The calibration of a Rigetti QPU connected to Quantum Machines’ control system was remotely automated by AI-powered tools from Quantum Elements and Qruise. This project was a component of the Israeli Quantum Computing Center’s “AI for Quantum Calibration Challenge.”
Today, Rigetti Computing (Nasdaq: RGTI), a leader in full-stack quantum-classical computing, and Quantum Machines, the top supplier of processor-based quantum controllers, reported the effective use of artificial intelligence (AI) to automate a quantum computer’s calibration. The project was finished as part of the Israeli Quantum Computing Center’s (IQCC) “AI for Quantum Calibration Challenge” (the “Challenge”), which was jointly issued by Rigetti and Quantum Machines.
A 9-qubit Rigetti NoveraTM QPU integrated with Quantum Machines’ cutting-edge OPX1000 control system and NVIDIA DGX Quantum, a unified system for quantum-classical computing that NVIDIA built with Quantum Machines, was automatically calibrated by the two Challenge participating companies, Quantum Elements and Qruise. This accomplishment demonstrates the growing maturity and cooperation within the quantum computing ecosystem in addition to the promise of AI in quantum computer calibration.
With the help of their software, Quantum Elements was able to set both individual and pair qubits with excellent accuracy, achieving 98.5% two-qubit gate fidelity and 99.9% single-qubit gate fidelity. To expedite the preparation process, Qruise used the control system to simultaneously fine-tune all nine qubits.
The industry struggles with more complicated calibration requirements as the number of qubits in quantum computing systems increases. According to Dr. Yonatan Cohen, co-founder and CTO of Quantum Machines, “one of the most important bottlenecks in scaling quantum systems is calibrating quantum computers.” Even a small quantum processor can now require weeks to months of expert effort to construct the model needed to calibrate it correctly, and hours—sometimes even days—are needed to repeat repeated calibrations.
Dozens of parameters need to be carefully adjusted for each qubit, and these parameters interact in intricate ways and drift over time. The problem gets significantly harder as we approach devices with thousands of qubits.
Our goals for larger quantum computers simply cannot be met by the conventional method of manual calibration by quantum physicists.
