Variational quantum algorithms show significant potential in effectively operating on noisy intermediatescale quantum (NISQ) devices. Using classical optimizers and parameterized quantum circuits, these hybrid quantumclassical algorithms can adapt to and mitigate noiseinduced errors and imperfections in quantum hardware. A novel variational approach to reinforcement learning has been recently proposed, incorporating linearoptical interferometers and a classical learning model known as projective simulation (PS). PS is a decisionmaking tool for reinforcement learning and can be classically represented as a random walk on a graph that describes the agent’s memory. In its optical quantum version, this approach utilizes quantum walks of single photons on a mesh of tunable MachZehnder interferometers to select actions. In this thesis, we physically implement this algorithm on a photonic chip provided by Quandela, using Perceval, a framework to simulate and interface with discretevariable photonic quantum computers. Finally, we evaluate the algorithm on a use case of industrial interest, taking the first steps towards the application of interpretable quantum learning agents.
Variational quantum algorithms show significant potential in effectively operating on noisy intermediatescale quantum (NISQ) devices. Using classical optimizers and parameterized quantum circuits, these hybrid quantumclassical algorithms can adapt to and mitigate noiseinduced errors and imperfections in quantum hardware. A novel variational approach to reinforcement learning has been recently proposed, incorporating linearoptical interferometers and a classical learning model known as projective simulation (PS). PS is a decisionmaking tool for reinforcement learning and can be classically represented as a random walk on a graph that describes the agent’s memory. In its optical quantum version, this approach utilizes quantum walks of single photons on a mesh of tunable MachZehnder interferometers to select actions. In this thesis, we physically implement this algorithm on a photonic chip provided by Quandela, using Perceval, a framework to simulate and interface with discretevariable photonic quantum computers. Finally, we evaluate the algorithm on a use case of industrial interest, taking the first steps towards the application of interpretable quantum learning agents.
Quantum Optical Projective Simulation for Reinforcement Learning
FRANCESCHETTO, GIACOMO
2022/2023
Abstract
Variational quantum algorithms show significant potential in effectively operating on noisy intermediatescale quantum (NISQ) devices. Using classical optimizers and parameterized quantum circuits, these hybrid quantumclassical algorithms can adapt to and mitigate noiseinduced errors and imperfections in quantum hardware. A novel variational approach to reinforcement learning has been recently proposed, incorporating linearoptical interferometers and a classical learning model known as projective simulation (PS). PS is a decisionmaking tool for reinforcement learning and can be classically represented as a random walk on a graph that describes the agent’s memory. In its optical quantum version, this approach utilizes quantum walks of single photons on a mesh of tunable MachZehnder interferometers to select actions. In this thesis, we physically implement this algorithm on a photonic chip provided by Quandela, using Perceval, a framework to simulate and interface with discretevariable photonic quantum computers. Finally, we evaluate the algorithm on a use case of industrial interest, taking the first steps towards the application of interpretable quantum learning agents.File  Dimensione  Formato  

Franceschetto_Giacomo.pdf
accesso riservato
Dimensione
9.82 MB
Formato
Adobe PDF

9.82 MB  Adobe PDF 
The text of this website © Università degli studi di Padova. Full Text are published under a nonexclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/54701