Quantum Machine Learning

Overview

Quantum Machine Learning (QML) is an emerging field that combines the power of quantum computing with the capabilities of machine learning to solve complex problems in various domains. QML leverages the principles of quantum mechanics, such as superposition and entanglement, to speed up certain computational tasks and develop novel algorithms.

In traditional machine learning, classical computers process data using linear operations, which can be slow and inefficient for large datasets. Quantum computers, on the other hand, can perform many calculations simultaneously due to their ability to exist in multiple states (superposition) and be connected in ways that defy classical physics (entanglement). This allows QML algorithms to:

  1. Speed up training: Quantum computers can efficiently process large datasets and perform complex calculations, reducing training times for machine learning models.
  2. Improve accuracy: Quantum computers can solve problems with higher precision than classical computers, leading to more accurate predictions and decision-making.
  3. Discover new patterns: Quantum computers can explore vast solution spaces and identify novel relationships between data features, enabling the discovery of new patterns and insights.

As QML continues to evolve, it has the potential to revolutionise various industries and transform the way we approach machine learning and artificial intelligence.

Our Interests

Aside from the development of novel quantum artificial intelligence frameworks, comparing quantum machine learning algorithms against their classical counterparts represents a significant challenge in and of itself. How can we be sure that a quantum algorithm performs better than classical because of an intrinsic quantum enhancement? How do we know that our classical network isn’t a straw man? How can we compare apples to oranges?

In our group, we adopt the trial by fire philosophy, and propose that algorithms can be continuously benchmarked against one another in a standardised environment: the classical computer game Pong. Think of it like the machine learning Olympics! Take a sneak peak of our environment in action below as a reinforcement learning algorithm (The Terminator) battles against a dyed-in-the-wool “follow the ball” algorithm (Occam’s Chainsaw).