mariaschuld(She/Her)

I have spent the past ten years thinking about what quantum computers mean for machine learning: Does quantum information fundamentally change the way in which computers learn from data? It took me a long time to understand how the way we phrase our questions influences the answers we are able to find, and me and my small team of researchers at Xanadu are looking for the right questions to make quantum machine learning a reality one day. Classical machine learning also fascinates me, in particular the theoretical riddles it poses, the possibilities it creates to study social behaviour, and the commercial potential it has for South Africa - my home of choice for the last decade.

Demos

An equivariant graph embedding

Computing gradients in parallel with Amazon Braket

Function fitting with a photonic quantum neural network

How to optimize a QML model using JAX and JAXopt

How to optimize a QML model using JAX and Optax

Kernel-based training of quantum models with scikit-learn

Quantum models as Fourier series

Variational classifier

Blog Posts

Why measuring performance is our biggest blind spot in quantum machine learning

Benchmarks, exponential speedups, expressivity…the question of how we measure if a quantum machine learning method is "good" can be a lot more tricky than one may think.

Taking stock of quantum machine learning—a critical perspective

PennyLane's very own Maria Schuld recently gave a talk at the inaugural QML Meetup, summarizing some critical quantum machine learning insights. Check out the video below to learn why variational quantum models have a scaling problem, why the way we encode data must become more than an incidental choice …