This project will explore the use of analogue quantum simulation to provide information on models from quantum chemistry, especially when combined with machine learning techniques. Our starting point will be the formulation of relevant models based on harmonic oscillators and/or coupled spins and lattice models that can be studied with analogue quantum simulators, especially with neutral atoms. We will study these models using state-of-the art classical techniques, and understand how to implement them on near and medium-term quantum hardware, and to identify regimes where we might be able to identify a quantum advantage.

We will then combine this with classical machine learning techniques, understanding how quantum computing can assist in the generation of training or benchmarking data for machine learning techniques, relevant to the chemistry models identified at the beginning of the project. This could involve, e.g., training of machine learning models for larger systems on data generated for smaller systems by quantum computers, or optimisation of the AI based on matching outputs to solutions generated on a quantum computer to models that are beyond the capabilities of simulation algorithms running on known conventional computers. Ultimately, we aim to identify ways that analogue quantum simulators might usefully be used in this context, and on what timescale the relevant hardware will be available to impact this sector.

For the PhD candidate, the project will provide opportunities to develop expertise in both quantum computing and state-of-the art classical computation techniques. These will include tensor networks, machine learning methods, and classical simulators for medium-scale implementations of Quantum Computing.