Quantum Science and Engineering Center researchers awarded ERC Grants

Jean-Philippe Brantut and Giuseppe Carleo were each awarded one of the highly competitive and prestigious Consolidator Grants from the ERC, which grant researchers 2 million euros apiece. This year, of the 2,652 applicants, 313 were chosen, with a success rate of only roughly 12%. Of these 313 awardees, 6 are researchers at EPFL.

Both Brantut and Carleo are excited about what these grants mean to quantum research and appreciate the opportunity to follow their scientific curiosity wherever it leads them.

“That’s the thing with the ERC,” says Brantut. “This is perhaps the only place where you can propose anything you like, and the only judgement is about the scientific excellence and not about how relevant some people might think it is.”

Jean-Phillippe Brantut

Brantut began the work that provides the basis of his project, DDisQS: Driven and Dissipative Quantum Simulator, at EPFL in 2016. He is now able to use this ERC grant to develop experiments where his group will couple matter and light and control the interaction, applying this method to quantum simulation.

This project works to develop specialized quantum machines that can be used to model how atoms behave in certain materials, and leverages driving, what is done to a system from the outside, and dissipation, when a system leaks to the outside, to extend the scope of quantum simulation to problems that were not previously accessible.

“Usually these models are not easy to solve by traditional theoretical methods,” Brantut explains. “The idea is to try and force atoms to behave in accordance with a specific model, and then read out what they do.” The main advantage to this is Brantut and his team can manipulate and observe these atoms in a more effective way than in naturally occurring systems.

In addition, dissipation is typically a hindrance that everyone fights against in quantum technology, so being able to leverage dissipation instead of have it as an annoyance will be an important development in the field.

“If someone can better understand how dissipation operates at the fundamental level and use it as a tool, that would be highly beneficial for anything you want to do with quantum,” Brantut says, “be it sensing, computing, and of course quantum simulation.”

This project also has the broader goal of furthering our understanding of fundamental science, because there are a lot of open questions in quantum physics that we don’t know how to address besides through quantum simulations, and driving and dissipation are good tools to deal with these.

Gisueppe Carleo

Carleo’s project, NEQS:Neural Quantum Simulation, has the aim of simulating a group of problems that hasn’t been possible before. So far, existing machine learning methods have been able to study quantum systems at static state, which is when something described by quantum mechanics, such as a molecule or material, is at equilibrium and doesn’t change in time. What is more difficult is describing and predicting what happens when you disturb these quantum systems and force them out of equilibrium.

“This is where we want to develop new techniques based on deep learning and machine learning that would allow us to study more accurately and make predictions in this setting,” Carleo says.

To do this, Carleo and his team will study an isolated quantum system by simulating what would happen over time to a molecule described by quantum mechanics in a perfectly closed and isolated box, so that it was at equilibrium and then it was given one ‘kick’, or disturbance.

“We take something that is normally at equilibrium, we hit it hard, then this thing evolves, but without contact with the rest of the environment,” Carleo explains. “The only interaction is at the beginning and the rest is controlled.” This allows him to compare what happens on long time scales when the system is allowed to evolve and eventually relax again to an equilibrium state. The goal is to understand whether thermodynamics is a good descriptor of equilibrium properties with strong interactions.

The first part of this work will be to run the simulations on classical super computers. In the second part, Carleo and his team will combine the power of these advanced classical methods with the quantum computers that are currently available.

“These guys are small and noisy,” Carleo says. “They’re not good, they’re just the best we have right now, and we have to try and see what we can do with them.” To help these computers work better, they use what are called hybrid computation schemes, which means they run some things back and forth between the classical and quantum computers so they can work together to solve problems.

The ERC at EPFL

These grants will last for 5 years, and for Brantut, it comes at the end of an ERC Starter Grant he was awarded a few years ago, building this further research from the project he began with that initial grant. Both projects will mostly use the funds to hire people, as well as work on developing the machinery in Brantut’s case and paying for computing time in Carleo’s.

It's really exciting, as it’s so competitive," says Brantut. “I think the fact that I was at EPFL with my own group played a big role”Being at a good place with good students and good infrastructure is something that is highly beneficial”