Quantum Computing utilises quantum phenomena including superposition, entanglement, and quantum tunneling to potentially perform certain computations exponentially faster and with far greater complexity than classical systems. Thus, possibly solving problems including molecular simulations, drug discovery, complex optimisation tasks, and machine learning.

Our work at the QUantum Algorithmic Intelligence Laboratory (QUAIL) encompasses the following three main research areas:

Quantum Optimisation

We aim to investigate novel quantum heuristics for both classical discrete and continuous optimisation. A major obstacle to the application of quantum optimisation algorithms is the mismatch between the size of the encoded optimisation problem and the computational resources available to quantum processors. Our research has led to the development of algorithms that decomposes large QUBOs into smaller subproblems that can be solved iteratively on quantum or classical backends.

Machine Learning for Quantum Computing

Machine Learning techniques can support quantum-device tuning, calibration, control error-related workflows, compilation, and other hardware-aware activities. This approach is favourable when there is data to support a recurring inference or decision problem like estimating drift, selection a control action, predicting execution quality, or choosing an implementation for a changing backend. Our research currently explores techniques like gradient descent, evolutionary computation, reinforcement learning, and multi-task approaches for quantum control applicable to realistic noisy quantum devices.

Quantum Computing for Machine Learning

Quantum Computing for machine learning includes parameterised quantum circuits as trainable models, feature maps, or kernels within a classical optimisation loop. It also includes hybrid transfer-learning approaches that use classical models to reduce a high-dimensional representation before a quantum component. Our research aims to answer whether a quantum component provides a reproducible benefit, a useful inductive bias, or scientific insight under a controlled resource budget, while ultimately identifying which learning problems align with operations that quantum computers perform natively and remain difficult to approximate classically.

We are currently looking for PhD students, and interns to join us! Don’t stay in superposition, entangle with us! Contact us for more details.