Дисертації з теми "Quantum Machine Learning (QML)"
Щоб переглянути інші типи публікацій з цієї теми, перейдіть за посиланням: Quantum Machine Learning (QML).
Оформте джерело за APA, MLA, Chicago, Harvard та іншими стилями
Ознайомтеся з топ-49 дисертацій для дослідження на тему "Quantum Machine Learning (QML)".
Біля кожної праці в переліку літератури доступна кнопка «Додати до бібліографії». Скористайтеся нею – і ми автоматично оформимо бібліографічне посилання на обрану працю в потрібному вам стилі цитування: APA, MLA, «Гарвард», «Чикаго», «Ванкувер» тощо.
Також ви можете завантажити повний текст наукової публікації у форматі «.pdf» та прочитати онлайн анотацію до роботи, якщо відповідні параметри наявні в метаданих.
Переглядайте дисертації для різних дисциплін та оформлюйте правильно вашу бібліографію.
1
Huembeli, Patrick. "Machine learning for quantum physics and quantum physics for machine learning." Doctoral thesis, Universitat Politècnica de Catalunya, 2021. http://hdl.handle.net/10803/672085.
Повний текст джерелаАнотація:
Research at the intersection of machine learning (ML) and quantum physics is a recent growing field due to the enormous expectations and the success of both fields. ML is arguably one of the most promising technologies that has and will continue to disrupt many aspects of our lives. The way we do research is almost certainly no exception and ML, with its unprecedented ability to find hidden patterns in data, will be assisting future scientific discoveries. Quantum physics on the other side, even though it is sometimes not entirely intuitive, is one of the most successful physical theories and we are on the verge of adopting some quantum technologies in our daily life. Quantum many-body physics is a subfield of quantum physics where we study the collective behavior of particles or atoms and the emergence of phenomena that are due to this collective behavior, such as phases of matter. The study of phase transitions of these systems often requires some intuition of how we can quantify the order parameter of a phase. ML algorithms can imitate something similar to intuition by inferring knowledge from example data. They can, therefore, discover patterns that are invisible to the human eye, which makes them excellent candidates to study phase transitions. At the same time, quantum devices are known to be able to perform some computational task exponentially faster than classical computers and they are able to produce data patterns that are hard to simulate on classical computers. Therefore, there is the hope that ML algorithms run on quantum devices show an advantage over their classical analog.
This thesis is devoted to study two different paths along the front lines of ML and quantum physics. On one side, we study the use of neural networks (NN) to classify phases of mater in many-body quantum systems. On the other side, we study ML algorithms that run on quantum computers. The connection between ML for quantum physics and quantum physics for ML in this thesis is an emerging subfield in ML, the interpretability of learning algorithms. A crucial ingredient in the study of phase transitions with NNs is a better understanding of the predictions of the NN, to eventually infer a model of the quantum system and interpretability can assist us in this endeavor. The interpretability method that we study analyzes the influence of the training points on a test prediction and it depends on the curvature of the NN loss landscape. This further inspired an in-depth study of the loss of quantum machine learning (QML) applications which we as well will discuss.
In this thesis, we give answers to the questions of how we can leverage NNs to classify phases of matter and we use a method that allows to do domain adaptation to transfer the learned "intuition" from systems without noise onto systems with noise. To map the phase diagram of quantum many-body systems in a fully unsupervised manner, we study a method known from anomaly detection that allows us to reduce the human input to a mini mum. We will as well use interpretability methods to study NNs that are trained to distinguish phases of matter to understand if the NNs are learning something similar to an order parameter and if their way of learning can be made more accessible to humans. And finally, inspired by the interpretability of classical NNs, we develop tools to study the loss landscapes of variational quantum circuits to identify possible differences between classical and quantum ML algorithms that might be leveraged for a quantum advantage.La investigación en la intersección del aprendizaje automático (machine learning, ML) y la física cuántica es una área en crecimiento reciente debido al éxito y las enormes expectativas de ambas áreas. ML es posiblemente una de las tecnologías más prometedoras que ha alterado y seguirá alterando muchos aspectos de nuestras vidas. Es casi seguro que la forma en que investigamos no es una excepción y el ML, con su capacidad sin precedentes para encontrar patrones ocultos en los datos ayudará a futuros descubrimientos científicos. La física cuántica, por otro lado, aunque a veces no es del todo intuitiva, es una de las teorías físicas más exitosas, y además estamos a punto de adoptar algunas tecnologías cuánticas en nuestra vida diaria. La física cuántica de los muchos cuerpos (many-body) es una subárea de la física cuántica donde estudiamos el comportamiento colectivo de partículas o átomos y la aparición de fenómenos que se deben a este comportamiento colectivo, como las fases de la materia. El estudio de las transiciones de fase de estos sistemas a menudo requiere cierta intuición de cómo podemos cuantificar el parámetro de orden de una fase. Los algoritmos de ML pueden imitar algo similar a la intuición al inferir conocimientos a partir de datos de ejemplo. Por lo tanto, pueden descubrir patrones que son invisibles para el ojo humano, lo que los convierte en excelentes candidatos para estudiar las transiciones de fase. Al mismo tiempo, se sabe que los dispositivos cuánticos pueden realizar algunas tareas computacionales exponencialmente más rápido que los ordenadores clásicos y pueden producir patrones de datos que son difíciles de simular en los ordenadores clásicos. Por lo tanto, existe la esperanza de que los algoritmos ML que se ejecutan en dispositivos cuánticos muestren una ventaja sobre su analógico clásico. Estudiamos dos caminos diferentes a lo largo de la vanguardia del ML y la física cuántica. Por un lado, estudiamos el uso de redes neuronales (neural network, NN) para clasificar las fases de la materia en sistemas cuánticos de muchos cuerpos. Por otro lado, estudiamos los algoritmos ML que se ejecutan en ordenadores cuánticos. La conexión entre ML para la física cuántica y la física cuántica para ML en esta tesis es un subárea emergente en ML: la interpretabilidad de los algoritmos de aprendizaje. Un ingrediente crucial en el estudio de las transiciones de fase con NN es una mejor comprensión de las predicciones de la NN, para inferir un modelo del sistema cuántico. Así pues, la interpretabilidad de la NN puede ayudarnos en este esfuerzo. El estudio de la interpretabilitad inspiró además un estudio en profundidad de la pérdida de aplicaciones de aprendizaje automático cuántico (quantum machine learning, QML) que también discutiremos. En esta tesis damos respuesta a las preguntas de cómo podemos aprovechar las NN para clasificar las fases de la materia y utilizamos un método que permite hacer una adaptación de dominio para transferir la "intuición" aprendida de sistemas sin ruido a sistemas con ruido. Para mapear el diagrama de fase de los sistemas cuánticos de muchos cuerpos de una manera totalmente no supervisada, estudiamos un método conocido de detección de anomalías que nos permite reducir la entrada humana al mínimo. También usaremos métodos de interpretabilidad para estudiar las NN que están entrenadas para distinguir fases de la materia para comprender si las NN están aprendiendo algo similar a un parámetro de orden y si su forma de aprendizaje puede ser más accesible para los humanos. Y finalmente, inspirados por la interpretabilidad de las NN clásicas, desarrollamos herramientas para estudiar los paisajes de pérdida de los circuitos cuánticos variacionales para identificar posibles diferencias entre los algoritmos ML clásicos y cuánticos que podrían aprovecharse para obtener una ventaja cuántica.
2
De, Bonis Gianluca. "Rassegna su Quantum Machine Learning." Bachelor's thesis, Alma Mater Studiorum - Università di Bologna, 2021. http://amslaurea.unibo.it/24652/.
Повний текст джерелаАнотація:
Il Quantum Computing (QC) e il Machine Learning (ML) sono due dei settori più promettenti dell’informatica al giorno d’oggi. Il primo riguarda l’utilizzo di proprietà fisiche di sistemi quantistici per realizzare computazioni, mentre il secondo algoritmi di apprendimento automatizzati capaci di riconoscere pattern nei dati. In questo elaborato vengono esposti alcuni dei principali algoritmi di Quantum Machine Learning (QML), ovvero versioni quantistiche dei classici algoritmi di ML. Il tutto è strutturato come un’introduzione all’argomento: inizialmente viene introdotto il QC spiegandone le proprietà più rilevanti, successivamente vengono descritti gli algoritmi di QML confrontandoli con le loro controparti classiche e infine vengono discusse le principali tecnologie attuali, mostrando alcune implementazioni degli algoritmi precedentemente discussi.
3
Du, Yuxuan. "The Power of Quantum Neural Networks in The Noisy Intermediate-Scale Quantum Era." Thesis, The University of Sydney, 2021. https://hdl.handle.net/2123/24976.
Повний текст джерелаАнотація:
Machine learning (ML) has revolutionized the world in recent years. Despite the success, the huge computational overhead required by ML models makes them approach the limits of Moore’s law. Quantum machine learning (QML) is a promising way to conquer this issue, empowered by Google's demonstration of quantum computational supremacy. Meanwhile, another cornerstone in QML is validating that quantum neural networks (QNNs) implemented on the noisy intermediate-scale quantum (NISQ) chips can accomplish classification and image generation tasks. Despite the experimental progress, little is known about the theoretical advances of QNNs. In this thesis, we explore the power of QNNs to fill this knowledge gap.
First, we consider the potential advantages of QNNs in generative learning. We demonstrate that QNNs possess a stronger expressive power than that of classical neural networks in the measure of computational complexity and entanglement entropy. Moreover, we employ QNNs to tackle synthetic generation tasks with state-of-the-art performance.
Next, we propose a Grover-search based quantum classifier, which can tackle specific classification tasks with quadratic runtime speedups. Furthermore, we exhibit that the proposed scheme allows batch gradient descent optimization, which is different from previous studies. This property is crucial to train large-scale datasets.
Then, we study the capabilities and limitations of QNNs in the view of optimization theory and learning theory. The achieved results imply that a large system noise can destroy the trainability of QNNs. Meanwhile, we show that QNNs can tackle parity learning and juntas learning with provable advantages.
Last, we devise a quantum auto-ML scheme to enhance the trainability QNNs under the NISQ setting. The achieved results indicate that our proposal effectively mitigates system noise and alleviates barren plateaus for both conventional machine learning and quantum chemistry tasks.
4
Macaluso, Antonio <1990>. "A Novel Framework for Quantum Machine Learning." Doctoral thesis, Alma Mater Studiorum - Università di Bologna, 2021. http://amsdottorato.unibo.it/9791/2/Antonio_Macaluso_tesi.pdf.
Повний текст джерелаАнотація:
Quantum computation is an emerging computing paradigm with the potential to revolutionise the world of information technology. It leverages the laws of quantum mechanics to endow quantum machines with tremendous computing power, thus enabling the solution of problems impossible to address with classical devices. For this reason, the field is attracting ever-increasing attention from both academic and private sectors, and its full potential is still to be understood. This dissertation investigates how classical machine learning can benefit from quantum computing and provides several contributions to the emerging field of Quantum Machine Learning.
The idea is to provide a universal and efficient framework that can reproduce the output of a plethora of classical machine learning algorithms exploiting quantum computation’s advantages. The proposed framework is named Multiple Aggregator Quantum Algorithm (MAQA) due to its capability to combine multiple functions to solve typical supervised learning tasks. Thanks to this property, in its general formulation MAQA can be potentially adopted as the quantum counterpart of all those models falling into the scheme of aggregation of multiple functions. The theoretical design of the quantum algorithm and the corresponding circuit’s implementation are presented. As a second meaningful addition, two practical applications are illustrated: the quantum version of ensemble methods and neural networks.
The final contribution addresses the restriction to linear operations imposed by quantum mechanics. The idea is to exploit a quantum transposition of classical Splines to approximate non-linear functions, thus overcoming this limitation and introducing significant advantages in terms of computational complexity theory.
5
Rodriguez, Fernandez Carlos Gustavo. "Machine learning quantum error correction codes : learning the toric code /." São Paulo, 2018. http://hdl.handle.net/11449/180319.
Повний текст джерелаАнотація:
Orientador: Mario Leandro AolitaBanca:Alexandre Reily Rocha
Banca: Juan Felipe Carrasquilla
Resumo: Usamos métodos de aprendizagem supervisionada para estudar a decodificação de erros em códigos tóricos de diferentes tamanhos. Estudamos múltiplos modelos de erro, e obtemos figuras da eficácia de decodificação como uma função da taxa de erro de um único qubit. Também comentamos como o tamanho das redes neurais decodificadoras e seu tempo de treinamento aumentam com o tamanho do código tórico.
Abstract: We use supervised learning methods to study the error decoding in toric codes ofdifferent sizes. We study multiple error models, and obtain figures of the decoding efficacyas a function of the single qubit error rate. We also comment on how the size of thedecoding neural networks and their training time scales with the size of the toric code
Mestre
6
TACCHINO, FRANCESCO. "Digital quantum simulations and machine learning on near-term quantum processors." Doctoral thesis, Università degli studi di Pavia, 2020. http://hdl.handle.net/11571/1317093.
Повний текст джерелаАнотація:
Quantum mechanics is the gateway towards novel and potentially disruptive approaches to scientific and technical computing. In this thesis we explore, from a number of different perspectives, the effects of such strong relationship between the physical nature of information and the informational side of physical processes, with a focus on the digital quantum computing paradigm.
After an extensive introduction to the theory of universal quantum simulation techniques, we review the main achievements in the field and, in parallel, we outline the state of the art of near-term architectures for quantum information processing. We then move on to present novel and scalable procedures for the study of paradigmatic spin models on intermediate-scale noisy quantum processors. Through an innovative combination of quantum algorithms with classical post-processing and error mitigation protocols, we demonstrate in practice the full digital quantum simulation of spin-spin dynamical correlation functions, reporting experimental results obtained on superconducting cloud-accessible IBM Q devices. We also exhibit a practical use-case by successfully reproducing, from quantum computed data, cross section calculations for four-dimensional inelastic neutron scattering, a common tool employed in the analysis of molecular magnetic clusters.
The central part of the thesis is dedicated to the exploration of perspective hardware solutions for quantum computing. As it is not yet clear whether the currently dominant platforms, namely trapped ions and superconducting circuits, will eventually allow to reach the final goal of a fully-fledged architecture for general-purpose quantum information processing, the search for alternative technologies is at least as urgent as the improvement of existing ones or the development of new algorithms. After providing an overview of some recent proposals, including hybrid set-ups, we introduce quantum electromechanics as a promising candidate platform for future realizations of digital quantum simulators and we predict competitive performances for an elementary building block featuring nanomechanical qubits integrated within superconducting circuitry.
In the final part, we extend the reach of quantum information protocols beyond its traditional areas of application, and we account for the birth and rapid development of Quantum Machine Learning, a discipline aimed at establishing a productive interplay between the parallel revolutions brought about by quantum computing and artificial intelligence. In particular, we describe an original procedure for implementing, on a quantum architecture, the behavior of binary-valued artificial neurons. Formally exact and platform-independent, our data encoding and processing scheme guarantees in principle an exponential memory advantage over classical counterparts and is particularly well suited for pattern and image recognition purposes. We test our algorithm on IBM Q quantum processors, discussing possible training schemes for single nodes and reporting a proof-of-principle demonstration of a 2-layer, 3-neuron feed-forward neural network computation run on 7 active qubits. The latter is, in terms of the total size of the quantum register, one of the largest quantum neural network computation reported to date on real quantum hardware.Quantum mechanics is the gateway towards novel and potentially disruptive approaches to scientific and technical computing. In this thesis we explore, from a number of different perspectives, the effects of such strong relationship between the physical nature of information and the informational side of physical processes, with a focus on the digital quantum computing paradigm. After an extensive introduction to the theory of universal quantum simulation techniques, we review the main achievements in the field and, in parallel, we outline the state of the art of near-term architectures for quantum information processing. We then move on to present novel and scalable procedures for the study of paradigmatic spin models on intermediate-scale noisy quantum processors. Through an innovative combination of quantum algorithms with classical post-processing and error mitigation protocols, we demonstrate in practice the full digital quantum simulation of spin-spin dynamical correlation functions, reporting experimental results obtained on superconducting cloud-accessible IBM Q devices. We also exhibit a practical use-case by successfully reproducing, from quantum computed data, cross section calculations for four-dimensional inelastic neutron scattering, a common tool employed in the analysis of molecular magnetic clusters. The central part of the thesis is dedicated to the exploration of perspective hardware solutions for quantum computing. As it is not yet clear whether the currently dominant platforms, namely trapped ions and superconducting circuits, will eventually allow to reach the final goal of a fully-fledged architecture for general-purpose quantum information processing, the search for alternative technologies is at least as urgent as the improvement of existing ones or the development of new algorithms. After providing an overview of some recent proposals, including hybrid set-ups, we introduce quantum electromechanics as a promising candidate platform for future realizations of digital quantum simulators and we predict competitive performances for an elementary building block featuring nanomechanical qubits integrated within superconducting circuitry. In the final part, we extend the reach of quantum information protocols beyond its traditional areas of application, and we account for the birth and rapid development of Quantum Machine Learning, a discipline aimed at establishing a productive interplay between the parallel revolutions brought about by quantum computing and artificial intelligence. In particular, we describe an original procedure for implementing, on a quantum architecture, the behavior of binary-valued artificial neurons. Formally exact and platform-independent, our data encoding and processing scheme guarantees in principle an exponential memory advantage over classical counterparts and is particularly well suited for pattern and image recognition purposes. We test our algorithm on IBM Q quantum processors, discussing possible training schemes for single nodes and reporting a proof-of-principle demonstration of a 2-layer, 3-neuron feed-forward neural network computation run on 7 active qubits. The latter is, in terms of the total size of the quantum register, one of the largest quantum neural network computation reported to date on real quantum hardware.
7
Lukac, Martin. "Quantum Inductive Learning and Quantum Logic Synthesis." PDXScholar, 2009. https://pdxscholar.library.pdx.edu/open_access_etds/2319.
Повний текст джерелаАнотація:
Since Quantum Computer is almost realizable on large scale and Quantum Technology is one of the main solutions to the Moore Limit, Quantum Logic Synthesis (QLS) has become a required theory and tool for designing Quantum Logic Circuits. However, despite its growth, there is no any unified aproach to QLS as Quantum Computing is still being discovered and novel applications are being identified.
The intent of this study is to experimentally explore principles of Quantum Logic Synthesis and its applications to Inductive Machine Learning. Based on algorithmic approach, I first design a Genetic Algorithm for Quantum Logic Synthesis that is used to prove and verify the methods proposed in this work.
Based on results obtained from the evolutionary experimentation, I propose a fast, structure and cost based exhaustive search that is used for the design of a novel, least expensive universal family of quantum gates.
The results form both the evolutionary and heuristic search are used to formulate an Inductive Learning Approach based on Quantum Logic Synthesis with the intended application being the humanoid behavioral robotics.
The presented approach illustrates a successful algorithmic approach, where the search algorithm was able to invent/discover novel quantum circuits as well as novel principles in Quantum Logic Synthesis.
8
Orazi, Filippo. "Quantum machine learning: development and evaluation of the Multiple Aggregator Quantum Algorithm." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2022. http://amslaurea.unibo.it/25062/.
Повний текст джерелаАнотація:
Human society has always been shaped by its technology, so much that even ages and parts of our history are often named after the discoveries of that time. The growth of modern society is largely derived from the introduction of classical computers that brought us innovations like repeated tasks automatization and long-distance communication. However, this explosive technological advancement could be subjected to a heavy stop when computers reach physical limitations and the empirical law known as Moore Law comes to an end. Foreshadowing these limits and hoping for an even more powerful technology, forty years ago the branch of quantum computation was born. Quantum computation uses at its advantage the same quantum effects that could stop the progress of traditional computation and aim to deliver hardware and software capable of even greater computational power. In this context, this thesis presents the implementation of a quantum variational machine learning algorithm called quantum single-layer perceptron. We start by briefly explaining the foundation of quantum computing and machine learning, to later dive into the theoretical approach of the multiple aggregator quantum algorithms, and finally deliver a versatile implementation of the quantum counterparts of a single hidden layer perceptron. To conclude we train the model to perform binary classification using standard benchmark datasets, alongside three baseline quantum machine learning models taken from the literature. We then perform tests on both simulated quantum hardware and real devices to compare the performances of the various models.
9
Cangini, Nicolò. "Quantum Supervised Learning: Algoritmi e implementazione." Master's thesis, Alma Mater Studiorum - Università di Bologna, 2019. http://amslaurea.unibo.it/17694/.
Повний текст джерелаАнотація:
Il Quantum Computing non riguarda più soltanto la scienza della Fisica, negli ultimi anni infatti la ricerca in questo campo ha subito una notevole espansione dimostrando l'enorme potenziale di cui dispongono questi nuovi calcolatori che in un futuro prossimo potranno rivoluzionare il concetto di Computer Science così come lo conosciamo. Ad oggi, siamo già in grado di realizzare algoritmi su piccola scala eseguibili in un quantum device grazie ai quali è possibile sperimentare uno speed-up notevole (in alcuni casi esponenziale) su diversi task tipici della computazione classica. In questo elaborato vengono discusse le basi del Quantum Computing, con un focus particolare sulla possibilità di eseguire alcuni algoritmi supervisionati di Machine Learning in un quantum device per ottenere uno speed-up sostanziale nella fase di training. Oltre che una impostazione teorica del problema, vengono effettuati diversi esperimenti utilizzando le funzionalità dell'ambiente Qiskit, grazie al quale è possibile sia simulare il comportamento di un computer quantistico in un calcolatore classico, sia eseguirlo in cloud sui computer messi a disposizione da IBM.
10
Gupta, Riddhi Swaroop. "Robotic control and machine learning for the characterization and control of qubits." Thesis, University of Sydney, 2020. https://hdl.handle.net/2123/23519.
Повний текст джерелаАнотація:
Decoherence remains a major challenge in Near term, Intermediate scale Noisy Quantum (NISQ) computers. This thesis proposes techniques for characterizing classical noise correlations and performance variations in NISQ devices using single-qubit projective measurements. The central tasks of qubit state estimation and prediction are considered for measurements that are either temporally or spatially correlated. Firstly, this thesis focuses on timeseries prediction for single-qubit measurement outcomes. Focusing on repeated Ramsey measurements performed on a single qubit subject to temporally correlated dephasing, the key challenge is to predict qubit state dynamics by learning temporal noise correlations in a real-time stream of incoming measurements. Autoregressive or Fourier-based Kalman Filtering (KF) protocols are investigated for maximizing the forward prediction horizon and autoregressive approaches demonstrate superior predictive capabilities. Secondly, this thesis investigates the utility of nonlinear stochastic methods to characterize quantum systems using spatially correlated single-qubit projective measurements, via particle filters. A novel likelihood function is presented and incorporated into conventional particle filters. An adaptive procedure, `NMQA', is proposed to characterize a spatially inhomogeneous dephasing field in 2D by allocating measurements on a multi-qubit array. NMQA outperforms brute-force mapping in simulations and using experimental data. Finally, spatial prediction in 2D is compared with bivariate interpolation on a geometric arrangement of points known as the Padua points. By measuring a spatially inhomogeneous dephasing field on dedicated `sensor qubits', the objective is to predict the value of the field on unmeasured, proximal `data qubits'. The number and arrangement of sensor qubits relative to a fixed lattice of data qubits is investigated, providing insights for a general purpose predictive mapping protocol.
11
Tingley, Michael Alan. "Towards the Quantum Machine: Using Scalable Machine Learning Methods to Predict Photovoltaic Efficacy of Organic Molecules." Thesis, Harvard University, 2014. http://nrs.harvard.edu/urn-3:HUL.InstRepos:12553271.
Повний текст джерелаАнотація:
Recent advances in machine learning have resulted in an upsurge of interest in developing a “quantum machine”, a technique of simulating and predicting quantum-chemical properties on the molecular level. This paper explores the development of a large-scale quantum machine in the context of accurately and rapidly classifying molecules to determine photovoltaic efficacy through machine learning. Specifically, this paper proposes several novel representations of molecules that are amenable to learning, in addition to extending and improving existing representations. This paper also proposes and implements extensions to scalable distributed learning algorithms, in order to perform large scale molecular regression. This paper leverages Harvard’s Odyssey supercomputer in order to train various kinds of predictive algorithms over millions of molecules, and assesses cross-validated test performance of these models for predicting photovoltaic efficacy. The study suggests combinations of representations and learning models that may be most desirable in constructing a large-scale system designed to classify molecules by photovoltaic efficacy.
12
Sriarunothai, Theeraphot [Verfasser]. "Multi-qubit gates and quantum-enhanced deliberation for machine learning using a trapped-ion quantum processor / Theeraphot Sriarunothai." Siegen : Universitätsbibliothek der Universität Siegen, 2019. http://d-nb.info/1177366320/34.
Повний текст джерела
13
Mills, Matthew. "A multipolar polarisable force field method from quantum chemical topology and machine learning." Thesis, University of Manchester, 2012. https://www.research.manchester.ac.uk/portal/en/theses/a-multipolar-polarisable-force-field-method-from-quantum-chemical-topology-and-machine-learning(3fb1e55c-0d4c-4d11-932b-71706bdbeb8b).html.
Повний текст джерелаАнотація:
Force field methods are used to investigate the properties of a wide variety of chemical systems on a routine basis. The expression for the electrostatic energy typically does not take into account the anisotropic nature of the atomic electron distribution or the dependence of that distribution on the system geometry. This has been suggested as a cause of the failure of force field methods to reliably predict the behaviour of chemical systems. A method for incorporation of anisotropy and polarisation is described in this work. Anisotropy is modelled by the inclusion of multipole moments centred at atoms whose values are determined by application of the methods of Quantum Chemical Topology. Polarisation, the dependence of the electron distribution on system geometry, is modelled by training machine learning models to predict atomic multipole moments from knowledge of the nuclear positions of a system. The resulting electrostatic method can be implemented for any chemical system. An application to progressively more complex systems is reported, including small organic molecules and larger molecules of biological importance. The accuracy of the method is rigorously assessed by comparison of its predictions to exact interaction energy values. A procedure for generating transferable atomic multipole moment models is defined and tested. The electrostatic method can be combined with the empirical expressions used in force field calculations to describe total system energies by fitting parameters against ab initio conformational energies. Derivatives of the energy are given and the resulting multipolar polarisable force field can be used to perform geometry optimisation calculations. Future applications to conformational searching and problems requiring dynamic descriptions of a system are feasible.
14
Wu, Jiaxin. "Topics in Cold Atoms Related to Quantum Information Processing and A Machine Learning Approach to Condensed Matter Physics." The Ohio State University, 2019. http://rave.ohiolink.edu/etdc/view?acc_num=osu156320039156199.
Повний текст джерела
15
Bauer, Carsten [Verfasser], Simon [Gutachter] Trebst, and Achim [Gutachter] Rosch. "Simulating and machine learning quantum criticality in a nearly antiferromagnetic metal / Carsten Bauer ; Gutachter: Simon Trebst, Achim Rosch." Köln : Universitäts- und Stadtbibliothek Köln, 2020. http://d-nb.info/1228071888/34.
Повний текст джерела
16
Li, Zhenwei. "On-the-fly machine learning of quantum mechanical forces and its potential applications for large scale molecular dynamics." Thesis, King's College London (University of London), 2014. http://kclpure.kcl.ac.uk/portal/en/theses/onthefly-machine-learning-of-quantum-mechanical-forces-and-its-potential-applications-for-large-scale-molecular-dynamics(2a2f33a6-fa0c-44e3-8689-f4cf3f1c9198).html.
Повний текст джерелаАнотація:
Material simulation using molecular dynamics (MD) at the quantum mechanical (QM) accuracy level has gained great interest in the community. However, the bottleneck arising from the O(N3) scaling of QM calculation has enormously limited its investigation scope. As an approach to address this issue, in this thesis, I proposed a machine-learning (ML) MD scheme based on Bayesian inference from CPU-intensive QM force database. In this scheme, QM calculations are only performed when necessary and used to augment the ML database for more challenging prediction case. The scheme is generally transferable to new chemical situations and database completeness is never required. To achieve the maximal ML eciency, I use a symmetrically reduced internal-vector representation for the atomic congurations. Signicant speed-up factor is achieved under controllable accuracy tolerance in the MD simulation on test case of Silicon at dierent temperatures. As the database grows in conguration space, the extrapolative capability systematically increases and QM calculations are nally not needed for simple chemical processes. In the on-the-y ML force calculation scheme, sorting/selecting out the closest data congurations is used to enhance the overall eciency to scale as O(N). The potential application of this methodology for large-scale simulation (e.g. fracture, amorphous, defect), where chemical accuracy and computational eciency are required at the same time, can be anticipated. In the context of fracture simulations, a typical multi-scale system, interesting events happen near the crack tips beyond the description of classical potentials. The simulation results by machine-learning potential derived from a xed database with no enforced QM accuracy inspire a theoretical model which is further used to investigate the atomic bond breaking process during fracture propagation as well as its relation with the initialised vibration modes, crack speed, and bonding structure.
17
Zhang, Wei. "Many-Body Localization in Disordered Quantum Spin Chain and Finite-Temperature Gutzwiller Projection in Two-Dimensional Hubbard Model:." Thesis, Boston College, 2019. http://hdl.handle.net/2345/bc-ir:108695.
Повний текст джерелаАнотація:
Thesis advisor: Ziqiang . WangThe transition between many-body localized states and the delocalized thermal states is an eigenstate phase transition at finite energy density outside the scope of conventional quantum statistical mechanics. We apply support vector machine (SVM) to study the phase transition between many-body localized and thermal phases in a disordered quantum Ising chain in a transverse external field. The many-body eigenstate energy E is bounded by a bandwidth W=Eₘₐₓ-Eₘᵢₙ. The transition takes place on a phase diagram spanned by the energy density ϵ=2(Eₘₐₓ-Eₘᵢₙ)/W and the disorder strength ẟJ of the spin interaction uniformly distributed within [-ẟJ, ẟJ], formally parallel to the mobility edge in Anderson localization. In our study we use the labeled probability density of eigenstate wavefunctions belonging to the deeply localized and thermal regimes at two different energy densities (ϵ's) as the training set, i.e., providing labeled data at four corners of the phase diagram. Then we employ the trained SVM to predict the whole phase diagram. The obtained phase boundary qualitatively agrees with previous work using entanglement entropy to characterize these two phases. We further analyze the decision function of the SVM to interpret its physical meaning and find that it is analogous to the inverse participation ratio in configuration space. Our findings demonstrate the ability of the SVM to capture potential quantities that may characterize the many-body localization phase transition. To further investigate the properties of the transition, we study the behavior of the entanglement entropy of a subsystem of size L_A in a system of size L > L_A near the critical regime of the many-body localization transition. The many-body eigenstates are obtained by exact diagonalization of a disordered quantum spin chain under twisted boundary conditions to reduce the finite-size effect. We present a scaling theory based on the assumption that the transition is continuous and use the subsystem size L_A/ξ as the scaling variable, where ξ is the correlation length. We show that this scaling theory provides an effective description of the critical behavior and that the entanglement entropy follows the thermal volume law at the transition point. We extract the critical exponent governing the divergence of ξ upon approaching the transition point. We again study the participation entropy in the spin-basis of the domain wall excitations and show that the transition point and the critical exponent agree with those obtained from finite size scaling of the entanglement entropy. Our findings suggest that the many-body localization transition in this model is continuous and describable as a localization transition in the many-body configuration space. Besides the many-body localization transition driven by disorder, We also study the Coulomb repulsion and temperature driving phase transitions. We apply a finite-temperature Gutzwiller projection to two-dimensional Hubbard model by constructing a "Gutzwiller-type" density matrix operator to approximate the real interacting density matrix, which provides the upper bound of free energy of the system. We firstly investigate half filled Hubbard model without magnetism and obtain the phase diagram. The transition line is of first order at finite temperature, ending at 2 second order points, which shares qualitative agreement with dynamic mean field results. We derive the analytic form of the free energy and therefor the equation of states, which benefits the understanding of the different phases. We later extend our approach to take anti-ferromagnetic order into account. We determine the Neel temperature and explore its interesting behavior when varying the Coulomb repulsion
Thesis (PhD) — Boston College, 2019
Submitted to: Boston College. Graduate School of Arts and Sciences
Discipline: Physics
18
Mori, Yuto. "Path optimization with neural network for sign problem in quantum field theories." Doctoral thesis, Kyoto University, 2021. http://hdl.handle.net/2433/263466.
Повний текст джерела
19
Fiderer, Lukas J. [Verfasser], and Daniel [Akademischer Betreuer] Braun. "New Concepts in Quantum Metrology : Dynamics, Machine Learning, and Bounds on Measurement Precision / Lukas J. Fiderer ; Betreuer: Daniel Braun." Tübingen : Universitätsbibliothek Tübingen, 2020. http://d-nb.info/1212025334/34.
Повний текст джерела
20
Eisenhart, Andrew. "Quantum Simulations of Specific Ion Effects in Organic Solvents." University of Cincinnati / OhioLINK, 2021. http://rave.ohiolink.edu/etdc/view?acc_num=ucin1626356392775228.
Повний текст джерела
21
Pérez, Salinas Adrián. "Algorithmic strategies for seizing quantum computing." Doctoral thesis, Universitat de Barcelona, 2021. http://hdl.handle.net/10803/673255.
Повний текст джерелаАнотація:
Quantum computing is an emergent technology with prospects to solve problems nowadays intractable. For this purpose it is a requirement to build computers capable to store and control quantum systems without losing their quantum properties. However, these computers are hard to achieve, and in the near term there will only be Noisy Intermediate-Scale Quantum (NISQ) computers with limited performance. In order to seize quantum computing during the NISQ era, algorithms with low resource demands and capable to return approximate solutions are explored.
This thesis presents two different algorithmic strategies aiming to contribute to the plethora of algorithms available for NISQ devices, namely re-uploading and strategy. Each strategy takes advantage of different features of quantum computing, namely the superposition and the density of the Hilbert space in re-uploading, and entanglement among different partitions of the system in unary, to overcome a variety of obstacles. In both cases, the strategies are general and can be applied in a range of scenarios. Some examples are also provided in this thesis.
First, the re-uploading is designed as a meeting point between quantum computing and machine learning. Machine learning is a set of techniques to build computer programs capable to learn how to solve a problem through experience, without being explicitly programmed for it. Even though the re-uploading is not the first attempt to join quantum computers and machine learning, this approach has certain properties that make it different from other methods.
In particular, the re-uploading approach consists in introducing data into a classical algorithms in different stages along the process. This is a main difference with respect to standard methods, where data is uploaded at the beginning of the procedure. In the re-uploading, data is accompanied by tunable classical parameters that are optimized by a classical method according to a cost function defining the problem. The joint action of data and tunable parameters grant the quantum algorithm a great flexibility to learn a given behavior from sampling target data. The more re- uploadings are used, the better results can be obtained.
In this thesis, re-uploading is presented by means of a set of theoretical results supporting its capabilities, and simulations and experiments to benchmark its performance in a variety of problems.
The second algorithmic strategy is unary. This strategy describes a problem making use of only a small part of the available computational space. Thus, the computational capabilites of the computer are not optimal. In exchange, the operations required to execute a certain task become simpler. As a consequence, the retrieved results are more resilient to noise and decoherence, and meaningful. Therefore, a trade-off between efficiency and resillience against noise arises. NISQ computers benefit from this circumstance, especially in the case of small problems, where even quantum advantage and advantage over standard algorithms can be achieved.
In this thesis, unary is used to solve a typical problem in finance called option pricing, which is of interest for real world applications. Options are contracts to buy the right to buy/sell a given asset at certain time and price. The holder of the option will only exercise this right in case of profit. Option pricing concists in estimating this profit by handling stochastic evolution models.
This thesis aims to contribute to the growing number of algorithms available for NISQ computers and pave the way towards new quantum technologies.La computación cuántica es una tecnología emergente con potencial para resolver problemas hoy impracticables. Para ello son necesarios ordenadores capaces de mantener sistemas cuánticos y controlarlos con precisión. Sin embargo, construir estos ordenadores es complejo y a corto plazo solo habrá ordenadores pequeños afectados por el ruido y sujetos a ruido (NISQ). Para aprovechar los ordenadores NISQ se exploran algoritmos que requieran pocos recursos cuánticos mientras proporcionan soluciones aproximadas a los problemas que enfrentan. En esta tesis se estudian dos propuestas para algoritmos NISQ: re-uploading y unary. Cada estrategia busca tomar ventaja de diferentes características de la computación cuántica para superar diferentes obstáculos. Ambas estrategias son generales y aplicables en diversos escenarios. En primer lugar, re-uploading está diseñado como un puente entre la computación cuántica y el aprendizaje automático (Machine Learning). Aunque no es el primer intento de aplicar la cuántica al aprendizaje automático, re-uploading tiene ciertas características que lo distinguen de otros métodos. En concreto, re-uploading consiste en introducir datos en un algoritmo cuántico en diferentes puntos a lo largo del proceso. Junto a los datos se utilizan también parámetros optimizables clásicamente que permiten al circuito aprender cualquier comportamiento. Los resultados mejoran cuantas más veces se introducen los datos. El re-uploading cuenta con teoremas matemáticos que sustentan sus capacidades, y ha sido comprobado con éxito en diferentes situaciones tanto simuladas como experimentales. La segunda estrategia algorítmica es unary. Consiste en describir los problemas utilizando solo parte del espacio de computación disponible dentro del ordenador. Así, las capacidades computacionales del ordenador no son óptimas, pero a cambio las operaciones necesarias para una cierta tarea se simplifican. Los resultados obtenidos son resistentes al ruido, y mantienen su significado, y se produce una compensación entre eficiencia y resistencia a errores. Los ordenadores NISQ se ven beneficiados de esta situación para problemas pequeños. En esta tesis, unary se utiliza para resolver un problema tíıpico de finanzas, incluso obteniendo ventajas cuánticas en un problema aplicable al mundo real. Con esta tesis se espera contribuir al crecimiento de los algoritmos disponibles para ordenadores cuánticos NISQ y allanar el camino para las tecnologías venideras.
22
Maxwell, Peter. "FFLUX : towards a force field based on interacting quantum atoms and kriging." Thesis, University of Manchester, 2017. https://www.research.manchester.ac.uk/portal/en/theses/fflux-towards-a-force-field-based-on-interacting-quantum-atoms-and-kriging(72a8462a-6907-4f3d-82da-4c182e5a644d).html.
Повний текст джерелаАнотація:
Force fields have been an integral part of computational chemistry for decades, providing invaluable insight and facilitating the better understanding of biomolecular system behaviour. Despite the many benefits of a force field, there continue to be deficiencies as a result of the classical architecture they are based upon. Some deficiencies, such as a point charge electrostatic description instead of a multipole moment description, have been addressed over time, permitted by the ever-increasing computational power available. However, whilst incorporating such significant improvements has improved force field accuracy, many still fail to describe several chemical effects including polarisation, non-covalent interactions and secondary/tertiary structural effects. Furthermore, force fields often fail to provide consistency when compared with other force fields. In other words, no force field is reliably performing more accurately than others, when applied to a variety of related problems. The work presented herein develops a next-generation force field entitled FFLUX, which features a novel architecture very different to any other force field. FFLUX is designed to capture the relationship between geometry and energy through a machine learning method known as kriging. Instead of a series of parameterised potentials, FFLUX uses a collection of atomic energy kriging models to make energy predictions. The energies describing atoms within FFLUX are obtained from the Interacting Quantum Atoms (IQA) energy partitioning approach, which in turn derives the energies from the electron density and nuclear charges of topological atoms described by Quantum Chemical Topology (QCT). IQA energies are shown to provide a unique insight into the relationship between geometry and energy, allowing the identification of explicit atoms and energies contributing towards torsional barriers within various systems. The IQA energies can be modelled to within 2.6% accuracy, as shown for a series of small systems including weakly bound complexes. The energies also allow an interpretation of how an atom feels its surrounding environment through intra-atomic, covalent and electrostatic energetic descriptions, which typically are seen to converge within a ~7 - 8 A horizon radius around an atom or small system. These energy convergence results are particularly relevant to tackling the transferability theme within force field development. Where energies are seen to converge, a proximity limit on the geometrical description needed for a transferable energy model is defined. Finally, the FFLUX force field is validated through successfully optimising distorted geometries of a series of small molecules, to near-ab initio accuracy.
23
Österberg, Viktor. "Using Machine Learning to Develop a Quantum-Accurate Inter-Atomic Potential for Large Scale Molecular Dynamics Simulations of Iron under Earth’s Core Conditions." Thesis, KTH, Fysik, 2021. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-298848.
Повний текст джерелаАнотація:
Measurements of iron at extreme pressures do not agree on the melting temperature for conditions comparable with those believed to hold at Earth's core. To attempt to determine the stability of relevant lattices, simulations involving a huge amount of particles are needed. In this thesis, a machine learned model is trained to yield results from density functional theory. Different machine learning models are compared. The trained model is then used in molecular dynamics simulations of relevant lattices at a scale too large for density functional theory.Mätningar av järns smälttemperatur under påfrestningar jämförbara med desom tros gälla i jordens kärna överensstämmer ej. För att försöka bestämma stabiliteten av relevanta gitter krävs simulationer av enorma mängder partiklar. I denna tes tränas en maskininlärd modell att återge resultat från Täthetsfunktionalteori. Olika maskininlärningsmodeller jämförs. Den tränade modellen används sedan i Molekyldynamik-simulationer av relevanta gitter som är förstora för Täthetsfunktionalteori.
24
Zauleck, Julius Philipp Paul [Verfasser], and Regina de [Akademischer Betreuer] Vivie-Riedle. "Improving grid based quantum dynamics : from the inclusion of solvents to the utilization of machine learning / Julius Philipp Paul Zauleck ; Betreuer: Regina de Vivie-Riedle." München : Universitätsbibliothek der Ludwig-Maximilians-Universität, 2018. http://d-nb.info/1151818461/34.
Повний текст джерела
25
Linn, Hanna. "Detecting quantum speedup for random walks with artificial neural networks." Thesis, KTH, Skolan för elektroteknik och datavetenskap (EECS), 2020. http://urn.kb.se/resolve?urn=urn:nbn:se:kth:diva-289347.
Повний текст джерелаАнотація:
Random walks on graphs are an essential base for crucial algorithms for solving problems, like the boolean satisfiability problem. A speedup of random walks could improve these algorithms. The quantum version of the random walk, quantum walk, is faster than random walks in specific cases, e.g., on some linear graphs. An analysis of when the quantum walk is faster than the random walk can be accomplished analytically or by simulating both the walks on the graph. The problem arises when the graphs grow in size and connectivity. There are no known general rules for what an arbitrary graph not having explicit symmetries should exhibit to promote the quantum walk. Simulations will only answer the question for one single case, and will not provide any general rules for properties the graph should have. Using artificial neural networks (ANNs) as an aid for detecting when the quantum walk is faster on average than random walk on graphs, going from an initial node to a target node, has been done before. The quantum speedup may not be more than polynomial if the initial state of the quantum walk is purely in the initial node of the graph. We investigate starting the quantum walk in various superposition states, with an additional auxiliary node, to maybe achieve a larger quantum speedup. We suggest different ways to add the auxiliary node and select one of these schemes for use in this thesis. The superposition states examined are two stabiliser states and two magic states, inspired by the Gottesman-Knill theorem. According to this theorem, starting a quantum algorithm in a magic state may give an exponential speedup, but starting in a stabilizer state cannot give an exponential speedup, given that only gates from the Clifford group are used in the algorithm, as well as measurements are performed in the Pauli basis. We show that it is possible to train an ANN to classify graphs into what quantum walk was the fastest for various initial states of the quantum walk. The ANN classifies linear graphs and random graphs better than a random guess. We also show that a convolutional neural network (CNN) with a deeper architecture than earlier proposed for the task, is better at classifying the graphs than before. Our findings pave the way for automated research in novel quantum walk-based algorithms.Slumpvandringar på grafer är essensiella i viktiga algoritmer för att lösa olika problem, till exempel SAT, booleska uppfyllningsproblem (the satisfiability problem). Genom att göra slumpvandringar snabbare går det att förbättra dessa algoritmer. Kvantversionen av slumpvandringar, kvantvandringar, har visats vara snabbare än klassiska slumpvandringar i specifika fall, till exempel på vissa linjära grafer. Det går att analysera, analytiskt eller genom att simulera vandringarna på grafer, när kvantvandringen är snabbare än slumpvandingen. Problem uppstår dock när graferna blir större, har fler noder samt fler kanter. Det finns inga kända generella regler för vad en godtycklig graf, som inte har några explicita symmetrier, borde uppfylla för att främja kvantvandringen. Simuleringar kommer bara besvara frågan för ett enda fall. De kommer inte att ge några generella regler för vilka egenskaper grafer borde ha. Artificiella neuronnät (ANN) har tidigare används som hjälpmedel för att upptäcka när kvantvandringen är snabbare än slumpvandingen på grafer. Då jämförs tiden det tar i genomsnitt att ta sig från startnoden till slutnoden. Dock är det inte säkert att få kvantacceleration för vandringen om initialtillståndet för kvantvandringen är helt i startnoden. I det här projektet undersöker vi om det går att få en större kvantacceleration hos kvantvandringen genom att starta den i superposition med en extra nod. Vi föreslår olika sätt att lägga till den extra noden till grafen och sen väljer vi en för att använda i resen av projektet. De superpositionstillstånd som undersöks är två av stabilisatortillstånden och två magiska tillstång. Valen av dessa tillstånd är inspirerat av Gottesmann- Knill satsen. Enligt satsen så kan en algoritm som startar i ett magiskt tillstånd ha en exponetiell uppsnabbning, men att starta i någon stabilisatortillstånden inte kan ha det. Detta givet att grindarna som används i algoritmen är från Cliffordgruppen samt att alla mätningar är i Paulibasen. I projektet visar vi att det är möjligt att träna en ANN så att den kan klassificera grafer utifrån vilken kvantvandring, med olika initialtillstånd, som var snabbast. Artificiella neuronnätet kan klassificera linjära grafer och slumpmässiga grafer bättre än slumpen. Vi visar också att faltningsnätverk med en djupare arkitektur än tidigare föreslaget för uppgiften är bättre på att klassificera grafer än innan. Våra resultat banar vägen för en automatiserad forskning i nya kvantvandringsbaserade algoritmer.
26
Glasser, Ivan [Verfasser], Ignacio [Akademischer Betreuer] Cirac, Nora [Gutachter] Brambilla, and Ignacio [Gutachter] Cirac. "Tensor networks, conformal fields and machine learning: applications in the description of quantum many-body systems / Ivan Glasser ; Gutachter: Nora Brambilla, Ignacio Cirac ; Betreuer: Ignacio Cirac." München : Universitätsbibliothek der TU München, 2018. http://d-nb.info/1173899057/34.
Повний текст джерела
27
Pronobis, Wiktor Verfasser], Klaus-Robert [Akademischer Betreuer] [Gutachter] [Müller, Alexandre [Gutachter] Tkatchenko, and Manfred [Gutachter] Opper. "Towards more efficient and performant computations in quantum chemistry with machine learning / Wiktor Pronobis ; Gutachter: Klaus-Robert Müller, Alexandre Tkatchenko, Manfred Opper ; Betreuer: Klaus-Robert Müller." Berlin : Technische Universität Berlin, 2020. http://d-nb.info/1208764470/34.
Повний текст джерела
28
Pronobis, Wiktor [Verfasser], Klaus-Robert [Akademischer Betreuer] [Gutachter] Müller, Alexandre [Gutachter] Tkatchenko, and Manfred [Gutachter] Opper. "Towards more efficient and performant computations in quantum chemistry with machine learning / Wiktor Pronobis ; Gutachter: Klaus-Robert Müller, Alexandre Tkatchenko, Manfred Opper ; Betreuer: Klaus-Robert Müller." Berlin : Technische Universität Berlin, 2020. http://d-nb.info/1208764470/34.
Повний текст джерела
29
August, Moritz [Verfasser], Thomas [Akademischer Betreuer] Huckle, José Miguel [Gutachter] Hernández-Lobato, Steffen J. [Gutachter] Glaser, and Thomas [Gutachter] Huckle. "Tensor networks and machine learning for approximating and optimizing functions in quantum physics / Moritz August ; Gutachter: José Miguel Hernández-Lobato, Steffen J. Glaser, Thomas Huckle ; Betreuer: Thomas Huckle." München : Universitätsbibliothek der TU München, 2018. http://d-nb.info/1175091804/34.
Повний текст джерела
30
Perea, Ospina Jose Dario [Verfasser], Salvador León [Akademischer Betreuer] Cabanillas, and Christoph J. [Gutachter] Brabec. "Solubility and Miscibility of Organic Semiconductors for Efficient and Stable Organic Solar Cells Investigated via Machine Learning and Quantum Chemistry Methods / Jose Dario Perea Ospina ; Gutachter: Christoph J. Brabec ; Betreuer: Salvador León Cabanillas." Erlangen : Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), 2019. http://d-nb.info/1184575215/34.
Повний текст джерела
31
Théveniaut, Hugo. "Méthodes d'apprentissage automatique et phases quantiques de la matière." Thesis, Toulouse 3, 2020. http://www.theses.fr/2020TOU30228.
Повний текст джерелаАнотація:
Mon travail de thèse s'est articulé autour de trois manières d'utiliser les méthodes d'apprentissage automatique (machine learning) en physique de la matière condensée. Premièrement, j'expliquerai comment il est possible de détecter automatiquement des transitions de phase en reformulant cette tâche comme un problème de classification d'images. J'ai testé la fiabilité et relevé les limites de cette approche dans des modèles présentant des phases localisées à N corps (many-body localized - MBL) en dimension 1 et en dimension 2. Deuxièmement, j'introduirai une représentation variationnelle d'états fondamentaux sous la forme de réseaux de neurones (neural-network quantum states - NQS). Je présenterai nos résultats sur un modèle contraint de bosons de coeur dur en deux dimensions avec des méthodes variationnelles basées sur des NQS et de projection guidée. Nos travaux montrent notamment que les états NQS peuvent encoder avec précision des états solides et liquides de bosons. Enfin, je présenterai une nouvelle approche pour la recherche de stratégies de corrections d'erreur dans les codes quantiques, cette approche se base sur les techniques utilisées pour concevoir l'intelligence artificielle AlphaGo. Nous avons pu montrer que des stratégies efficaces peuvent être découvertes avec des algorithmes d'optimisation évolutionnistes. En particulier, nous avons observé que des réseaux de neurones peu profonds sont compétitifs avec les réseaux profonds utilisés dans des travaux antérieurs, représentant un gain d'un facteur 10000 en termes de nombre de paramètresMy PhD thesis presents three applications of machine learning to condensed matter theory. Firstly, I will explain how the problem of detecting phase transitions can be rephrased as an image classification task, paving the way to the automatic mapping of phase diagrams. I tested the reliability of this approach and showed its limits for models exhibiting a many-body localized phase in 1 and 2 dimensions. Secondly, I will introduce a variational representation of quantum many-body ground-states in the form of neural-networks and show our results on a constrained model of hardcore bosons in 2d using variational and projection methods. In particular, we confirmed the phase diagram obtained independently earlier and extends its validity to larger system sizes. Moreover we also established the ability of neural-network quantum states to approximate accurately solid and liquid bosonic phases of matter. Finally, I will present a new approach to quantum error correction based on the same techniques used to conceive the best Go game engine. We showed that efficient correction strategies can be uncovered with evolutionary optimization algorithms, competitive with gradient-based optimization techniques. In particular, we found that shallow neural-networks are competitive with deep neural-networks
32
Hoffmann, Guillaume. "Mise au point de nouveaux descripteurs théoriques pour la réactivité chimique Can molecular and atomic descriptors predict the electrophilicity of Michael acceptors? On the influence of dynamical effects on reactivity descriptors Predicting experimental electrophilicities from quantum and topological descriptors : a machine learning approach Electrophilicity indices and halogen bonds : some new alternatives to the molecular electrostatic potential." Thesis, Normandie, 2020. http://www.theses.fr/2020NORMR042.
Повний текст джерелаАнотація:
L’étude des descripteurs de réactivité globaux, locaux et non locaux d'un système réactif est d’une importance capitale afin de comprendre la réactivité de la totalité des processus chimiques lors d’une réaction. Le but de cette thèse a ainsi été de mettre au point de nouveaux descripteurs de réactivités, ainsi que des modèles de prédiction basés sur ces derniers, afin d’étudier la réactivité chimique. Les principale méthodes théoriques employées ont été la Théorie de la fonctionnelle de la densité conceptuelle (CDFT) et la théorie quantique « Atoms in Molecules » (QTAIM) qui sont toutes deux basées sur la densité électronique. Notre domaine d’étude se place principalement dans le cadre de l’échelle expérimentale de Mayr, qui permet par le biais de mesures cinétiques d’effectuer un classement des molécules par ordre de réactivité. Dans un premier temps, de grande avancées ont été réalisées durant cette thèse vis-à-vis de la prédiction théorique de l’électrophilie expérimentale des accepteurs de Michael. Puis dans un second temps, nous nous sommes intéressées à l’application des descripteurs de réactivité sur la liaison chimique, et particulièrement la liaison halogène. Enfin, une partie de synthèse réalisée au cours de cette thèse est présentée, en proposant une nouvelle voie de synthèse des cations iminium vinyloguesThe study of global, local and non-local reactivity descriptors of a reactive system is of paramount importance in order to understand the reactivity of all chemical processes during a reaction. The goal of this thesis was then to develop new reactivity descriptors, as well as prediction models based on them, in order to study chemical reactivity. The main theoretical methods used were the Conceptual Density Functional Theory (CDFT) and Quantum Theory of “Atoms in Molecules” (QTAIM), which are both based on electron density. Our field of study is mainly within the framework of the Mayr experimental scale, which allows, through kinetic measurements a classification of molecules in order of reactivity. In the first part, great advances were made during this thesis with respect to the theoretical prediction of experimental electrophilicity of Michael acceptors. Then in a second step, we looked at the application of reactivity descriptors on the chemical bond, and in particular the halogen bond. Finally, a part of synthesis carried out during the course of this thesis is presented, by proposing a new way of synthesis of vinylogous iminium cations
33
Leelar, Bhawani Shankar. "Machine Learning Algorithms Using Classical And Quantum Photonics." Thesis, 2017. http://etd.iisc.ac.in/handle/2005/4303.
Повний текст джерелаАнотація:
ABSTRACT
In the modern day , we are witnessing two complementary trends, exponential growth in data and shrinking of chip size. The Data is approaching to 44 zettabytes by 2020 and the chips are now available with 10nm technology. The hyperconnectivity between machine-to-machine and humanto- machine creates multi-dimensional data which is more complex. Our thesis addresses the quantum meta layer abstraction which provides the interface to the Application layer to design quantum and classical algorithms. The first part of the thesis addresses the quantum algorithms and second part address classical algorithms running on top of quantum meta layer. In the first part of our thesis we explored quantum stochastic algorithm for ranking Quantum Webpages, analogous to the classical Google PageRank. The architecture is a six-waveguide photonic lattice that runs finely-tuned quantum stochastic walk. The evolution of density matrix
solves the ranking of quantum webpages. We force the photon stochastic walk for quantum PageRank by matching the entries of Google matrix with parameters of the Kossakowski-Lindblad master equation. We have done extensive simulation to observe the density matrix evolution with different parameter settings. We have used noise in the Kossakowski-Lindblad master equation to break the symmetry (reciprocity) property of quantum system, which helps in distinguishable measurement of the quantum PageRank. We next propose a new quantum deep learning with photonic lattice waveguide as a feedforward neural network. The proposed deep photonic neural network uses the quantum properties for learning. The hidden layers of our deep photonic neural network can be designed to learn object representation and mentains the quantum quantum properties for longer time for optimal learning.
The second part of the thesis discusses the data based learning. We have used data graph method which captures the system representation. The proposed data graph model captures and encodes the data efficiently and then the data graph is updated and trained with new data to provide efficient predictions. The model retains the previously learned knowledge by transfer learning and improves it with new training. The proposed method is highly adaptive and scalable for different real-time scenarios. Data graph models the system where every node (object) is associated with data and if two objects are related then they are linked with a data edge. The proposed algorithm is an incremental algorithm which learns hidden objects and hidden relationships through the data pattern over time and updates the model accordingly. We have used algebraic graph transformation methods to trigger the mutation of the Data Graph. This new updated Data Graph behaves differently for the data it observes. We explore more into machine learning algorithms and have proposed a complete framework to predict the state of the system based on the system parameters. We have proposed the discretization of the data points using the symbol algebra and used Bayesian machine learning algorithm to select the best model to represent the new data. Symbol algebra provides unified language platform to different sensor data and it can process both, the discrete and continuous data. The portability of unified language platform in processing heterogeneous and homogeneous data increases the hypotheses space and Bayesian machine learning gets more degrees of freedom in choosing the best model with high measure of confidence level in the predicted state.
34
Pipicelli, Claudio. "Quantum Machine Learning: A Comparison Between Quantum and Classical Support Vector Machine." Tesi di dottorato, 2020. http://www.fedoa.unina.it/13259/1/PhDThesisPipicelliFinal.pdf.
Повний текст джерелаАнотація:
This thesis is mainly focused on the study of Quantum Support Vector Machine
(QSVM), a very important member of the recent and innovative Quantum Machine Learning
field, and its comparison with conventional Support Vector Machine (SVM).
In this paper, I have worked on the application of Quantum Support Vector Machine
algorithm, that runs on near term quantum processors from I.B.M., through IBM Quantum
Experience cloud service, to a set of supervised machine learning case studies and
I compared its performance with classical Support Vector Machine algorithm; net of the
enormous hype surrounding the proliferation of quantum technologies in recent years, are
we beginning to glimpse an application of real interest in which quantum systems, albeit
with limitations, offer concrete improvements already now?
35
Tranter, Aaron. "Machine learning for quantum and complex systems." Phd thesis, 2021. http://hdl.handle.net/1885/220395.
Повний текст джерелаАнотація:
Machine learning now plays a pivotal role in our society, providing solutions to problems that were previously thought intractable. The meteoric rise of this technology can no doubt be attributed to the information age that we now live in. As data is continually amassed, more efficient and scalable methods are required to yield functional models and accurate inferences.
Simultaneously we have also seen quantum technology come to the forefront of research and next generation systems. These technologies promise secure information transfer, efficient computation and high precision sensing, at levels unattainable by their classical counterparts. Although these technologies are powerful, they are necessarily more complicated and difficult to control.
The combination of these two advances yields an opportunity for study, namely leveraging the power of machine learning to control and optimise quantum (and more generally complex) systems. The work presented in thesis explores these avenues of investigation and demonstrates the potential success of machine learning methods in the domain of quantum and complex systems.
One of the most crucial potential quantum technologies is the quantum memory. If we are to one day harness the true power of quantum key distribution for secure transimission of information, and more general quantum computating tasks, it will almost certainly involve the use of quantum memorys.
We start by presenting the operation of the cold atom workhorse: the magneto-optical trap (MOT). To use a cold atomic ensemble as a quantum memory we are required to prepare the atoms using a specialised cooling sequence. During this we observe a stable, coherent optical emission exiting each end of the elongated ensemble. We characterise this behaviour and compare it to similar observations in previous work.
Following this, we use the ensemble to implement a backward Raman memory. Using this scheme we are able to demonstrate an increased efficiency over that of previous forward recall implementations. While we are limited by the optical depth of the system, we observe an efficiency more than double that of previous implementations.
The MOT provides an easily accessible test bed for the optimisation via some machine learning technique. As we require an efficient search method, we implement a new type of algorithm based on deep learning. We design this technique such that the artificial neural networks are placed in control of the online optimisation, rather than simply being used as surrogate models.
We experimentally optimise the optical depth of the MOT using this method, by parametrising the time varying compression sequence. We identify a new and unintuitive method for cooling the atomic ensemble which surpasses current methods.
Following this initial implementation we make substantial improvements to the deep learning approach. This extends the approach to be applicable to a far wider range of complex problems, which may contain extensive local minima and structure. We benchmark this algorithm against many of the conventional optimisation techniques and demonstrate superior capability to optimise problems with high dimensionality. Finally we apply this technique to a series of preliminary problems, namely the tuning of a single electron transistor and second-order correlations from a quantum dot source.
36
Buffoni, Lorenzo. "Machine learning applications in science." Doctoral thesis, 2021. http://hdl.handle.net/2158/1227616.
Повний текст джерелаАнотація:
Machine learning is a broad field of study, with multifaceted applications of cross-disciplinary breadth that ultimately aims at developing computer algorithms that improve automatically through experience. The core idea of artificial intelligence technology is that systems can learn from data, so as to identify distinctive patterns and make consequently decisions, with minimal human intervention. The range of applications of these methodologies is already extremely vast, and still growing at a steady pace due to the pressing need to cope with the efficiently handling of big data. In parallel scientists have increasingly become interested in the potential of Machine Learning for fundamental research, for example in physics, biology and engineering. To some extent, this is not too surprising, since both Machine Learning algorithms and scientists share some of their methods as well as goals. The two fields are both concerned about the process of gathering and analyzing data to design models that can predict the behavior of complex systems. However, the fields prominently differ in the way their fundamental goals are realized. On the one hand, scientists use knowledge, intelligence and intuition to inform their models, on the other hand, Machine Learning models are agnostic and the machine provides the intelligence by extracting it from data often giving little to no insight on the knowledge gathered. Machine learning tools in science are therefore welcomed enthusiastically by some, while being eyed with suspicions by others, albeit producing surprisingly good results in some cases. In this thesis we will argue, using practical cases and applications from biology, network theory and quantum physics, that the communication between these two fields can be not only beneficial but also necessary for the progress of both fields.
37
Huang, Yufeng. "Computational Heterogeneous Electrochemistry – From Quantum Mechanics to Machine Learning." Thesis, 2019. https://thesis.library.caltech.edu/11328/13/Yufeng_thesis_2019.02.11.pdf.
Повний текст джерелаАнотація:
Because of coulomb interactions and complex surface morphologies, rigorous methods for heterogeneous electrochemical catalysis were not well-established. Thus, for different types of electrochemical systems, a specific strategy must be adapted. In this thesis, we first used the cluster model to study the chemistry on a 1D chain of MoS2 edges. Then, a rigorous grand canonical potential kinetics (GCP-K) method was developed for general crystalline systems. Starting from quantum mechanical calculations, the method gave rise to a different picture from the traditional description given by the Butler-Volmer kinetics. Next, we studied the chemical selectivity of CO2 reduction on polycrystalline copper nanoparticles. Because of the complexity of the reaction sites, we combined the reactive force field, density functional theory, and machine learning method to predict the reactive sites on 20,000 sites on a roughly 200,000-atom nanoparticle. Such a strategy opens up new way to understand chemistries on a much wider range of complex structures that were impossible to study theoretically. Lastly, we formulated a machine learning force field strategy using atomic energies for amorphous systems. We have shown that such a method can be used to reproduce quantum mechanical accuracies for molecular dynamics. This method will enable the accurate study of the dynamics of heterogeneous systems during electrochemical reactions. In summary, we have developed quantum chemical methods and machine learning strategies to reformulate rigorous ways to study a wide range of heterogeneous electrochemical catalysts.
38
"Machine learning for optical communications, nonlinear optics, and quantum optics." Tulane University, 2020.
Знайти повний текст джерела
39
"Classical and quantum data sketching with applications in communication complexity and machine learning." 2014. http://repository.lib.cuhk.edu.hk/en/item/cuhk-1291567.
Повний текст джерелаАнотація:
Liu, Yang.Thesis Ph.D. Chinese University of Hong Kong 2014.
Includes bibliographical references (leaves 163-188).
Abstracts also in Chinese.
Title from PDF title page (viewed on 25, October, 2016).
40
"Control of classical & quantum multispatial modes of light for quantum networks through nonlinear optics and machine learning." Tulane University, 2020.
Знайти повний текст джерелаАнотація:
archives@tulane.eduWith the advent of lasers the spatial shape of paraxial light also became an avenue for information processing and transfer applications. The light sources that support multiple of these spatial modes as separate, multiplexed information channels are readily used through classical optical implementations such as free-space optical communication, and to enhance the capacity of these channels. Recently, the hot atomic vapour based non-linear optical systems showed promise for the usage of paraxial multiple spatial modes of light for quantum information applications such as quantum communication, quantum networking, quantum computation and various other quantum technologies. In this dissertation, we use analytical, numerical, statistical and experimental techniques to model the propagation of multi-spatial light through various classical and non-linear systems to be able to steer, optimize and control the quantum states generated for quantum technologies applications. In the first chapter, we give a general introduction to classical (both linear & non-linear) and quantum (both linear \& non-linear) optical systems we are going to analyze. In the second chapter, we use a numerical, Fourier transform based beam propagation technique to examine the self-healing of a generic beam that is generated through an atomic process. In the third chapter, we analyzed our hot atomic vapour four-wave mixing experiment that uses a special type of multiple paraxial spatial mode to drive the non-linear optical process through numerical modeling of Fraunhofer diffraction. In the fourth chapter, we devise a coherent, analytically and quantum mechanically motivated beam propagation method based on decomposing the paraxial beam into its constituent multiple spatial modes. We calibrate this method by using the numerical, experimental and theoretical results of the previous chapters to model how the multiple spatial modes propagate through spatial masks that represent apertures, obstructions, atmospheric turbulence. In the fifth chapter, we extend the beam decomposition formalism into semi-classical and full quantum mechanical optical systems to model seeded hot-atomic vapour four-wave mixing experiments. We again calibrate our numerical models of intensity difference squeezing using the previous experimental results. Next, we use these calibrated models to devise a scheme to optimally generate multi-mode squeezed states. Lastly, in the sixth chapter, we turn our attention into estimating the quantum state of discrete variable, polarization qubit systems using machine learning and various other stochastic techniques. We improve these well studied systems to detect the quantum states in real time, in the presence of noise, and in the absence of various measurements using machine learning. We study these discrete variable, polarization qubit systems both as a gateway and a complement to study the tomographic reconstruction of continuous variable quantum optical systems of the previous chapters, in order to achieve our general goal of having a general estimation, steering and control methodology for quantum networking applications.
1
ONUR DANACI
41
Mengoni, Riccardo. "Quantum Approaches to Data Science and Data Analytics." Doctoral thesis, 2020. http://hdl.handle.net/11562/1018231.
Повний текст джерелаАнотація:
In this thesis are explored different research directions related to both the use of classical data analysis techniques for the study of quantum systems and the employment of quantum computing to speed up hard Machine Learning tasks
42
Fu, chien wei, and 傅建維. "Using quantum chemistry, molecular simulation and machine learning techniques to study the enzymatic mechanism for several enzymes." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/j556k8.
Повний текст джерелаАнотація:
博士國立清華大學
分子醫學研究所
103
Chapter I: Site of metabolism prediction for FMO enzymes via machine learning and condensed Fukui function The flavin-containing monooxygenase (FMO) catalyzes xenobiotics with soft nucleophiles and also plays an important role in drug metabolism in Phase I enzymes. The site of metabolism (SOM) refers to the place where the reaction of metabolism occurs in a molecule. Identification of SOMs of a compound is not usually a low-cost task in drug discovery. Thus, a silico method to predict site of metabolism (SOMs) of FMOs would provide medical chemists information of SOMs before experiments. In this work, we developed a machine learning model combining quantum features (condensed Fukui function) and circular fingerprints to predict potential SOMs in a molecule. The final model via SVM was easily interpreted with only five features. In the training set with 10 CV showed an area under curve (AUC) value of ROC curve, 0.889, and the value of MCC,0.767. In the external validation, AUC value of the model was 0.801 and the accuracy (MCC) was 0.611. These showed the predictive power of our model and we wish such a research to assist medical chemists in the assessment of FMO metabolism at the preclinical stage of drug discovery. Chapter II: Interaction between Trehalose and MTHase from Sulfolobus solfataricus studied by theoretical computation and site-directed mutagenesis Maltooligosyltrehalose trehalohydrolase (MTHase) catalyzes the release of trehalose by cleaving the α-1,4-glucosidic linkage next to the α-1,1-linked terminal disaccharide of maltooligosyltrehalose. Computer simulation using the hydrogen bond analysis, free energy decomposition, and computational alanine scanning were employed to investigate the interaction between maltooligosyltrehalose and the enzyme. The same residues that were chosen for theoretical investigation were also studied by site-directed mutagenesis and enzyme kinetic analysis. The importance of residues determined either experimentally or computed theoretically were in good accord with each other. It was found that residues Y155, D156, and W218 of subsites -2 and -3 of the enzyme might play an important role in interacting with the ligand. The theoretically constructed structure of the enzyme-ligand complex was further validated through an ab initio quantum chemical calculation using the Gaussian09 package. The activation energy computed from this latter study was very similar to those reported in literatures for the same type of hydrolysis reactions. Chapter III: A Theoretical Study on the Alkaline Hydrolysis of Methyl Thioacetate in Aqueous Solution A base catalyzed hydrolysis reaction of thiolester has been studied in both gas and solution phases using two ab initio quantum mechanics calculations such as Gaussian09 and CPMD. The free energy surface along the reaction path is also constructed using a configuration sampling technique namely the metadynamics method. While there are two different reaction paths obtained for the potential profile of the base-catalyzed hydrolysis reaction for thiolester in gas phase, a triple-well reaction path is computed for the reaction in solution phase by both two quantum mechanics calculations. Unlike a SN2 mechanism (a concerted mechanism) found for the gas-phase reaction, a nucleophilic attack from the hydroxide ion on the carbonyl carbon to yield a tetrahedral intermediate (a stepwise mechanism) is observed for the solution phase reaction. Moreover, the energy profiles computed by these two theoretical calculations are found to be well comparable with those determined experimentally.
43
Slatyer, Harry James. "Multi-parameter optimisation of quantum optical systems." Phd thesis, 2018. http://hdl.handle.net/1885/146120.
Повний текст джерелаАнотація:
Quantum optical systems are poised to become integral components
of technologies of the future. While there is growing commercial
interest in these systems---for applications in information
processing, secure communication and precision metrology---there
remain significant technical challenges to overcome before
widespread adoption is possible. In this thesis we consider the
general problem of optimising quantum optical systems, with a
focus on sensing and information processing applications. We
investigate four different classes of system with varying degrees
of generality and complexity, and demonstrate four corresponding
optimisation techniques.
At the most specific end of the spectrum---where behaviour is
best understood---we consider the problem of interferometric
sensitivity enhancement, specifically in the context of
long-baseline gravitational wave detectors. We investigate the
use of an auxiliary optomechanical system to generate squeezed
light exhibiting frequency-dependent quadrature rotation. Such
rotation is necessary to evade the effect of quantum back action
and achieve broadband sensitivity beyond the standard quantum
limit. We find that a cavity optomechanical system is generally
unsuitable for this purpose, since the quadrature rotation occurs
in the opposite direction to that required for broadband
sensitivity improvement.
Next we introduce a general technique to engineer arbitrary
optical spring potentials in cavity optomechanical systems. This
technique has the potential to optimise many types of sensors
relying on the optical spring effect. As an example, we show that
this technique could yield an enhancement in sensitivity by a
factor of 5 when applied to a certain gravitational sensor based
on a levitated cavity mirror.
We then consider a particular nanowire-based optomechanical
system with potential applications in force sensing. We
demonstrate a variety of ways to improve its sensitivity to
transient forces. We first apply a non-stationary feedback
cooling protocol to the system, and achieve an improvement in
peak signal-to-noise ratio by a factor of 3, corresponding to a
force resolution of 0.2fN. We then implement two non-stationary
estimation schemes, which involve post-processing data taken in
the absence of physical feedback cooling, to achieve a comparable
enhancement in performance without the need for additional
experimental complexity.
Finally, to address the most complex of systems, we present a
general-purpose machine learning algorithm capable of
automatically modelling and optimising arbitrary physical systems
without human input. To demonstrate the potential of the
algorithm we apply it to a magneto-optical trap used for a
quantum memory, and achieve an improvement in optical depth from
138 to 448.
The four techniques presented differ significantly in their style
and the types of systems to which they are applicable.
Successfully harnessing the full range of such optimisation
procedures will be vital in unlocking the potential of quantum
optical systems in the technologies of the future
44
Swann, Ellen Therese. "Development and application of statistical and quantum mechanical methods for modelling molecular ensembles." Phd thesis, 2018. http://hdl.handle.net/1885/142784.
Повний текст джерелаАнотація:
The development of new quantum chemical methods requires
extensive benchmarking to establish the accuracy and limitations
of a method. Current benchmarking practices in computational
chemistry use test sets that are subject to human biases and as
such can be fundamentally flawed. This work presents a thorough
benchmark of diffusion Monte Carlo
(DMC) for a range of systems and properties as well as a novel
method for developing new, unbiased test sets using multivariate
statistical techniques. Firstly, the hydrogen abstraction of
methanol is used as a test system to develop a more efficient
protocol that minimises the computational cost of DMC without
compromising accuracy. This protocol is then applied to three
test sets of reaction energies, including 43 radical
stabilisation energies, 14 Diels-Alder reactions and 76 barrier
heights of hydrogen and non-hydrogen transfer reactions. The
average mean absolute error for all three databases is just 0.9
kcal/mol.
The accuracy of the explicitly correlated trial wavefunction used
in DMC is demonstrated using the ionisation potentials and
electron affinities of first- and second-row atoms. A
multi-determinant trial wavefunction reduces the errors for
systems with strong multi-configuration character, as well as
for predominantly single-reference systems. It is shown that the
use of pseudopotentials in place of all-electron basis sets
slightly increases the error for these systems. DMC is then
tested with a set of eighteen challenging reactions.
Incorporating more determinants in the trial wavefunction reduced
the errors for most systems but results are highly dependent on
the active space used in the CISD wavefunction. The accuracy of
multi-determinant DMC for strongly multi-reference systems is
tested for the isomerisation of diazene. In this case no method
was capable of reducing the error of the strongly-correlated
rotational transition state.
Finally, an improved method for selecting test sets is presented
using multivariate statistical techniques. Bias-free test sets
are constructed by selecting archetypes and prototypes based on
numerical representations of molecules. Descriptors based on the
one-, two- and three-dimensional structures of a molecule are
tested. These new test sets are
then used to benchmark a number of methods.
45
Hughes, Zak E., E. Ren, J. C. R. Thacker, B. C. B. Symons, A. F. Silva, and P. L. A. Popelier. "A FFLUX water model: flexible, polarizable and with a multipolar description of electrostatics." 2019. http://hdl.handle.net/10454/17932.
Повний текст джерелаАнотація:
YesKey to progress in molecular simulation is the development of advanced models that go beyond the limitations of traditional force fields that employ a fixed, point charge‐based description of electrostatics. Taking water as an example system, the FFLUX framework is shown capable of producing models that are flexible, polarizable and have a multipolar description of the electrostatics. The kriging machine‐learning methods used in FFLUX are able to reproduce the intramolecular potential energy surface and multipole moments of a single water molecule with chemical accuracy using as few as 50 training configurations. Molecular dynamics simulations of water clusters (25–216 molecules) using the new FFLUX model reveal that incorporating charge‐quadrupole, dipole–dipole, and quadrupole–charge interactions into the description of the electrostatics results in significant changes to the intermolecular structuring of the water molecules.
EPSRC. Grant Number: K005472
46
MARCUCCI, GIULIA. "Complex extreme nonlinear waves: classical and quantum theory for new computing models." Doctoral thesis, 2020. http://hdl.handle.net/11573/1353250.
Повний текст джерелаАнотація:
The historical role of nonlinear waves in developing the science of complexity, and also their physical feature of being a widespread paradigm in optics, establishes a bridge between two diverse and fundamental fields that can open an immeasurable number of new routes. In what follows, we present our most important results on nonlinear waves in classical and quantum nonlinear optics. About classical phenomenology, we lay the groundwork for establishing one uniform theory of dispersive shock waves, and for controlling complex nonlinear regimes through simple integer topological invariants. The second quantized field theory of optical propagation in nonlinear dispersive media allows us to perform numerical simulations of quantum solitons and the quantum nonlinear box problem. The complexity of light propagation in nonlinear media is here examined from all the main points of view: extreme phenomena, recurrence, control, modulation instability, and so forth. Such an analysis has a major, significant goal: answering the question can nonlinear waves do computation? For this purpose, our study towards the realization of an all-optical computer, able to do computation by implementing machine learning algorithms, is illustrated. The first all-optical realization of the Ising machine and the theoretical foundations of the random optical machine are here reported. We believe that this treatise is a fundamental study for the application of nonlinear waves to new computational techniques, disclosing new procedures to the control of extreme waves, and to the design of new quantum sources and non-classical state generators for future quantum technologies, also giving incredible insights about all-optical reservoir computing. Can nonlinear waves do computation? Our random optical machine draws the route for a positive answer to this question, substituting the randomness either with the uncertainty of quantum noise effects on light propagation or with the arbitrariness of classical, extremely nonlinear regimes, as similarly done by random projection methods and extreme learning machines.
47
Thacker, J. C. R., A. L. Wilson, Zak E. Hughes, M. J. Burn, P. I. Maxwell, and P. L. A. Popelier. "Towards the simulation of biomolecules: optimisation of peptide-capped glycine using FFLUX." 2018. http://hdl.handle.net/10454/15726.
Повний текст джерелаАнотація:
YesThe optimisation of a peptide-capped glycine using the novel force field FFLUX is presented. FFLUX is a force field based on the machine-learning method kriging and the topological energy partitioning method called Interacting Quantum Atoms. FFLUX has a completely different architecture to that of traditional force fields, avoiding (harmonic) potentials for bonded, valence and torsion angles. In this study, FFLUX performs an optimisation on a glycine molecule and successfully recovers the target density-functional theory energy with an error of 0.89 ± 0.03 kJ mol−1. It also recovers the structure of the global minimum with a root-mean-squared deviation of 0.05 Å (excluding hydrogen atoms). We also show that the geometry of the intra-molecular hydrogen bond in glycine is recovered accurately.
EPSRC Established Career Fellowship [grant number EP/K005472]
48
(9674882), Sayan Basak. "Hysteresis and Pattern Formation in Electronic Phase Transitions in Quantum Materials." Thesis, 2020.
Знайти повний текст джерелаАнотація:
We propose an order parameter theory of the quantum Hall nematic in high fractional Landau levels in terms of an Ising description. This new model solves a couple of extant problems in the literature: (1) The low-temperature behavior of the measured resistivity anisotropy is captured better by our model than previous theoretical treatments based on the electron nematic having XY symmetry. (2) Our model allows for the development of true long-range order at low temperature, consistent with the observation of anisotropic low-temperature transport.
We furthermore propose new experimental tests based on hysteresis that can distinguish whether any two-dimensional electron nematic is in the XY universality class (as previously proposed in high fractional Landau levels), or in the Ising universality class (as we propose). Given the growing interest in electron nematics in many materials, we expect our proposed test of universality class to be of broad interest.
Whereas the XY model in two dimensions does not have a long-range ordered phase, the addition of uniaxial random field disorder induces a long-range ordered phase in which the spontaneous magnetization points perpendicular to the random field direction, via an order-by-disorder transition. We have shown that this spontaneous magnetization is robust against a rotating driving field, up to a critical driving field amplitude. Thus we have found evidence for a new non-equilibrium phase transition that was unknown before in this model. Moreover, we have discovered an incredible anomaly at this nonequilibrium phase transition: the critical region is accompanied by a cascade of period multiplication events. This physics is reminiscent of the period bifurcation cascade signaling the transition to chaos in nonlinear systems, and of the approach to the irreversibility transition in models of yield in amorphous solids~\cite{reichhardt-dahmen,leishangthem_yielding_2017}. This period multiplication cascade is surprising to be present in a statistical mechanics model, and suggests that the non-equilibrium transition as a function of driving field amplitude is part of a larger class of transitions in dynamical systems.
Moreover, we show that this multi-period behavior represents a new emergent classical discrete time-crystal, since the new period is robust against changes to initial conditions and low-temperature fluctuations over hundreds of driving period cycles.
We expect this work to be of broad interest, further encouraging cross-fertilization between the rapidly growing field of time-crystals with the well-established fields of nonequilibrium phase transitions and dynamical systems.
Geometrical configurations gave us a better understanding of the multi-period behavior of the limit-cycles.
Moreover, surface probes are continually evolving and generating vast amounts of spatially resolved data of quantum materials, which reveal a lot of detail about the microscopic and macroscopic properties of the system.
Materials undergoing a transition between two distinct states, phase separate.
These phase-separated regions form intricate patterns on the observable surface, which can encode model-specific information, including interaction, dimensionality, and disorder.
While there are rigorous methods for understanding these patterns, they turn out to be time-consuming as well as requiring expertise.
We show that a well-tuned machine learning framework can decipher this information with minimal effort from the user.
We expect this to be widely used by the scientific community to fast-track comprehension of the underlying physics in these materials.
49
Lamarre, Aldo. "Apprentissage de circuits quantiques par descente de gradient classique." Thesis, 2020. http://hdl.handle.net/1866/24322.
Повний текст джерелаАнотація:
Nous présentons un nouvel algorithme d’apprentissage de circuits quantiques basé sur la descente de gradient classique. Comme ce sujet unifie deux disciplines, nous expliquons les deux domaines aux gens de l’autre discipline. Conséquemment, nous débutons par une
présentation du calcul quantique et des circuits quantiques pour les gens en apprentissage automatique suivi d’une présentation des algorithmes d’apprentissage automatique pour les
gens en informatique quantique. Puis, pour motiver et mettre en contexte nos résultats,
nous passons à une légère revue de littérature en apprentissage automatique quantique. Ensuite, nous présentons notre modèle, son algorithme, ses variantes et quelques résultats empiriques. Finalement, nous critiquons notre implémentation en montrant des extensions et
des nouvelles approches possibles. Les résultats principaux se situent dans ces deux dernières parties, qui sont respectivement les chapitres 4 et 5 de ce mémoire. Le code de l’algorithme
et des expériences que nous avons créé pour ce mémoire se trouve sur notre github à l’adresse suivante : https://github.com/AldoLamarre/quantumcircuitlearning.We present a new learning algorithm for quantum circuits based on gradient descent. Since this subject unifies two areas of research, we explain each field for people working in the other domain. Consequently, we begin by introducing quantum computing and quantum circuits to machine learning specialists, followed by an introduction of machine learning to quantum computing specialists. To give context and motivate our results we then give a light literature review on quantum machine learning. After this, we present our model, its algorithms and its variants, then discuss our currently achieved empirical results. Finally, we criticize our models by giving extensions and future work directions. These last two parts are our main results. They can be found in chapter 4 and 5 respectively. Our code which helped obtain these results can be found on github at this link : https://github.com/ AldoLamarre/quantumcircuitlearning.