Exploring the Possibilities of Memorizers for Quantum Computing and Artificial Intelligence – News

Ian Billiard and Dominic Druin, teachers Per Department of Electrical Engineering and Computer Engineering Faculty of Engineering, and member of the Quantum Institute of UdeS.

Photo: Michelle Caron – UdeS

Although the term biomimicry only entered Le Robert in 1975, the process has been used for a long time. “Learn from nature, you will find the future there,” said Leonardo da Vinci.

For many researchers at the Institut Quantique (IQ), the future is quantum computers. In addition to its architecture, algorithms and error correction, this extraordinary potential research tool will also require the participation of engineers who will ensure its control. This is a huge challenge, and there is still much work to be done, since maintaining quantum conditions requires cryogenic temperatures. So there is a need to create electronics that can operate at these catastrophic temperatures.

To build a quantum computer, we need to create electronics capable of working at cryogenic temperatures.

And, to do that, a research team from the Institut Quantique (IQ) and the Interdisciplinary Institute for Technological Innovation (3IT) is interested in Memoristor (compression). Memories And Resistant) This electronic component, theoretically predicted in 1971 by Professor Leon Chua of Berkeley, was implemented only forty years later. It is a nanoscopic electronic component whose resistance can be changed at will. This property makes the memoirist a very promising candidate for the realization of artificial synapses within the circuit structure optimized for artificial intelligence (AI).

What we’re trying to do is combine the skills found in 3IT and IQ. We direct our research projects into the interfaces of artificial intelligence, emerging nanoelectronics and quantum science; You need three at the same time.

Ian Billiard, professor of engineering and IQ member

At 3IT, we are developing resistant memory, also called “memoryister”, the researcher continues. These are new nanocomponents that make it possible to develop high-performance electronic circuits specifically for AI. Then, three years ago, the concept Emerged: Why not apply these technologies to quantum computer scaling? Specifically, the question of whether to contribute to the automatic control of qubits using AI, whether one chooses a quantum point or even a superconducting circuit in silicon. If we want to scale quantum technology with thousands or even millions of cubits, we need to automate processes using AI and use classical control electronics which is very powerful.

The research team identifies artificial intelligence as a promising way to automate certain control mechanisms of the quantum system, from reading qubits to tomography, including controlling qubits and quantum devices.

To operate the artificial intelligence, high-powered computer hardware is required to avoid overheating the cryostat. You need optimized electronics to make everything work efficiently.

“Our team is collaborating with Roger Melko, a professor at the University of Waterloo and a researcher at the Perimeter Institute, and Stephanie Szczecin, a PhD applicant at Quantum AI,” explained Dominic Druin, a professor in the Faculty of Engineering. This research project is funded, among other things, by IQ calls for projects. And, since we needed experimental data to train neural networks for quantum points in silicon, we relied on the research work of Sophie Rochett and Julien Camerand Lemaire when they were doing their doctorate in Professor Michel Piero-Ladrier’s group. In addition to working on a variety of disciplinary interfaces and utilizing the resources of 3IT and IQ, we have relied on a collaborative approach. By combining all of these components, we’ve been able to demonstrate quantum dot self-drawing by machine learning, which is one step closer to automating certain methods. “

Next job

Ian Billiard reminds us of the temperature challenge: “We need cryogenic resistant memory, especially to be able to apply high-performance AI-based control systems directly to the cryostat, adapting to the operating limitations of the quantum system. On the other hand, all the memories developed so far are managed at room temperature. So the next step is to design a memoryister adapted to cryogenic conditions specifically to unlock all applications. So work needs to be done on both the material and the architecture of the components, especially using superconducting materials, first in the case of Memistrist. A

This is where the limits of nature may lie and the resources of 3IT, IQ and associates will come into their own.

Leave a Comment