Non-equilibrium phenomena in quantum physics are ubiquitous in current research because of their extremely rich phenomenology. Among them, adiabatic processes are particularly interesting because they connect two of the cornerstone of modern condensed matter theory: topological phases of matter and quantum information. Indeed with adiabatic periodic driving, it is possible to engineer exotic non-equilibrium phases with tunable topological properties, which are of great interest for both fundamental research and future quantum technologies. Adiabatic processes can also be used to implement optimization processes on quantum hardware and lie at the basis of hybrid quantum-classical algorithms that are extremely promising for near-term quantum devices. In this thesis, I focus on these two aspects of adiabatic dynamics in quantum systems. In the first part, I investigate the robustness of topological phases, arising in periodically driven systems, with respect to the driving protocol and disorder. In particular, I show that quantized transport and Anderson localization coexist in one-dimensional systems displaying Thouless pumping, because of a delocalization-localization transition in the Floquet spectrum. This transition is linked to the topological nature of the adiabatic driving and disappears if the phase is trivial in the clean limit. In the second part, I study one of the most popular hybrid optimization method, the quantum approximate optimization algorithm (QAOA). I show that QAOA can tackle the first-order phase transition arising in the infinite range p-spin model with polynomial resources, in stark contrast with adiabatic quantum computation, which requires exponentially long evolution time to reach similar performances. Finally, I present an approach to QAOA based on reinforcement learning (RL). Interestingly, the RL agent automatically adopts strategies that converge towards optimal adiabatic schedules and that can be easily transferred between systems with different sizes, even in the presence of disorder.

Adiabatic approaches to non-equilibrium systems: Topology, Optimization, and Learning / Wauters, Matteo Michele. - (2020 Sep 29).

Adiabatic approaches to non-equilibrium systems: Topology, Optimization, and Learning

Wauters, Matteo Michele
2020-09-29

Abstract

Non-equilibrium phenomena in quantum physics are ubiquitous in current research because of their extremely rich phenomenology. Among them, adiabatic processes are particularly interesting because they connect two of the cornerstone of modern condensed matter theory: topological phases of matter and quantum information. Indeed with adiabatic periodic driving, it is possible to engineer exotic non-equilibrium phases with tunable topological properties, which are of great interest for both fundamental research and future quantum technologies. Adiabatic processes can also be used to implement optimization processes on quantum hardware and lie at the basis of hybrid quantum-classical algorithms that are extremely promising for near-term quantum devices. In this thesis, I focus on these two aspects of adiabatic dynamics in quantum systems. In the first part, I investigate the robustness of topological phases, arising in periodically driven systems, with respect to the driving protocol and disorder. In particular, I show that quantized transport and Anderson localization coexist in one-dimensional systems displaying Thouless pumping, because of a delocalization-localization transition in the Floquet spectrum. This transition is linked to the topological nature of the adiabatic driving and disappears if the phase is trivial in the clean limit. In the second part, I study one of the most popular hybrid optimization method, the quantum approximate optimization algorithm (QAOA). I show that QAOA can tackle the first-order phase transition arising in the infinite range p-spin model with polynomial resources, in stark contrast with adiabatic quantum computation, which requires exponentially long evolution time to reach similar performances. Finally, I present an approach to QAOA based on reinforcement learning (RL). Interestingly, the RL agent automatically adopts strategies that converge towards optimal adiabatic schedules and that can be easily transferred between systems with different sizes, even in the presence of disorder.
29-set-2020
Santoro, Giuseppe Ernesto
Wauters, Matteo Michele
File in questo prodotto:
File Dimensione Formato  
Thesis.pdf

accesso aperto

Descrizione: PhD Thesis
Tipologia: Tesi
Licenza: Non specificato
Dimensione 8.46 MB
Formato Adobe PDF
8.46 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.11767/114409
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact