Joseph Bowles

Quantum Machine Learning Researcher (Remote) - Xanadu - Career Page Apply to Quantum Machine Learning Researcher (Remote) at Xanadu in Toronto, ON, Canada.

New opening in the QML team @ Xanadu!

xanadu.applytojob.com/apply/v1gc8F...

GitHub - XanaduAI/iqpopt: Fast optimization of instantaneous quantum polynomial circuits in JAX Fast optimization of instantaneous quantum polynomial circuits in JAX - XanaduAI/iqpopt

If you want to experiment yourself, you can construct and train quantum generative models in the IQPopt package we recently released:

github.com/XanaduAI/iqp...

We did not encounter problems of barren plateaus, which may be due to the fact we use a data-dependent initialisation strategy. We still aren't sure if the ansatz/loss combo has barren plateaus under random initialisation.

Train on classical, deploy on quantum: scaling generative quantum machine learning to a thousand qubits We propose an approach to generative quantum machine learning that overcomes the fundamental scaling issues of variational quantum circuits. The core idea is to use a class of generative models based ...

Very excited to release our latest paper on generative QML. Much to our own surprise, we show that it is feasible to train quantum models with up to a thousand qubits and hundreds of thousands of parameters, all on classical hardware.

scirate.com/arxiv/2503.0...

IQPopt: Fast optimization of instantaneous quantum polynomial circuits in JAX IQPopt is a software package designed to optimize large-scale instantaneous quantum polynomial circuits on classical hardware. By exploiting an efficient classical simulation algorithm for expectation...

We recently released a software package to optimize IQP circuits that scales to millions of qubits and gates. Our main motivation was quantum machine learning (paper on the way!), but we believe there could be many other applications.

arxiv.org/abs/2501.04776
github.com/XanaduAI/iqp...

yes please :)