Master’s student in Aerospace Engineering at ISAE-SUPAERO with a strong academic background in physics and a deep interest in quantum computing. I completed a bachelor’s degree in physics with a dual program focused on research, which gave me early experience in scientific projects. My current work explores quantum machine learning algorithms, particularly a quantum version of k-means clustering, using PennyLane.

Quantum computing has emerged as a promising technology by leveraging fundamental principles of quantum mechanics, such as quantum parallelism and entanglement, with the aim of accelerating computational tasks and improving accuracy. In astrophysics, the massive amount of data generated by modern surveys makes unsupervised methods like clustering particularly valuable, especially to complement traditional visual galaxy classifications and to reveal underlying physical correlations.

This project aims to implement and test a quantum kmeans clustering algorithm on real galaxy data. To achieve this, a first algorithm was investigate and validated on the Iris dataset. It demonstrates that while the expected speed-up is not yet observable due to the number of shots required for accurate results, this quantum circuit correctly computes distances and performs clustering. A quantum circuit with a higher level of superposition, is still under development and has not yet yielded stable results. In parallel, a detailed study has been conducted to select the most relevant galaxy dataset and features for future application.

The next steps will be to finalise and validate the second level quantum k-means algorithm. Apply it to the chosen galaxy data, in order to investigate whether the resulting clusters can reproduce or refine the classical morphological classification, and show improvement compared to classical method.

This project is based on the algorithm and result used in “Quantum clustering with k-Means: A hybrid approach” from Alessandro Poggiali et al. (doi: 10.1016/j.tcs.2024.114466).

Arnaud Dion

Over the past few months, I’ve developed expertise in quantum machine learning using the PennyLane library from Xanadu. I worked on implementing a quantum version of the k-means algorithm by computing distances through measurement probabilities. This involved hands-on manipulation of qubits, gates, and quantum circuits such as the Deutsch-Jozsa and Grover algorithms. I’d be glad to support anyone working on quantum algorithms, quantum clustering tasks, or needing help setting up quantum simulations in PennyLane.

Quantum k-means clustering could be applied to high dimensional classification tasks.