Back Quantum Machine Learning for Healthcare

Quantum Machine Learning for Healthcare

Quantum Machine Learning for Healthcare
This project, structured as a multi-disciplinary PhD thesis, will be focused on the design, development and application of novel methods for data analysis within the paradigm of quantum computation, in what is known as quantum machine learning (QML)

This project, structured as a multi-disciplinary PhD thesis, will be focused on the design, development and application of novel methods for data analysis within the paradigm of quantum computation, in what is known as quantum machine learning (QML). Quantum computation is known to outperform classical computers in tasks such as unstructured search (Grover’s algorithm) and factorization (Shor’s algorithm), fundamental to applications such as cryptography. QML explores the potential benefits of quantum representations and mechanisms such as superposition and entanglement for massively parallel processing. Current research is limited by the state of quantum technology, which is in its infancy but developing at an impressive pace.

We have recently initiated a new research line within our group, exploring the potential of QML for healthcare. Initially, we have focused on quantum computing embeddings for medical imaging data and quantum machine learning algorithms for classification of lung cancer Current on-going research also includes preliminary work on quantum generative models, such as quantum GANs, quantum VAEs4 and quantum circuit Born machines. We will explore the use of these techniques for health data, as well as devising novel approaches based on quantum adaptations of algorithms such as optimal transport normalizing flows, which are being developed by our group to analyze medical imaging data. In this line of work, we will collaborate with Gustavo Deco (DTIC-CBC).

In a related parallel line of work, we will also explore quantum-inspired algorithms, such as those based on tensor networks, for which promising preliminary results have been reported for big data analysis tasks, such as those of the second use case, on pollution and policy evaluation (in collaboration with Josep M. Antó, from ISGlobal & DMELIS).

We will make extensive use of open access software libraries (e.g. Qiskit, Pennylane, Cirq), as well as open quantum simulators and hardware systems, such as those provided by IBM and Xanadu. Furthermore, we will work with existing collaborators in this line of research, such as the Barcelona Supercomputing Center, which lead the Quantum Spain initiative and will soon host two quantum computers.

Principal researchers

Miguel Ángel González Ballester

UPF group:

 

  • Prof. Josep M. Antó (ISGlobal, DMELIS & UPF Centre for Planetary Wellbeing), scientific collaborator, owner of pollution and policy evaluation use case.

  • Prof. Gustavo Deco (DTIC-CBC & ICREA), scientific collaborator, owner of neuroimaging and neuronal connectivity assessment use case.

  • Dr. Xavier Font (Tecnocampus), scientific collaborator / MSc student on quantum machine learning, supervised by the PI.


External:

  • Dr. Mario Ceresa (EU Joint Research Centre), leader of previous Planetary Wellbeing project IPER, on agent-based simulation for pollution and policy evaluation.

  • Dr. Artur Garcia (Barcelona Supercomputing Center, leader of Quantic group), collaborator on quantum computing theory and applied research.

The project will be supported by the PhD Fellowship program at the Department of Information and Communication Technologies at UPF.

Access to quantum computing resources (IBM, Pennylane, BSC), as well as non-quantum high-performance computing (DTIC cluster, BSC Mare Nostrum), will be also available for its execution.