladores quânticos têm tido um importante papel no estudo e desenvolvimento da computação quântica ao longo dos anos. A simulação de algoritmos quânticos em computadores clássicos é computacionalmente difícil, principalmente devido à natureza paralela dos sistemas quânticos. Para acelerar essas simulações, alguns trabalhos propõem usar hardware paralelo programável como FPGAs, o que diminui consideravelmente o tempo de execução. Contudo, essa abordagem tem três problemas principais: pouca escalabilidade, já que apenas transfere a complexidade do domínio do tempo para o domínio do espaço; a necessidade de re-síntese a cada mudança no algoritmo; e o esforço extra ao projetar o código RTL para simulação. Para lidar com esses problemas, uma arquitetura de um co-processador SIMD é proposta, cujas operações das portas quânticas está baseada no modelo Network of Butterflies. Com isso, eliminamos a necessidade de re-síntese com mudanças pequenas no algoritmo quântico simulado, e eliminamos a influência de um dos fatores que levam ao crescimento exponencial do uso de recursos da FPGA. Adicionamente, desenvolvemos uma ferramenta para geração automática do código RTL sintetizável do co-processador, reduzindo assim o esforço extra de projeto. / Quantum simulators have had a important role on the studying and development of quantum computing throughout the years. The simulation of quantum algorithms on classical computers is computationally hard, mainly due to the parallel nature of quantum systems. To speed up these simulations, some works have proposed to use programmable parallel hardware such as FPGAs, which considerably shorten the execution time. However this approach has three main problems: low scalability, since it only transfers the complexity from time domain to space domain; the need of re-synthesis on every change on the algorithm; and the extra effort on designing the RTL code for simulation. To handle these problems, an architecture of a SIMD co-processor is proposed, whose operations of quantum gates are based on Network of Butterflies model. Thus, we eliminate the need of re-synthesis on small changes on the simulated quantum algorithm, and we eliminated the influence of one of the factors that lead to the exponential growth on the consumption of FPGA resources. Aditionally, we developed a tool to automatically generate the synthesizable RTL code of the co-processor, thus reducing the extra design effort.

In this thesis, I present a microscopic theory of quantum circuits based on interacting electron spins in quantum dot molecules. We use the Linear Combination of Harmonic Orbitals-Configuration Interaction (LCHO-CI) formalism for microscopic calculations. We then derive effective Hubbard, t-J, and Heisenberg models. These models are used to predict the electronic, spin and transport properties of a triple quantum dot molecule (TQDM) as a function of topology, gate configuration, bias and magnetic field. With these theoretical tools and fully characterized TQDMs, we propose the following applications: 1. Voltage tunable qubit encoded in the chiral states of a half-filled TQDM. We show how to perform single qubit operations by pulsing voltages. We propose the "chirality-to-charge" conversion as the measurement scheme and demonstrate the robustness of the chirality-encoded qubit due to charge fluctuations. We derive an effective qubit-qubit Hamiltonian and demonstrate the two-qubit gate. This provides all the necessary operations for a quantum computer built with chirality-encoded qubits. 2. Berry's phase. We explore the prospect of geometric quantum computing with chirality-encoded qubit. We construct a Herzberg circuit in the voltage space and show the accumulation of Berry's phase. 3. Macroscopic quantum states on a semiconductor chip. We consider a linear chain of TQDMs, each with 4 electrons, obtained by nanostructuring a metallic gate in a field effect transistor. We theoretically show that the low energy spectrum of the chain maps onto that of a spin-1 chain. Hence, we show that macroscopic quantum states, protected by a Haldane gap from the continuum, emerge. In order to minimize decoherence of electron spin qubits, we consider using electron spins in the p orbitals of the valence band (valence holes) as qubits. We develop a theory of valence hole qubit within the 4-band k.p model. We show that static magnetic fields can be used to perform single qubit operations. We also show that the qubit-qubit interactions are sensitive to the geometry of a quantum dot network. For vertical qubit arrays, we predict that there exists an optimal qubit separation suitable for the voltage control of qubit-qubit interactions.

Quantum computing has emerged as a new field that may have the potential to revolutionize the landscape of information processing and computational power, although physically constructing quantum hardware has proven difficult,and quantum computers in the current Noisy Intermediate Scale Quantum (NISQ) era are error prone and limited in the number of qubits they contain.A sub-field within quantum algorithms research which holds potential for the NISQ era, and which has seen increasing activity in recent years, is quantum machine learning, where researchers apply approaches from classical machine learning to quantum computing algorithms and explore the interplay between the two. This master thesis investigates feature selection and autoencoding algorithms for quantum computers. Our review of the prior art led us to focus on contributing to three sub-problems: A) Embedded feature selection on quantum annealers, B) short depth quantum autoencoder circuits, and C)embedded compressed feature representation for quantum classifier circuits.For problem A, we demonstrate a working example by converting ridge regression to the Quadratic Unconstrained Binary Optimization (QUBO) problem formalism native to quantum annealers, and solving it on a simulated backend. For problem B we develop a novel quantum convolutional autoencoder architecture and successfully run simulation experiments to study its performance.For problem C, we choose a classifier quantum circuit ansatz based on theoretical considerations from the prior art, and experimentally study it in parallel with a classical benchmark method for the same classification task,then show a method from embedding compressed feature representation onto that quantum circuit. / Kvantberäkning är ett framväxande område som potentiellt kan revolutionera informationsbehandling och beräkningskraft. Dock är praktisk konstruktion av kvantdatorer svårt, och nuvarande kvantdatorer i den s.k. NISQ-eran lider av fel och begränsningar i antal kvantbitar de kan hantera. Ett lovande delområde inom kvantalgoritmer är kvantmaskininlärning, där forskare tillämpar klassiska maskininlärningsmetoder på kvantalgoritmer och utforskar samspelet mellande två områdena.. Denna avhandling fokuserar på kvantalgoritmer för funktionsval,och datakompression (i form av s.k. “autoencoders”). Vi undersöker tre delproblem: A) Inbäddat funktionsval på en kvantannealer, B) autoencoder-kvantkretsar för datakompression, och C) inbyggt funktionsval för kvantkretsar för klassificering. För problem A demonstrerar vi ett fungerande exempel genom att omvandla ridge regression till problemformuleringen "Quadratic Unconstrained Binary Optimization"(QUBO) som är nativ för kvantannealers,och löser det på en simulerad backend. För problem B utvecklar vi en ny konvolutionerande autoencoder-kvantkrets-arkitektur och utför simuleringsexperimentför att studera dess prestanda. För problem C väljer vi en kvantkrets-ansats för klassificering baserad på teoretiska överväganden från tidigare forskning och studerar den experimentellt parallellt med en klassisk benchmark-metod församma klassificeringsuppgift, samt visar en metod för inbyggt funktionsval (i form av datakompression) i denna kvantkrets.

Le sujet de thèse est une étude théorique des modes normaux d’oscillations plasma dans des chaînes de jonctions Josephson supra-conductrices. Les propriétés de ces modes normaux peuvent être contrôlés en choisissant une modulation spatiale appropriée de paramètres des jonctions le long de la chaîne et/ou un couplage approprié à l'environnement extérieur. Le travail théorique au sein du LPMMC se fait en étroite collaboration avec l'équipe expérimentale"Cohérence Quantique" à l'Institut Néel. Les problèmes spécifiques étudiés dans la thèse sont : modélisation détaillée du couplage des modes normaux à l'environnement pour leur caractérisation dans une expérience de transmission de micro-ondes, dissipation intrinsèque des oscillations du plasma à cause de quasi-particules hors équilibre, l'optimisation de la structure spatiale de la chaîne de jonctions Josephson pour son utilisation en tant qu'une super-inductance. / The subject of thesis is a theorerical study of normal modes of plasma oscillations in superconducting Josephson junction chains. The properties of these normal modes can be controlled by choosing an appropriate spatial modulation of the junction parameters along the chain and/or an appropriate coupling to the external environment. The theoretical work at LPMMC is performed in a close collaboration with the experimental Quantum Coherence group at Néel Institute. The specific problems studied in this thesis are : detailed modeling of the normal mode coupling to the environment for probing them in a microwave transmission experiment, intrinsic dissipation of plasma oscillations due to the presence of non-equilibrium quasi-particles, optimization of the spatial structure of the Josephson junction chain for its use as a super-inductance.

L'informatique quantique exploite les phénomènes de la théorie quantique pour le traitement de l'information, tandis que l'apprentissage automatique s'intéresse aux algorithmes qui peuvent s'améliorer en fonction des expériences passées. L'informatique quantique a produit des algorithmes qui dépassent de loin les capacités des ordinateurs classiques que nous utilisons tous les jours. Cependant, l'identification de nouveaux algorithmes quantiques fut moins prolifique que dans le cas classique. Ces dernières années, on a cherché à combiner l'informatique quantique et l'apprentissage automatique. Le cadre de l'apprentissage automatique a servi à apprendre les paramètres de circuits quantiques paramétrés dans l'espoir d'apprendre à résoudre des problèmes où les phénomènes quantiques peuvent aider grâce au traitement de l'information quantique. L'objectif principal de ce mémoire est de pousser plus loin cette idée d'apprentissage de circuits quantiques et de fonder solidement ses capacités en développant une architecture universelle de circuit quantique paramétré. La première contribution est une évaluation d'algorithmes d'optimisation itératifs actuels pour les circuits quantiques paramétrés en tant que modèles d'apprentissage automatique, ainsi que la présentation d'un algorithme d'optimisation itératif simple, mais robuste. La deuxième contribution est une architecture de circuit quantique dans laquelle une famille de petits circuits avec des connexions arbitraires peut être intégrée. / Quantum information processing leverages the phenomena of quantum theory for information processing, while machine learning concerns itself with algorithms that can improve based on past experiences. Quantum information processing has produced algorithms that go far past the capabilities of the classical computers we use every day. However, the identification of new quantum algorithms has been much slower than during the early days of classical computing. In recent years, there has been a push to combine quantum information processing and machine learning. The framework of machine learning has been used to learn quantum circuits in the hopes of learning to solve problems where quantum phenomena can help through the use of quantum information processing. The main goal of this thesis is to further push this idea of learning quantum circuits and to solidly ground its capabilities by developing a learnable parametrized universal quantum circuit. The first contribution is an assessment of current optimization methods for parametrized quantum circuits as machine learning models. The second contribution is a quantum circuit architecture in which a family of smaller circuits with arbitrary connections can be embedded.

Les amplificateurs paramétriques Josephson (JPA) se sont révélés être un outil indispensablepour l’étude expérimentale de dispositifs quantiques dans le régime micro-onde ; car ilsrajoutent uniquement le minimum de bruit imposé par la mécanique quantique[1]. Cependant,ces amplificateurs sont beaucoup plus difficiles à utiliser et optimiser que leurs homologuesclassiques. Récemment, plusieurs expériences réalisées avec des circuits supraconducteurs,composés d’une jonction Josephson polarisée en tension en série avec un résonateur microonde,ont montré qu’une paire de Cooper peut traverser la barrière de la jonction par effettunnel en émettant un ou plusieurs photons avec une énergie totale de 2e fois la tensionappliquée. Dans cette thèse, nous montrerons qu’un tel circuit permet de mettre en place unamplificateur micro-onde préservant la phase que nous appelons « Amplificateur basé sur letunneling inélastique de paires de Cooper » (ICTA). Il est alimenté par une tension continueet peut fonctionner avec un bruit très proche de la limite quantique.Nous commencerons en présentant le principe du fonctionnement de l’ICTA. Par analogieavec la théorie quantique des JPAs[2], nous avons étudié les performances de cet amplificateurcomme le gain, la bande passante et le bruit. Ensuite, nous présenterons la premièrepreuve expérimentale d’une amplification proche de la limite quantique sans utilisation d’unepompe micro-onde externe, mais simplement d’une tension continue dans une configurationextrêmement simple. Ces mesures ont été faites sur des échantillons avec des jonctionsen aluminium, dénommés ICTA de première génération. Selon nos résultats théoriques etexpérimentaux, nous avons conçu des circuits hyperfréquences où l’impédance présentéeà la jonction dépend de fréquences spécifiques afin d’optimiser les performances de notreamplificateur. Ces échantillons, dénommés ICTA de seconde génération, ont été fabriquésavec du nitrure de niobium. Une amélioration significative du gain et du bruit a été prouvée.Un tel amplificateur, alimenté par une simple tension continue, pourrait rendre la mesurede signaux micro-ondes au niveau du photon unique beaucoup plus faciles et permettred’intégrer plusieurs amplificateurs sur une seule puce. Il pourrait donc être un élémentimportant pour la lecture de qubit dans les processeurs quantiques à grande échelle. / Josephson parametric amplifiers (JPA), have proven to be an indispensable tool for awide range of experiments on quantum devices in the microwave frequency regime, becausethey provide the lowest possible noise. However, JPAs remain much more difficult to use andoptimize than conventional microwave amplifiers. Recent experiments with superconductingcircuits consisting of a DC voltage-biased Josephson junction in series with a resonator,showed that a tunneling Cooper pair can emit one or several photons with a total energyof 2e times the applied voltage. In this thesis we show that such q circuit can be used toimplement a new type of phase preserving microwave amplifier, which we call InelasticCooper pair Tunneling Amplifier (ICTA). It is powered by a simple DC bias and offers nearquantum-limited noise performance.We start this work by presenting a brief and simple picture of the basic ICTA operatingprinciples. In analogy with the quantum theory of JPAs, we calculate the performances ofthis amplifier such as the gain, bandwidth and noise. Then, we present the first experimentalproof that amplification close to the quantum limit is possible without microwave drive inan extremely simple setup. These measurements are made on a first generation of samplesbased on aluminium junctions. According to our theoretical and experimental results, wehave designed microwave circuits presenting specific frequency-dependent impedances tothe junction in order to optimize the performances of our amplifier. This second generationof ICTA samples is fabricated from niobium nitride and provide significantly lower noiseand higher gain.We expect that once fully optimized, such an amplifier, powered by simple DC voltagescould then make measuring microwave signals at the single photon level much easier andallow to deploy many amplifiers on a chip. It could therefore be an important ingredient forqubit readout in large-scale quantum processors.

Submitted by Johnny Rodrigues (johnnyrodrigues@ufcg.edu.br) on 2018-08-06T19:17:03Z No. of bitstreams: 1 HERON ARAGÃO MONTEIRO - DISSERTAÇÃO PPGCC 2012..pdf: 15948168 bytes, checksum: e445512265f530700a45c3924f68aa02 (MD5) / Made available in DSpace on 2018-08-06T19:17:03Z (GMT). No. of bitstreams: 1 HERON ARAGÃO MONTEIRO - DISSERTAÇÃO PPGCC 2012..pdf: 15948168 bytes, checksum: e445512265f530700a45c3924f68aa02 (MD5) Previous issue date: 2012-05-31 / Com o avanço da nanotecnologia, a computação quântica tem recebido grande destaque no meio científico. Utilizando os fundamentos da mecânica quântica, têm sido propostos diversos algoritmos quânticos. E, até então, os mesmos têm apresentado ganhos significativos com relação às suas versões clássicas. Na intenção de poder ser verificada a eficiência dos algoritmos quânticos, diversos simuladores vêm sendo desenvolvidos, visto que a confecção de um computador quântico ainda não foi possível. Há duas grandes vertentes de simuladores: os simuladores por software e os simuladores por hardware, chamados de emuladores. Na primeira classe se encontram os programas desenvolvidos em um computador clássico, procurando implementar os fundamentos da mecânica quântica, fazendo uso das linguagens de programação clássicas. Na segunda, são utilizados recursos que não estejam vinculados à plataforma do computador clássico. Dentre os emuladores, particularmente, estudos têm sido realizados fazendo uso de hardware dedicado (mais especificamente, FPGAV). O presente trabalho propõem a verificação da real utilidade da plataforma FPGA, com a intenção de se desenvolver um emulador universal, que permita a emulação de qualquer classe de circuitos, e que os mesmos possam ser implementados com um maior número de q-bits em relação aos circuitos tratados nos trabalhos anteriores. / With the progress of nanotechnology, quantum computing has received great emphasis in scientific circles. Using the basis of quantum mechanics, different quantum algorithms have been proposed. And so far, they have presented significant gains with respect to its classic versions. In order to verify the efficiency of quantum algorithms, several simulators have been developed, since the construction of a quantum computer is not yet possible. There are two major classes of simulators, simulators via software and via hardware. The latter being also called emulators. In the first class, programs are developed in a classical computer, attempting to implement the fundamentais of quantum mechanics, making use of classic programming languages. In the second, resources are used that are not related to the classic computer platform. Among the emulators, in particular, studies have been made using dedicated hardware (more specifically, FPGA's2). The present work proposes the use of the FPGA boards in emulation of quantum circuits aiming a gain scale in relation to the alternatives presented so far. The present work proposes checking the usefulness of the FPGA with the intention of developing an universal emulator that is able to emulate any type of circuit, and that they can be implemented with a larger number of q-bit in respect to the circuits treated in the previous works.

Many quantum computation algorithms, and processes like measurement based quantum computing, require the initial state of the quantum computer to be an eigenstate of a specific unitary operator. Here we study how quantum states that are eigenstates of finite dimensional irreducible representations of the special unitary (SU(d)) and the permutation (S_n) groups can be efficiently constructed in the computational basis formed by tensor products of the qudit states. The procedure is a unitary transform, which first uses Schur-Weyl duality to map every eigenstate to a unique Schur basis state, and then recursively uses the Clebsch - Gordan transform to rotate the Schur basis state to the computational basis. We explicitly provide an efficient quantum algorithm, and the corresponding quantum logic circuit, to generate any desired eigenstate of SU(2) and SU(3) irreducible representations in the computational basis.